{
  "best_global_step": 34344,
  "best_metric": 0.44693058729171753,
  "best_model_checkpoint": "saves_multiple/lntuning/llama-3-8b-instruct/train_codealpacapy_456_1765320135/checkpoint-34344",
  "epoch": 20.0,
  "eval_steps": 1908,
  "global_step": 38160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002620545073375262,
      "grad_norm": 0.8576380014419556,
      "learning_rate": 5.241090146750525e-08,
      "loss": 1.0078,
      "num_input_tokens_seen": 3808,
      "step": 5
    },
    {
      "epoch": 0.005241090146750524,
      "grad_norm": 1.7742892503738403,
      "learning_rate": 1.1792452830188679e-07,
      "loss": 1.4007,
      "num_input_tokens_seen": 6432,
      "step": 10
    },
    {
      "epoch": 0.007861635220125786,
      "grad_norm": 2.3607945442199707,
      "learning_rate": 1.8343815513626837e-07,
      "loss": 1.4172,
      "num_input_tokens_seen": 9120,
      "step": 15
    },
    {
      "epoch": 0.010482180293501049,
      "grad_norm": 1.459890365600586,
      "learning_rate": 2.489517819706499e-07,
      "loss": 1.2247,
      "num_input_tokens_seen": 11776,
      "step": 20
    },
    {
      "epoch": 0.01310272536687631,
      "grad_norm": 1.109479546546936,
      "learning_rate": 3.144654088050315e-07,
      "loss": 1.0888,
      "num_input_tokens_seen": 15968,
      "step": 25
    },
    {
      "epoch": 0.015723270440251572,
      "grad_norm": 0.8207584619522095,
      "learning_rate": 3.79979035639413e-07,
      "loss": 0.9561,
      "num_input_tokens_seen": 20704,
      "step": 30
    },
    {
      "epoch": 0.018343815513626835,
      "grad_norm": 1.782291293144226,
      "learning_rate": 4.454926624737945e-07,
      "loss": 1.33,
      "num_input_tokens_seen": 23264,
      "step": 35
    },
    {
      "epoch": 0.020964360587002098,
      "grad_norm": 4.1997199058532715,
      "learning_rate": 5.110062893081761e-07,
      "loss": 1.5667,
      "num_input_tokens_seen": 26784,
      "step": 40
    },
    {
      "epoch": 0.02358490566037736,
      "grad_norm": 1.0960667133331299,
      "learning_rate": 5.765199161425577e-07,
      "loss": 1.0799,
      "num_input_tokens_seen": 30144,
      "step": 45
    },
    {
      "epoch": 0.02620545073375262,
      "grad_norm": 1.0470774173736572,
      "learning_rate": 6.420335429769392e-07,
      "loss": 1.1201,
      "num_input_tokens_seen": 33216,
      "step": 50
    },
    {
      "epoch": 0.028825995807127882,
      "grad_norm": 1.0686659812927246,
      "learning_rate": 7.075471698113208e-07,
      "loss": 1.0534,
      "num_input_tokens_seen": 36768,
      "step": 55
    },
    {
      "epoch": 0.031446540880503145,
      "grad_norm": 1.2562153339385986,
      "learning_rate": 7.730607966457024e-07,
      "loss": 1.2616,
      "num_input_tokens_seen": 40288,
      "step": 60
    },
    {
      "epoch": 0.034067085953878404,
      "grad_norm": 1.2569166421890259,
      "learning_rate": 8.38574423480084e-07,
      "loss": 1.2013,
      "num_input_tokens_seen": 43520,
      "step": 65
    },
    {
      "epoch": 0.03668763102725367,
      "grad_norm": 1.805755853652954,
      "learning_rate": 9.040880503144654e-07,
      "loss": 1.0492,
      "num_input_tokens_seen": 46816,
      "step": 70
    },
    {
      "epoch": 0.03930817610062893,
      "grad_norm": 0.8431106209754944,
      "learning_rate": 9.69601677148847e-07,
      "loss": 1.0394,
      "num_input_tokens_seen": 49664,
      "step": 75
    },
    {
      "epoch": 0.041928721174004195,
      "grad_norm": 0.7984536290168762,
      "learning_rate": 1.0351153039832286e-06,
      "loss": 1.1739,
      "num_input_tokens_seen": 53088,
      "step": 80
    },
    {
      "epoch": 0.044549266247379454,
      "grad_norm": 1.0905178785324097,
      "learning_rate": 1.10062893081761e-06,
      "loss": 1.0417,
      "num_input_tokens_seen": 56608,
      "step": 85
    },
    {
      "epoch": 0.04716981132075472,
      "grad_norm": 1.7987223863601685,
      "learning_rate": 1.1661425576519918e-06,
      "loss": 1.241,
      "num_input_tokens_seen": 59904,
      "step": 90
    },
    {
      "epoch": 0.04979035639412998,
      "grad_norm": 1.2431687116622925,
      "learning_rate": 1.2316561844863732e-06,
      "loss": 0.9958,
      "num_input_tokens_seen": 63392,
      "step": 95
    },
    {
      "epoch": 0.05241090146750524,
      "grad_norm": 1.4009389877319336,
      "learning_rate": 1.2971698113207548e-06,
      "loss": 1.2893,
      "num_input_tokens_seen": 65952,
      "step": 100
    },
    {
      "epoch": 0.055031446540880505,
      "grad_norm": 1.9212751388549805,
      "learning_rate": 1.3626834381551364e-06,
      "loss": 1.1346,
      "num_input_tokens_seen": 68672,
      "step": 105
    },
    {
      "epoch": 0.057651991614255764,
      "grad_norm": 1.3107408285140991,
      "learning_rate": 1.4281970649895177e-06,
      "loss": 1.2134,
      "num_input_tokens_seen": 71712,
      "step": 110
    },
    {
      "epoch": 0.06027253668763103,
      "grad_norm": 1.447460651397705,
      "learning_rate": 1.4937106918238995e-06,
      "loss": 1.4448,
      "num_input_tokens_seen": 74656,
      "step": 115
    },
    {
      "epoch": 0.06289308176100629,
      "grad_norm": 1.556396245956421,
      "learning_rate": 1.559224318658281e-06,
      "loss": 1.0668,
      "num_input_tokens_seen": 78208,
      "step": 120
    },
    {
      "epoch": 0.06551362683438156,
      "grad_norm": 1.7310025691986084,
      "learning_rate": 1.6247379454926625e-06,
      "loss": 1.1614,
      "num_input_tokens_seen": 81376,
      "step": 125
    },
    {
      "epoch": 0.06813417190775681,
      "grad_norm": 1.0190327167510986,
      "learning_rate": 1.690251572327044e-06,
      "loss": 1.3953,
      "num_input_tokens_seen": 84288,
      "step": 130
    },
    {
      "epoch": 0.07075471698113207,
      "grad_norm": 2.203998565673828,
      "learning_rate": 1.7557651991614257e-06,
      "loss": 1.485,
      "num_input_tokens_seen": 86976,
      "step": 135
    },
    {
      "epoch": 0.07337526205450734,
      "grad_norm": 1.2534780502319336,
      "learning_rate": 1.821278825995807e-06,
      "loss": 1.1821,
      "num_input_tokens_seen": 89952,
      "step": 140
    },
    {
      "epoch": 0.0759958071278826,
      "grad_norm": 0.8065716624259949,
      "learning_rate": 1.8867924528301887e-06,
      "loss": 1.0346,
      "num_input_tokens_seen": 93120,
      "step": 145
    },
    {
      "epoch": 0.07861635220125786,
      "grad_norm": 2.4441771507263184,
      "learning_rate": 1.9523060796645703e-06,
      "loss": 1.2785,
      "num_input_tokens_seen": 96640,
      "step": 150
    },
    {
      "epoch": 0.08123689727463312,
      "grad_norm": 1.0667129755020142,
      "learning_rate": 2.017819706498952e-06,
      "loss": 1.0591,
      "num_input_tokens_seen": 99744,
      "step": 155
    },
    {
      "epoch": 0.08385744234800839,
      "grad_norm": 1.5128192901611328,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 1.1376,
      "num_input_tokens_seen": 102720,
      "step": 160
    },
    {
      "epoch": 0.08647798742138364,
      "grad_norm": 1.031844973564148,
      "learning_rate": 2.148846960167715e-06,
      "loss": 1.2913,
      "num_input_tokens_seen": 105344,
      "step": 165
    },
    {
      "epoch": 0.08909853249475891,
      "grad_norm": 1.8251255750656128,
      "learning_rate": 2.2143605870020966e-06,
      "loss": 1.3727,
      "num_input_tokens_seen": 108576,
      "step": 170
    },
    {
      "epoch": 0.09171907756813417,
      "grad_norm": 1.7584495544433594,
      "learning_rate": 2.279874213836478e-06,
      "loss": 1.0894,
      "num_input_tokens_seen": 111520,
      "step": 175
    },
    {
      "epoch": 0.09433962264150944,
      "grad_norm": 1.3092713356018066,
      "learning_rate": 2.34538784067086e-06,
      "loss": 1.4876,
      "num_input_tokens_seen": 114400,
      "step": 180
    },
    {
      "epoch": 0.09696016771488469,
      "grad_norm": 1.9502735137939453,
      "learning_rate": 2.410901467505241e-06,
      "loss": 0.9748,
      "num_input_tokens_seen": 118464,
      "step": 185
    },
    {
      "epoch": 0.09958071278825996,
      "grad_norm": 1.1691211462020874,
      "learning_rate": 2.476415094339623e-06,
      "loss": 1.3059,
      "num_input_tokens_seen": 121376,
      "step": 190
    },
    {
      "epoch": 0.10220125786163523,
      "grad_norm": 1.8256959915161133,
      "learning_rate": 2.5419287211740044e-06,
      "loss": 1.2696,
      "num_input_tokens_seen": 124800,
      "step": 195
    },
    {
      "epoch": 0.10482180293501048,
      "grad_norm": 0.8637256622314453,
      "learning_rate": 2.6074423480083857e-06,
      "loss": 1.1302,
      "num_input_tokens_seen": 127840,
      "step": 200
    },
    {
      "epoch": 0.10744234800838574,
      "grad_norm": 1.0522990226745605,
      "learning_rate": 2.6729559748427675e-06,
      "loss": 1.2698,
      "num_input_tokens_seen": 131296,
      "step": 205
    },
    {
      "epoch": 0.11006289308176101,
      "grad_norm": 2.038963556289673,
      "learning_rate": 2.738469601677149e-06,
      "loss": 1.3719,
      "num_input_tokens_seen": 133760,
      "step": 210
    },
    {
      "epoch": 0.11268343815513626,
      "grad_norm": 1.0614789724349976,
      "learning_rate": 2.8039832285115303e-06,
      "loss": 1.1731,
      "num_input_tokens_seen": 137536,
      "step": 215
    },
    {
      "epoch": 0.11530398322851153,
      "grad_norm": 0.8436765074729919,
      "learning_rate": 2.869496855345912e-06,
      "loss": 1.1073,
      "num_input_tokens_seen": 141568,
      "step": 220
    },
    {
      "epoch": 0.1179245283018868,
      "grad_norm": 0.6172121167182922,
      "learning_rate": 2.935010482180294e-06,
      "loss": 1.0602,
      "num_input_tokens_seen": 145248,
      "step": 225
    },
    {
      "epoch": 0.12054507337526206,
      "grad_norm": 1.1988259553909302,
      "learning_rate": 3.0005241090146753e-06,
      "loss": 1.1589,
      "num_input_tokens_seen": 148576,
      "step": 230
    },
    {
      "epoch": 0.12316561844863731,
      "grad_norm": 1.5790709257125854,
      "learning_rate": 3.0660377358490567e-06,
      "loss": 1.2804,
      "num_input_tokens_seen": 151360,
      "step": 235
    },
    {
      "epoch": 0.12578616352201258,
      "grad_norm": 1.8650583028793335,
      "learning_rate": 3.131551362683438e-06,
      "loss": 1.1058,
      "num_input_tokens_seen": 154144,
      "step": 240
    },
    {
      "epoch": 0.12840670859538783,
      "grad_norm": 0.7069394588470459,
      "learning_rate": 3.1970649895178203e-06,
      "loss": 1.0773,
      "num_input_tokens_seen": 157600,
      "step": 245
    },
    {
      "epoch": 0.1310272536687631,
      "grad_norm": 1.2779381275177002,
      "learning_rate": 3.2625786163522012e-06,
      "loss": 0.9816,
      "num_input_tokens_seen": 160832,
      "step": 250
    },
    {
      "epoch": 0.13364779874213836,
      "grad_norm": 1.3946716785430908,
      "learning_rate": 3.3280922431865826e-06,
      "loss": 1.2122,
      "num_input_tokens_seen": 163968,
      "step": 255
    },
    {
      "epoch": 0.13626834381551362,
      "grad_norm": 0.8945034146308899,
      "learning_rate": 3.393605870020965e-06,
      "loss": 1.0559,
      "num_input_tokens_seen": 168992,
      "step": 260
    },
    {
      "epoch": 0.1388888888888889,
      "grad_norm": 1.033284306526184,
      "learning_rate": 3.459119496855346e-06,
      "loss": 1.1413,
      "num_input_tokens_seen": 171616,
      "step": 265
    },
    {
      "epoch": 0.14150943396226415,
      "grad_norm": 1.5818067789077759,
      "learning_rate": 3.524633123689727e-06,
      "loss": 1.0234,
      "num_input_tokens_seen": 175680,
      "step": 270
    },
    {
      "epoch": 0.1441299790356394,
      "grad_norm": 0.8729956746101379,
      "learning_rate": 3.5901467505241094e-06,
      "loss": 1.008,
      "num_input_tokens_seen": 178880,
      "step": 275
    },
    {
      "epoch": 0.14675052410901468,
      "grad_norm": 1.0367937088012695,
      "learning_rate": 3.6556603773584908e-06,
      "loss": 1.0543,
      "num_input_tokens_seen": 182368,
      "step": 280
    },
    {
      "epoch": 0.14937106918238993,
      "grad_norm": 1.3457082509994507,
      "learning_rate": 3.7211740041928726e-06,
      "loss": 1.0406,
      "num_input_tokens_seen": 185792,
      "step": 285
    },
    {
      "epoch": 0.1519916142557652,
      "grad_norm": 1.676633596420288,
      "learning_rate": 3.786687631027254e-06,
      "loss": 0.9836,
      "num_input_tokens_seen": 189664,
      "step": 290
    },
    {
      "epoch": 0.15461215932914046,
      "grad_norm": 1.0386903285980225,
      "learning_rate": 3.852201257861635e-06,
      "loss": 1.2058,
      "num_input_tokens_seen": 193632,
      "step": 295
    },
    {
      "epoch": 0.15723270440251572,
      "grad_norm": 0.8702308535575867,
      "learning_rate": 3.917714884696017e-06,
      "loss": 1.2492,
      "num_input_tokens_seen": 196800,
      "step": 300
    },
    {
      "epoch": 0.159853249475891,
      "grad_norm": 1.896450161933899,
      "learning_rate": 3.9832285115303985e-06,
      "loss": 1.0738,
      "num_input_tokens_seen": 199744,
      "step": 305
    },
    {
      "epoch": 0.16247379454926625,
      "grad_norm": 1.2631577253341675,
      "learning_rate": 4.04874213836478e-06,
      "loss": 1.1527,
      "num_input_tokens_seen": 202624,
      "step": 310
    },
    {
      "epoch": 0.1650943396226415,
      "grad_norm": 1.240329623222351,
      "learning_rate": 4.114255765199162e-06,
      "loss": 1.2458,
      "num_input_tokens_seen": 205632,
      "step": 315
    },
    {
      "epoch": 0.16771488469601678,
      "grad_norm": 1.3951010704040527,
      "learning_rate": 4.1797693920335435e-06,
      "loss": 0.9968,
      "num_input_tokens_seen": 209280,
      "step": 320
    },
    {
      "epoch": 0.17033542976939203,
      "grad_norm": 0.6393210291862488,
      "learning_rate": 4.245283018867925e-06,
      "loss": 1.0812,
      "num_input_tokens_seen": 213344,
      "step": 325
    },
    {
      "epoch": 0.17295597484276728,
      "grad_norm": 2.5017921924591064,
      "learning_rate": 4.310796645702306e-06,
      "loss": 1.0162,
      "num_input_tokens_seen": 216448,
      "step": 330
    },
    {
      "epoch": 0.17557651991614256,
      "grad_norm": 1.3614442348480225,
      "learning_rate": 4.376310272536688e-06,
      "loss": 1.2816,
      "num_input_tokens_seen": 218976,
      "step": 335
    },
    {
      "epoch": 0.17819706498951782,
      "grad_norm": 1.4935790300369263,
      "learning_rate": 4.441823899371069e-06,
      "loss": 0.9566,
      "num_input_tokens_seen": 222080,
      "step": 340
    },
    {
      "epoch": 0.18081761006289307,
      "grad_norm": 1.17710542678833,
      "learning_rate": 4.507337526205451e-06,
      "loss": 1.2288,
      "num_input_tokens_seen": 225120,
      "step": 345
    },
    {
      "epoch": 0.18343815513626835,
      "grad_norm": 1.2805399894714355,
      "learning_rate": 4.572851153039833e-06,
      "loss": 1.4487,
      "num_input_tokens_seen": 227872,
      "step": 350
    },
    {
      "epoch": 0.1860587002096436,
      "grad_norm": 1.049477458000183,
      "learning_rate": 4.638364779874214e-06,
      "loss": 1.0933,
      "num_input_tokens_seen": 232320,
      "step": 355
    },
    {
      "epoch": 0.18867924528301888,
      "grad_norm": 1.8528871536254883,
      "learning_rate": 4.703878406708595e-06,
      "loss": 1.1348,
      "num_input_tokens_seen": 235104,
      "step": 360
    },
    {
      "epoch": 0.19129979035639413,
      "grad_norm": 1.229369044303894,
      "learning_rate": 4.769392033542977e-06,
      "loss": 1.3333,
      "num_input_tokens_seen": 237792,
      "step": 365
    },
    {
      "epoch": 0.19392033542976939,
      "grad_norm": 1.3280261754989624,
      "learning_rate": 4.834905660377359e-06,
      "loss": 0.9536,
      "num_input_tokens_seen": 241728,
      "step": 370
    },
    {
      "epoch": 0.19654088050314467,
      "grad_norm": 2.8456172943115234,
      "learning_rate": 4.90041928721174e-06,
      "loss": 1.091,
      "num_input_tokens_seen": 245696,
      "step": 375
    },
    {
      "epoch": 0.19916142557651992,
      "grad_norm": 0.7012678980827332,
      "learning_rate": 4.965932914046122e-06,
      "loss": 0.9418,
      "num_input_tokens_seen": 249408,
      "step": 380
    },
    {
      "epoch": 0.20178197064989517,
      "grad_norm": 1.7613924741744995,
      "learning_rate": 5.031446540880504e-06,
      "loss": 1.1821,
      "num_input_tokens_seen": 252160,
      "step": 385
    },
    {
      "epoch": 0.20440251572327045,
      "grad_norm": 1.660179615020752,
      "learning_rate": 5.096960167714885e-06,
      "loss": 1.3011,
      "num_input_tokens_seen": 255552,
      "step": 390
    },
    {
      "epoch": 0.2070230607966457,
      "grad_norm": 1.5295683145523071,
      "learning_rate": 5.162473794549266e-06,
      "loss": 1.1396,
      "num_input_tokens_seen": 259072,
      "step": 395
    },
    {
      "epoch": 0.20964360587002095,
      "grad_norm": 1.0136443376541138,
      "learning_rate": 5.227987421383648e-06,
      "loss": 0.9934,
      "num_input_tokens_seen": 263008,
      "step": 400
    },
    {
      "epoch": 0.21226415094339623,
      "grad_norm": 1.1560345888137817,
      "learning_rate": 5.2935010482180295e-06,
      "loss": 1.075,
      "num_input_tokens_seen": 267424,
      "step": 405
    },
    {
      "epoch": 0.2148846960167715,
      "grad_norm": 0.8486748933792114,
      "learning_rate": 5.359014675052411e-06,
      "loss": 1.3975,
      "num_input_tokens_seen": 270720,
      "step": 410
    },
    {
      "epoch": 0.21750524109014674,
      "grad_norm": 3.523170232772827,
      "learning_rate": 5.424528301886793e-06,
      "loss": 1.3013,
      "num_input_tokens_seen": 273568,
      "step": 415
    },
    {
      "epoch": 0.22012578616352202,
      "grad_norm": 1.5573949813842773,
      "learning_rate": 5.4900419287211744e-06,
      "loss": 1.1609,
      "num_input_tokens_seen": 276224,
      "step": 420
    },
    {
      "epoch": 0.22274633123689727,
      "grad_norm": 2.1238062381744385,
      "learning_rate": 5.555555555555556e-06,
      "loss": 1.3745,
      "num_input_tokens_seen": 279040,
      "step": 425
    },
    {
      "epoch": 0.22536687631027252,
      "grad_norm": 1.213826060295105,
      "learning_rate": 5.621069182389937e-06,
      "loss": 1.1753,
      "num_input_tokens_seen": 282048,
      "step": 430
    },
    {
      "epoch": 0.2279874213836478,
      "grad_norm": 1.1245131492614746,
      "learning_rate": 5.686582809224319e-06,
      "loss": 0.968,
      "num_input_tokens_seen": 285184,
      "step": 435
    },
    {
      "epoch": 0.23060796645702306,
      "grad_norm": 1.080130696296692,
      "learning_rate": 5.752096436058701e-06,
      "loss": 0.9999,
      "num_input_tokens_seen": 289152,
      "step": 440
    },
    {
      "epoch": 0.23322851153039834,
      "grad_norm": 0.7767438888549805,
      "learning_rate": 5.817610062893082e-06,
      "loss": 1.122,
      "num_input_tokens_seen": 292768,
      "step": 445
    },
    {
      "epoch": 0.2358490566037736,
      "grad_norm": 0.5710167288780212,
      "learning_rate": 5.8831236897274636e-06,
      "loss": 1.0067,
      "num_input_tokens_seen": 296128,
      "step": 450
    },
    {
      "epoch": 0.23846960167714884,
      "grad_norm": 1.7413545846939087,
      "learning_rate": 5.948637316561846e-06,
      "loss": 0.9036,
      "num_input_tokens_seen": 300544,
      "step": 455
    },
    {
      "epoch": 0.24109014675052412,
      "grad_norm": 1.4621224403381348,
      "learning_rate": 6.014150943396227e-06,
      "loss": 1.0581,
      "num_input_tokens_seen": 303200,
      "step": 460
    },
    {
      "epoch": 0.24371069182389937,
      "grad_norm": 0.7455990314483643,
      "learning_rate": 6.079664570230608e-06,
      "loss": 1.0435,
      "num_input_tokens_seen": 306368,
      "step": 465
    },
    {
      "epoch": 0.24633123689727462,
      "grad_norm": 1.1539311408996582,
      "learning_rate": 6.14517819706499e-06,
      "loss": 1.074,
      "num_input_tokens_seen": 309952,
      "step": 470
    },
    {
      "epoch": 0.2489517819706499,
      "grad_norm": 1.4379374980926514,
      "learning_rate": 6.210691823899371e-06,
      "loss": 1.0562,
      "num_input_tokens_seen": 312928,
      "step": 475
    },
    {
      "epoch": 0.25157232704402516,
      "grad_norm": 1.0077040195465088,
      "learning_rate": 6.2762054507337535e-06,
      "loss": 1.1057,
      "num_input_tokens_seen": 316160,
      "step": 480
    },
    {
      "epoch": 0.25419287211740044,
      "grad_norm": 1.056850552558899,
      "learning_rate": 6.341719077568134e-06,
      "loss": 0.9344,
      "num_input_tokens_seen": 319136,
      "step": 485
    },
    {
      "epoch": 0.25681341719077566,
      "grad_norm": 1.021744966506958,
      "learning_rate": 6.407232704402516e-06,
      "loss": 1.0499,
      "num_input_tokens_seen": 321824,
      "step": 490
    },
    {
      "epoch": 0.25943396226415094,
      "grad_norm": 3.329357862472534,
      "learning_rate": 6.472746331236898e-06,
      "loss": 1.6399,
      "num_input_tokens_seen": 324864,
      "step": 495
    },
    {
      "epoch": 0.2620545073375262,
      "grad_norm": 2.3047826290130615,
      "learning_rate": 6.538259958071278e-06,
      "loss": 1.3541,
      "num_input_tokens_seen": 327264,
      "step": 500
    },
    {
      "epoch": 0.26467505241090145,
      "grad_norm": 0.8381253480911255,
      "learning_rate": 6.60377358490566e-06,
      "loss": 1.1758,
      "num_input_tokens_seen": 330656,
      "step": 505
    },
    {
      "epoch": 0.2672955974842767,
      "grad_norm": 0.6899804472923279,
      "learning_rate": 6.669287211740043e-06,
      "loss": 0.8043,
      "num_input_tokens_seen": 335776,
      "step": 510
    },
    {
      "epoch": 0.269916142557652,
      "grad_norm": 1.4705612659454346,
      "learning_rate": 6.734800838574423e-06,
      "loss": 1.3665,
      "num_input_tokens_seen": 339392,
      "step": 515
    },
    {
      "epoch": 0.27253668763102723,
      "grad_norm": 0.9645290970802307,
      "learning_rate": 6.800314465408805e-06,
      "loss": 1.1238,
      "num_input_tokens_seen": 342848,
      "step": 520
    },
    {
      "epoch": 0.2751572327044025,
      "grad_norm": 0.9003207087516785,
      "learning_rate": 6.865828092243188e-06,
      "loss": 1.1249,
      "num_input_tokens_seen": 345600,
      "step": 525
    },
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 0.8326305747032166,
      "learning_rate": 6.931341719077568e-06,
      "loss": 1.1735,
      "num_input_tokens_seen": 348960,
      "step": 530
    },
    {
      "epoch": 0.280398322851153,
      "grad_norm": 1.9467405080795288,
      "learning_rate": 6.9968553459119495e-06,
      "loss": 1.0548,
      "num_input_tokens_seen": 351520,
      "step": 535
    },
    {
      "epoch": 0.2830188679245283,
      "grad_norm": 1.7447394132614136,
      "learning_rate": 7.062368972746332e-06,
      "loss": 1.3153,
      "num_input_tokens_seen": 355456,
      "step": 540
    },
    {
      "epoch": 0.2856394129979036,
      "grad_norm": 1.2435569763183594,
      "learning_rate": 7.127882599580712e-06,
      "loss": 1.531,
      "num_input_tokens_seen": 358144,
      "step": 545
    },
    {
      "epoch": 0.2882599580712788,
      "grad_norm": 0.9154888391494751,
      "learning_rate": 7.1933962264150945e-06,
      "loss": 0.7643,
      "num_input_tokens_seen": 363104,
      "step": 550
    },
    {
      "epoch": 0.2908805031446541,
      "grad_norm": 1.2529715299606323,
      "learning_rate": 7.258909853249477e-06,
      "loss": 0.9385,
      "num_input_tokens_seen": 367200,
      "step": 555
    },
    {
      "epoch": 0.29350104821802936,
      "grad_norm": 0.8947097659111023,
      "learning_rate": 7.324423480083858e-06,
      "loss": 1.2087,
      "num_input_tokens_seen": 370176,
      "step": 560
    },
    {
      "epoch": 0.29612159329140464,
      "grad_norm": 2.761025905609131,
      "learning_rate": 7.3899371069182395e-06,
      "loss": 1.3822,
      "num_input_tokens_seen": 372832,
      "step": 565
    },
    {
      "epoch": 0.29874213836477986,
      "grad_norm": 0.8652147650718689,
      "learning_rate": 7.455450733752621e-06,
      "loss": 1.118,
      "num_input_tokens_seen": 375296,
      "step": 570
    },
    {
      "epoch": 0.30136268343815514,
      "grad_norm": 1.2671217918395996,
      "learning_rate": 7.520964360587003e-06,
      "loss": 0.9242,
      "num_input_tokens_seen": 378464,
      "step": 575
    },
    {
      "epoch": 0.3039832285115304,
      "grad_norm": 2.4024317264556885,
      "learning_rate": 7.586477987421384e-06,
      "loss": 0.9831,
      "num_input_tokens_seen": 381664,
      "step": 580
    },
    {
      "epoch": 0.30660377358490565,
      "grad_norm": 1.4057327508926392,
      "learning_rate": 7.651991614255765e-06,
      "loss": 1.0043,
      "num_input_tokens_seen": 385248,
      "step": 585
    },
    {
      "epoch": 0.30922431865828093,
      "grad_norm": 1.5841363668441772,
      "learning_rate": 7.717505241090148e-06,
      "loss": 1.0741,
      "num_input_tokens_seen": 387840,
      "step": 590
    },
    {
      "epoch": 0.3118448637316562,
      "grad_norm": 1.4186307191848755,
      "learning_rate": 7.783018867924528e-06,
      "loss": 0.9385,
      "num_input_tokens_seen": 391424,
      "step": 595
    },
    {
      "epoch": 0.31446540880503143,
      "grad_norm": 1.6323543787002563,
      "learning_rate": 7.848532494758911e-06,
      "loss": 0.9774,
      "num_input_tokens_seen": 394656,
      "step": 600
    },
    {
      "epoch": 0.3170859538784067,
      "grad_norm": 1.275531530380249,
      "learning_rate": 7.914046121593292e-06,
      "loss": 1.1123,
      "num_input_tokens_seen": 397824,
      "step": 605
    },
    {
      "epoch": 0.319706498951782,
      "grad_norm": 1.606722354888916,
      "learning_rate": 7.979559748427674e-06,
      "loss": 1.0916,
      "num_input_tokens_seen": 400544,
      "step": 610
    },
    {
      "epoch": 0.3223270440251572,
      "grad_norm": 1.222265362739563,
      "learning_rate": 8.045073375262055e-06,
      "loss": 1.0187,
      "num_input_tokens_seen": 403552,
      "step": 615
    },
    {
      "epoch": 0.3249475890985325,
      "grad_norm": 1.9850631952285767,
      "learning_rate": 8.110587002096436e-06,
      "loss": 1.0671,
      "num_input_tokens_seen": 406464,
      "step": 620
    },
    {
      "epoch": 0.3275681341719078,
      "grad_norm": 2.0015807151794434,
      "learning_rate": 8.176100628930818e-06,
      "loss": 0.8142,
      "num_input_tokens_seen": 412032,
      "step": 625
    },
    {
      "epoch": 0.330188679245283,
      "grad_norm": 0.9350792169570923,
      "learning_rate": 8.241614255765199e-06,
      "loss": 0.9357,
      "num_input_tokens_seen": 415104,
      "step": 630
    },
    {
      "epoch": 0.3328092243186583,
      "grad_norm": 1.2293148040771484,
      "learning_rate": 8.307127882599582e-06,
      "loss": 0.9541,
      "num_input_tokens_seen": 418368,
      "step": 635
    },
    {
      "epoch": 0.33542976939203356,
      "grad_norm": 1.669934630393982,
      "learning_rate": 8.372641509433962e-06,
      "loss": 1.1969,
      "num_input_tokens_seen": 420928,
      "step": 640
    },
    {
      "epoch": 0.3380503144654088,
      "grad_norm": 2.0326383113861084,
      "learning_rate": 8.438155136268345e-06,
      "loss": 0.8785,
      "num_input_tokens_seen": 423840,
      "step": 645
    },
    {
      "epoch": 0.34067085953878407,
      "grad_norm": 0.8375669717788696,
      "learning_rate": 8.503668763102726e-06,
      "loss": 0.8626,
      "num_input_tokens_seen": 427904,
      "step": 650
    },
    {
      "epoch": 0.34329140461215935,
      "grad_norm": 1.2467238903045654,
      "learning_rate": 8.569182389937106e-06,
      "loss": 1.2738,
      "num_input_tokens_seen": 430304,
      "step": 655
    },
    {
      "epoch": 0.34591194968553457,
      "grad_norm": 1.7424036264419556,
      "learning_rate": 8.634696016771489e-06,
      "loss": 1.1578,
      "num_input_tokens_seen": 433088,
      "step": 660
    },
    {
      "epoch": 0.34853249475890985,
      "grad_norm": 1.0932422876358032,
      "learning_rate": 8.70020964360587e-06,
      "loss": 0.909,
      "num_input_tokens_seen": 437120,
      "step": 665
    },
    {
      "epoch": 0.35115303983228513,
      "grad_norm": 0.7657503485679626,
      "learning_rate": 8.765723270440252e-06,
      "loss": 0.6692,
      "num_input_tokens_seen": 441440,
      "step": 670
    },
    {
      "epoch": 0.35377358490566035,
      "grad_norm": 1.9715607166290283,
      "learning_rate": 8.831236897274633e-06,
      "loss": 1.0365,
      "num_input_tokens_seen": 444000,
      "step": 675
    },
    {
      "epoch": 0.35639412997903563,
      "grad_norm": 0.9620529413223267,
      "learning_rate": 8.896750524109015e-06,
      "loss": 1.0494,
      "num_input_tokens_seen": 447232,
      "step": 680
    },
    {
      "epoch": 0.3590146750524109,
      "grad_norm": 1.0355297327041626,
      "learning_rate": 8.962264150943396e-06,
      "loss": 0.8914,
      "num_input_tokens_seen": 450432,
      "step": 685
    },
    {
      "epoch": 0.36163522012578614,
      "grad_norm": 1.157413363456726,
      "learning_rate": 9.027777777777777e-06,
      "loss": 0.8638,
      "num_input_tokens_seen": 453440,
      "step": 690
    },
    {
      "epoch": 0.3642557651991614,
      "grad_norm": 1.2370481491088867,
      "learning_rate": 9.09329140461216e-06,
      "loss": 0.8908,
      "num_input_tokens_seen": 456352,
      "step": 695
    },
    {
      "epoch": 0.3668763102725367,
      "grad_norm": 1.4509947299957275,
      "learning_rate": 9.15880503144654e-06,
      "loss": 0.8893,
      "num_input_tokens_seen": 459712,
      "step": 700
    },
    {
      "epoch": 0.3694968553459119,
      "grad_norm": 0.787177324295044,
      "learning_rate": 9.224318658280923e-06,
      "loss": 0.7778,
      "num_input_tokens_seen": 463168,
      "step": 705
    },
    {
      "epoch": 0.3721174004192872,
      "grad_norm": 1.7847453355789185,
      "learning_rate": 9.289832285115305e-06,
      "loss": 1.0689,
      "num_input_tokens_seen": 465824,
      "step": 710
    },
    {
      "epoch": 0.3747379454926625,
      "grad_norm": 1.8353153467178345,
      "learning_rate": 9.355345911949686e-06,
      "loss": 1.1545,
      "num_input_tokens_seen": 468032,
      "step": 715
    },
    {
      "epoch": 0.37735849056603776,
      "grad_norm": 0.8285340070724487,
      "learning_rate": 9.420859538784067e-06,
      "loss": 0.7687,
      "num_input_tokens_seen": 471392,
      "step": 720
    },
    {
      "epoch": 0.379979035639413,
      "grad_norm": 1.8423426151275635,
      "learning_rate": 9.486373165618449e-06,
      "loss": 0.9454,
      "num_input_tokens_seen": 474400,
      "step": 725
    },
    {
      "epoch": 0.38259958071278827,
      "grad_norm": 1.578715205192566,
      "learning_rate": 9.551886792452832e-06,
      "loss": 0.8995,
      "num_input_tokens_seen": 477792,
      "step": 730
    },
    {
      "epoch": 0.38522012578616355,
      "grad_norm": 1.3625986576080322,
      "learning_rate": 9.617400419287211e-06,
      "loss": 1.001,
      "num_input_tokens_seen": 481216,
      "step": 735
    },
    {
      "epoch": 0.38784067085953877,
      "grad_norm": 0.75567626953125,
      "learning_rate": 9.682914046121595e-06,
      "loss": 0.9721,
      "num_input_tokens_seen": 484928,
      "step": 740
    },
    {
      "epoch": 0.39046121593291405,
      "grad_norm": 1.1038917303085327,
      "learning_rate": 9.748427672955976e-06,
      "loss": 0.8701,
      "num_input_tokens_seen": 488224,
      "step": 745
    },
    {
      "epoch": 0.39308176100628933,
      "grad_norm": 1.0793794393539429,
      "learning_rate": 9.813941299790357e-06,
      "loss": 0.9505,
      "num_input_tokens_seen": 491840,
      "step": 750
    },
    {
      "epoch": 0.39570230607966456,
      "grad_norm": 2.502938747406006,
      "learning_rate": 9.879454926624739e-06,
      "loss": 1.0154,
      "num_input_tokens_seen": 495232,
      "step": 755
    },
    {
      "epoch": 0.39832285115303984,
      "grad_norm": 0.817970335483551,
      "learning_rate": 9.94496855345912e-06,
      "loss": 0.8388,
      "num_input_tokens_seen": 498336,
      "step": 760
    },
    {
      "epoch": 0.4009433962264151,
      "grad_norm": 1.2297344207763672,
      "learning_rate": 1.0010482180293501e-05,
      "loss": 1.1694,
      "num_input_tokens_seen": 500448,
      "step": 765
    },
    {
      "epoch": 0.40356394129979034,
      "grad_norm": 0.7389943599700928,
      "learning_rate": 1.0075995807127883e-05,
      "loss": 0.824,
      "num_input_tokens_seen": 503520,
      "step": 770
    },
    {
      "epoch": 0.4061844863731656,
      "grad_norm": 1.3234213590621948,
      "learning_rate": 1.0141509433962266e-05,
      "loss": 0.9714,
      "num_input_tokens_seen": 506208,
      "step": 775
    },
    {
      "epoch": 0.4088050314465409,
      "grad_norm": 1.3816746473312378,
      "learning_rate": 1.0207023060796646e-05,
      "loss": 0.8534,
      "num_input_tokens_seen": 509152,
      "step": 780
    },
    {
      "epoch": 0.4114255765199161,
      "grad_norm": 0.9273484945297241,
      "learning_rate": 1.0272536687631027e-05,
      "loss": 0.9049,
      "num_input_tokens_seen": 513184,
      "step": 785
    },
    {
      "epoch": 0.4140461215932914,
      "grad_norm": 1.5392911434173584,
      "learning_rate": 1.033805031446541e-05,
      "loss": 1.1008,
      "num_input_tokens_seen": 516128,
      "step": 790
    },
    {
      "epoch": 0.4166666666666667,
      "grad_norm": 0.7226734161376953,
      "learning_rate": 1.040356394129979e-05,
      "loss": 0.7469,
      "num_input_tokens_seen": 519264,
      "step": 795
    },
    {
      "epoch": 0.4192872117400419,
      "grad_norm": 1.392201542854309,
      "learning_rate": 1.0469077568134173e-05,
      "loss": 1.0483,
      "num_input_tokens_seen": 522304,
      "step": 800
    },
    {
      "epoch": 0.4219077568134172,
      "grad_norm": 1.6442248821258545,
      "learning_rate": 1.0534591194968554e-05,
      "loss": 0.7602,
      "num_input_tokens_seen": 525600,
      "step": 805
    },
    {
      "epoch": 0.42452830188679247,
      "grad_norm": 1.0574039220809937,
      "learning_rate": 1.0600104821802936e-05,
      "loss": 0.9605,
      "num_input_tokens_seen": 528608,
      "step": 810
    },
    {
      "epoch": 0.4271488469601677,
      "grad_norm": 1.1016442775726318,
      "learning_rate": 1.0665618448637317e-05,
      "loss": 0.7984,
      "num_input_tokens_seen": 531232,
      "step": 815
    },
    {
      "epoch": 0.429769392033543,
      "grad_norm": 1.1845842599868774,
      "learning_rate": 1.0731132075471698e-05,
      "loss": 0.7452,
      "num_input_tokens_seen": 534208,
      "step": 820
    },
    {
      "epoch": 0.43238993710691825,
      "grad_norm": 1.3613451719284058,
      "learning_rate": 1.079664570230608e-05,
      "loss": 0.8217,
      "num_input_tokens_seen": 537792,
      "step": 825
    },
    {
      "epoch": 0.4350104821802935,
      "grad_norm": 0.9158235788345337,
      "learning_rate": 1.0862159329140461e-05,
      "loss": 1.0275,
      "num_input_tokens_seen": 540448,
      "step": 830
    },
    {
      "epoch": 0.43763102725366876,
      "grad_norm": 1.0936684608459473,
      "learning_rate": 1.0927672955974844e-05,
      "loss": 0.7745,
      "num_input_tokens_seen": 543072,
      "step": 835
    },
    {
      "epoch": 0.44025157232704404,
      "grad_norm": 0.465942919254303,
      "learning_rate": 1.0993186582809224e-05,
      "loss": 0.7065,
      "num_input_tokens_seen": 547456,
      "step": 840
    },
    {
      "epoch": 0.44287211740041926,
      "grad_norm": 0.6566649675369263,
      "learning_rate": 1.1058700209643607e-05,
      "loss": 0.8171,
      "num_input_tokens_seen": 551968,
      "step": 845
    },
    {
      "epoch": 0.44549266247379454,
      "grad_norm": 1.4592140913009644,
      "learning_rate": 1.1124213836477988e-05,
      "loss": 0.86,
      "num_input_tokens_seen": 554400,
      "step": 850
    },
    {
      "epoch": 0.4481132075471698,
      "grad_norm": 0.9253778457641602,
      "learning_rate": 1.118972746331237e-05,
      "loss": 0.8138,
      "num_input_tokens_seen": 557184,
      "step": 855
    },
    {
      "epoch": 0.45073375262054505,
      "grad_norm": 0.7528541684150696,
      "learning_rate": 1.1255241090146751e-05,
      "loss": 0.7111,
      "num_input_tokens_seen": 559840,
      "step": 860
    },
    {
      "epoch": 0.4533542976939203,
      "grad_norm": 1.251615285873413,
      "learning_rate": 1.1320754716981132e-05,
      "loss": 0.7302,
      "num_input_tokens_seen": 562624,
      "step": 865
    },
    {
      "epoch": 0.4559748427672956,
      "grad_norm": 1.1750283241271973,
      "learning_rate": 1.1386268343815514e-05,
      "loss": 0.6836,
      "num_input_tokens_seen": 565888,
      "step": 870
    },
    {
      "epoch": 0.4585953878406709,
      "grad_norm": 1.3334319591522217,
      "learning_rate": 1.1451781970649895e-05,
      "loss": 0.6709,
      "num_input_tokens_seen": 569088,
      "step": 875
    },
    {
      "epoch": 0.4612159329140461,
      "grad_norm": 1.023711085319519,
      "learning_rate": 1.1517295597484278e-05,
      "loss": 0.772,
      "num_input_tokens_seen": 571904,
      "step": 880
    },
    {
      "epoch": 0.4638364779874214,
      "grad_norm": 1.189740777015686,
      "learning_rate": 1.158280922431866e-05,
      "loss": 0.6806,
      "num_input_tokens_seen": 575232,
      "step": 885
    },
    {
      "epoch": 0.46645702306079667,
      "grad_norm": 1.8585017919540405,
      "learning_rate": 1.164832285115304e-05,
      "loss": 0.8676,
      "num_input_tokens_seen": 578432,
      "step": 890
    },
    {
      "epoch": 0.4690775681341719,
      "grad_norm": 1.3898049592971802,
      "learning_rate": 1.1713836477987422e-05,
      "loss": 0.9324,
      "num_input_tokens_seen": 582112,
      "step": 895
    },
    {
      "epoch": 0.4716981132075472,
      "grad_norm": 0.9158449172973633,
      "learning_rate": 1.1779350104821804e-05,
      "loss": 1.0649,
      "num_input_tokens_seen": 584608,
      "step": 900
    },
    {
      "epoch": 0.47431865828092246,
      "grad_norm": 1.1009880304336548,
      "learning_rate": 1.1844863731656185e-05,
      "loss": 0.6302,
      "num_input_tokens_seen": 587808,
      "step": 905
    },
    {
      "epoch": 0.4769392033542977,
      "grad_norm": 0.8030794858932495,
      "learning_rate": 1.1910377358490566e-05,
      "loss": 0.6772,
      "num_input_tokens_seen": 590848,
      "step": 910
    },
    {
      "epoch": 0.47955974842767296,
      "grad_norm": 0.7327920794487,
      "learning_rate": 1.197589098532495e-05,
      "loss": 0.7799,
      "num_input_tokens_seen": 595232,
      "step": 915
    },
    {
      "epoch": 0.48218029350104824,
      "grad_norm": 0.6426371932029724,
      "learning_rate": 1.204140461215933e-05,
      "loss": 0.6641,
      "num_input_tokens_seen": 598560,
      "step": 920
    },
    {
      "epoch": 0.48480083857442346,
      "grad_norm": 0.9936130046844482,
      "learning_rate": 1.210691823899371e-05,
      "loss": 0.6642,
      "num_input_tokens_seen": 606336,
      "step": 925
    },
    {
      "epoch": 0.48742138364779874,
      "grad_norm": 1.0440080165863037,
      "learning_rate": 1.2172431865828094e-05,
      "loss": 0.8439,
      "num_input_tokens_seen": 609728,
      "step": 930
    },
    {
      "epoch": 0.490041928721174,
      "grad_norm": 1.4837956428527832,
      "learning_rate": 1.2237945492662473e-05,
      "loss": 0.7696,
      "num_input_tokens_seen": 612256,
      "step": 935
    },
    {
      "epoch": 0.49266247379454925,
      "grad_norm": 0.8804078698158264,
      "learning_rate": 1.2303459119496856e-05,
      "loss": 0.8108,
      "num_input_tokens_seen": 615520,
      "step": 940
    },
    {
      "epoch": 0.49528301886792453,
      "grad_norm": 0.785793125629425,
      "learning_rate": 1.2368972746331238e-05,
      "loss": 0.7385,
      "num_input_tokens_seen": 618880,
      "step": 945
    },
    {
      "epoch": 0.4979035639412998,
      "grad_norm": 0.5949389934539795,
      "learning_rate": 1.243448637316562e-05,
      "loss": 0.7686,
      "num_input_tokens_seen": 621664,
      "step": 950
    },
    {
      "epoch": 0.500524109014675,
      "grad_norm": 0.8506843447685242,
      "learning_rate": 1.25e-05,
      "loss": 0.8816,
      "num_input_tokens_seen": 624736,
      "step": 955
    },
    {
      "epoch": 0.5031446540880503,
      "grad_norm": 1.5489593744277954,
      "learning_rate": 1.2565513626834382e-05,
      "loss": 0.7759,
      "num_input_tokens_seen": 628576,
      "step": 960
    },
    {
      "epoch": 0.5057651991614256,
      "grad_norm": 0.9176573157310486,
      "learning_rate": 1.2631027253668765e-05,
      "loss": 0.6945,
      "num_input_tokens_seen": 632288,
      "step": 965
    },
    {
      "epoch": 0.5083857442348009,
      "grad_norm": 1.209862470626831,
      "learning_rate": 1.2696540880503146e-05,
      "loss": 0.7362,
      "num_input_tokens_seen": 635008,
      "step": 970
    },
    {
      "epoch": 0.5110062893081762,
      "grad_norm": 0.8188832402229309,
      "learning_rate": 1.2762054507337526e-05,
      "loss": 0.5979,
      "num_input_tokens_seen": 638720,
      "step": 975
    },
    {
      "epoch": 0.5136268343815513,
      "grad_norm": 0.8520936369895935,
      "learning_rate": 1.2827568134171907e-05,
      "loss": 0.7511,
      "num_input_tokens_seen": 641344,
      "step": 980
    },
    {
      "epoch": 0.5162473794549266,
      "grad_norm": 1.0380951166152954,
      "learning_rate": 1.289308176100629e-05,
      "loss": 0.6494,
      "num_input_tokens_seen": 645600,
      "step": 985
    },
    {
      "epoch": 0.5188679245283019,
      "grad_norm": 0.6712779402732849,
      "learning_rate": 1.2958595387840672e-05,
      "loss": 0.9481,
      "num_input_tokens_seen": 648672,
      "step": 990
    },
    {
      "epoch": 0.5214884696016772,
      "grad_norm": 1.8206876516342163,
      "learning_rate": 1.3024109014675053e-05,
      "loss": 0.8136,
      "num_input_tokens_seen": 652416,
      "step": 995
    },
    {
      "epoch": 0.5241090146750524,
      "grad_norm": 0.7451065182685852,
      "learning_rate": 1.3089622641509436e-05,
      "loss": 0.77,
      "num_input_tokens_seen": 655712,
      "step": 1000
    },
    {
      "epoch": 0.5267295597484277,
      "grad_norm": 1.021485447883606,
      "learning_rate": 1.3155136268343814e-05,
      "loss": 0.712,
      "num_input_tokens_seen": 657760,
      "step": 1005
    },
    {
      "epoch": 0.5293501048218029,
      "grad_norm": 0.9769853949546814,
      "learning_rate": 1.3220649895178197e-05,
      "loss": 0.931,
      "num_input_tokens_seen": 660640,
      "step": 1010
    },
    {
      "epoch": 0.5319706498951782,
      "grad_norm": 1.2124098539352417,
      "learning_rate": 1.3286163522012579e-05,
      "loss": 0.7421,
      "num_input_tokens_seen": 666496,
      "step": 1015
    },
    {
      "epoch": 0.5345911949685535,
      "grad_norm": 0.9430646896362305,
      "learning_rate": 1.3351677148846962e-05,
      "loss": 0.7598,
      "num_input_tokens_seen": 669088,
      "step": 1020
    },
    {
      "epoch": 0.5372117400419287,
      "grad_norm": 0.7804273366928101,
      "learning_rate": 1.3417190775681343e-05,
      "loss": 0.7471,
      "num_input_tokens_seen": 672192,
      "step": 1025
    },
    {
      "epoch": 0.539832285115304,
      "grad_norm": 0.7935082912445068,
      "learning_rate": 1.3482704402515725e-05,
      "loss": 0.7233,
      "num_input_tokens_seen": 675104,
      "step": 1030
    },
    {
      "epoch": 0.5424528301886793,
      "grad_norm": 0.5651184320449829,
      "learning_rate": 1.3548218029350104e-05,
      "loss": 0.6257,
      "num_input_tokens_seen": 678368,
      "step": 1035
    },
    {
      "epoch": 0.5450733752620545,
      "grad_norm": 0.7643490433692932,
      "learning_rate": 1.3613731656184486e-05,
      "loss": 0.7335,
      "num_input_tokens_seen": 681600,
      "step": 1040
    },
    {
      "epoch": 0.5476939203354297,
      "grad_norm": 0.8937487006187439,
      "learning_rate": 1.3679245283018869e-05,
      "loss": 0.6799,
      "num_input_tokens_seen": 684576,
      "step": 1045
    },
    {
      "epoch": 0.550314465408805,
      "grad_norm": 1.2376718521118164,
      "learning_rate": 1.374475890985325e-05,
      "loss": 0.6841,
      "num_input_tokens_seen": 688832,
      "step": 1050
    },
    {
      "epoch": 0.5529350104821803,
      "grad_norm": 0.9899259805679321,
      "learning_rate": 1.3810272536687632e-05,
      "loss": 0.7569,
      "num_input_tokens_seen": 691520,
      "step": 1055
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 0.8967555165290833,
      "learning_rate": 1.3875786163522015e-05,
      "loss": 0.6203,
      "num_input_tokens_seen": 694336,
      "step": 1060
    },
    {
      "epoch": 0.5581761006289309,
      "grad_norm": 0.4839145839214325,
      "learning_rate": 1.3941299790356394e-05,
      "loss": 0.5463,
      "num_input_tokens_seen": 698016,
      "step": 1065
    },
    {
      "epoch": 0.560796645702306,
      "grad_norm": 0.6956073641777039,
      "learning_rate": 1.4006813417190776e-05,
      "loss": 0.715,
      "num_input_tokens_seen": 700896,
      "step": 1070
    },
    {
      "epoch": 0.5634171907756813,
      "grad_norm": 0.7053921222686768,
      "learning_rate": 1.4072327044025157e-05,
      "loss": 0.5298,
      "num_input_tokens_seen": 704416,
      "step": 1075
    },
    {
      "epoch": 0.5660377358490566,
      "grad_norm": 1.1298633813858032,
      "learning_rate": 1.413784067085954e-05,
      "loss": 0.8246,
      "num_input_tokens_seen": 707648,
      "step": 1080
    },
    {
      "epoch": 0.5686582809224319,
      "grad_norm": 1.0721114873886108,
      "learning_rate": 1.4203354297693922e-05,
      "loss": 0.7671,
      "num_input_tokens_seen": 711200,
      "step": 1085
    },
    {
      "epoch": 0.5712788259958071,
      "grad_norm": 0.6540836691856384,
      "learning_rate": 1.4268867924528303e-05,
      "loss": 0.6332,
      "num_input_tokens_seen": 715200,
      "step": 1090
    },
    {
      "epoch": 0.5738993710691824,
      "grad_norm": 0.7539855241775513,
      "learning_rate": 1.4334381551362683e-05,
      "loss": 0.599,
      "num_input_tokens_seen": 718080,
      "step": 1095
    },
    {
      "epoch": 0.5765199161425576,
      "grad_norm": 0.827911913394928,
      "learning_rate": 1.4399895178197064e-05,
      "loss": 0.7098,
      "num_input_tokens_seen": 722112,
      "step": 1100
    },
    {
      "epoch": 0.5791404612159329,
      "grad_norm": 0.6060814261436462,
      "learning_rate": 1.4465408805031447e-05,
      "loss": 0.7354,
      "num_input_tokens_seen": 725696,
      "step": 1105
    },
    {
      "epoch": 0.5817610062893082,
      "grad_norm": 0.6188961863517761,
      "learning_rate": 1.4530922431865828e-05,
      "loss": 0.6888,
      "num_input_tokens_seen": 729824,
      "step": 1110
    },
    {
      "epoch": 0.5843815513626834,
      "grad_norm": 0.6155710220336914,
      "learning_rate": 1.4596436058700211e-05,
      "loss": 0.6283,
      "num_input_tokens_seen": 734528,
      "step": 1115
    },
    {
      "epoch": 0.5870020964360587,
      "grad_norm": 0.46240320801734924,
      "learning_rate": 1.4661949685534593e-05,
      "loss": 0.5493,
      "num_input_tokens_seen": 737568,
      "step": 1120
    },
    {
      "epoch": 0.589622641509434,
      "grad_norm": 1.1781225204467773,
      "learning_rate": 1.4727463312368974e-05,
      "loss": 0.7468,
      "num_input_tokens_seen": 740256,
      "step": 1125
    },
    {
      "epoch": 0.5922431865828093,
      "grad_norm": 0.5400687456130981,
      "learning_rate": 1.4792976939203354e-05,
      "loss": 0.5839,
      "num_input_tokens_seen": 743968,
      "step": 1130
    },
    {
      "epoch": 0.5948637316561844,
      "grad_norm": 1.5020153522491455,
      "learning_rate": 1.4858490566037735e-05,
      "loss": 0.8189,
      "num_input_tokens_seen": 746528,
      "step": 1135
    },
    {
      "epoch": 0.5974842767295597,
      "grad_norm": 0.6580325365066528,
      "learning_rate": 1.4924004192872118e-05,
      "loss": 0.7318,
      "num_input_tokens_seen": 749632,
      "step": 1140
    },
    {
      "epoch": 0.600104821802935,
      "grad_norm": 0.608379602432251,
      "learning_rate": 1.49895178197065e-05,
      "loss": 0.6729,
      "num_input_tokens_seen": 752832,
      "step": 1145
    },
    {
      "epoch": 0.6027253668763103,
      "grad_norm": 0.5147722959518433,
      "learning_rate": 1.5055031446540883e-05,
      "loss": 0.6543,
      "num_input_tokens_seen": 756032,
      "step": 1150
    },
    {
      "epoch": 0.6053459119496856,
      "grad_norm": 0.8933337926864624,
      "learning_rate": 1.5120545073375264e-05,
      "loss": 0.6025,
      "num_input_tokens_seen": 759488,
      "step": 1155
    },
    {
      "epoch": 0.6079664570230608,
      "grad_norm": 1.354962944984436,
      "learning_rate": 1.5186058700209644e-05,
      "loss": 0.5707,
      "num_input_tokens_seen": 762304,
      "step": 1160
    },
    {
      "epoch": 0.610587002096436,
      "grad_norm": 0.42547494173049927,
      "learning_rate": 1.5251572327044025e-05,
      "loss": 0.539,
      "num_input_tokens_seen": 765152,
      "step": 1165
    },
    {
      "epoch": 0.6132075471698113,
      "grad_norm": 0.9127871990203857,
      "learning_rate": 1.5317085953878407e-05,
      "loss": 0.6724,
      "num_input_tokens_seen": 768608,
      "step": 1170
    },
    {
      "epoch": 0.6158280922431866,
      "grad_norm": 0.6236212253570557,
      "learning_rate": 1.538259958071279e-05,
      "loss": 0.6093,
      "num_input_tokens_seen": 771328,
      "step": 1175
    },
    {
      "epoch": 0.6184486373165619,
      "grad_norm": 0.5849370360374451,
      "learning_rate": 1.5448113207547173e-05,
      "loss": 0.7782,
      "num_input_tokens_seen": 774464,
      "step": 1180
    },
    {
      "epoch": 0.6210691823899371,
      "grad_norm": 0.9342488050460815,
      "learning_rate": 1.5513626834381552e-05,
      "loss": 0.6939,
      "num_input_tokens_seen": 777824,
      "step": 1185
    },
    {
      "epoch": 0.6236897274633124,
      "grad_norm": 0.43122732639312744,
      "learning_rate": 1.5579140461215932e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 782464,
      "step": 1190
    },
    {
      "epoch": 0.6263102725366876,
      "grad_norm": 1.8347707986831665,
      "learning_rate": 1.5644654088050315e-05,
      "loss": 0.7063,
      "num_input_tokens_seen": 785632,
      "step": 1195
    },
    {
      "epoch": 0.6289308176100629,
      "grad_norm": 0.5259456038475037,
      "learning_rate": 1.5710167714884695e-05,
      "loss": 0.6088,
      "num_input_tokens_seen": 789344,
      "step": 1200
    },
    {
      "epoch": 0.6315513626834381,
      "grad_norm": 0.7159034013748169,
      "learning_rate": 1.5775681341719078e-05,
      "loss": 0.7101,
      "num_input_tokens_seen": 792928,
      "step": 1205
    },
    {
      "epoch": 0.6341719077568134,
      "grad_norm": 0.9090266823768616,
      "learning_rate": 1.584119496855346e-05,
      "loss": 0.6979,
      "num_input_tokens_seen": 796032,
      "step": 1210
    },
    {
      "epoch": 0.6367924528301887,
      "grad_norm": 0.4865121841430664,
      "learning_rate": 1.5906708595387844e-05,
      "loss": 0.619,
      "num_input_tokens_seen": 798464,
      "step": 1215
    },
    {
      "epoch": 0.639412997903564,
      "grad_norm": 0.7079228162765503,
      "learning_rate": 1.597222222222222e-05,
      "loss": 0.8474,
      "num_input_tokens_seen": 801088,
      "step": 1220
    },
    {
      "epoch": 0.6420335429769392,
      "grad_norm": 1.0318795442581177,
      "learning_rate": 1.6037735849056604e-05,
      "loss": 0.7406,
      "num_input_tokens_seen": 803872,
      "step": 1225
    },
    {
      "epoch": 0.6446540880503144,
      "grad_norm": 0.7425409555435181,
      "learning_rate": 1.6103249475890987e-05,
      "loss": 0.6276,
      "num_input_tokens_seen": 806784,
      "step": 1230
    },
    {
      "epoch": 0.6472746331236897,
      "grad_norm": 0.6568148136138916,
      "learning_rate": 1.6168763102725366e-05,
      "loss": 0.4778,
      "num_input_tokens_seen": 810656,
      "step": 1235
    },
    {
      "epoch": 0.649895178197065,
      "grad_norm": 0.5042169094085693,
      "learning_rate": 1.623427672955975e-05,
      "loss": 0.5687,
      "num_input_tokens_seen": 813568,
      "step": 1240
    },
    {
      "epoch": 0.6525157232704403,
      "grad_norm": 0.6947185397148132,
      "learning_rate": 1.6299790356394132e-05,
      "loss": 0.8601,
      "num_input_tokens_seen": 816800,
      "step": 1245
    },
    {
      "epoch": 0.6551362683438156,
      "grad_norm": 0.6329739093780518,
      "learning_rate": 1.6365303983228512e-05,
      "loss": 0.8206,
      "num_input_tokens_seen": 820512,
      "step": 1250
    },
    {
      "epoch": 0.6577568134171907,
      "grad_norm": 0.7516289949417114,
      "learning_rate": 1.6430817610062892e-05,
      "loss": 0.5864,
      "num_input_tokens_seen": 824096,
      "step": 1255
    },
    {
      "epoch": 0.660377358490566,
      "grad_norm": 0.7000110149383545,
      "learning_rate": 1.6496331236897275e-05,
      "loss": 0.6713,
      "num_input_tokens_seen": 827744,
      "step": 1260
    },
    {
      "epoch": 0.6629979035639413,
      "grad_norm": 0.6225433945655823,
      "learning_rate": 1.6561844863731658e-05,
      "loss": 0.6907,
      "num_input_tokens_seen": 830944,
      "step": 1265
    },
    {
      "epoch": 0.6656184486373166,
      "grad_norm": 0.8078488111495972,
      "learning_rate": 1.6627358490566038e-05,
      "loss": 0.5345,
      "num_input_tokens_seen": 834176,
      "step": 1270
    },
    {
      "epoch": 0.6682389937106918,
      "grad_norm": 0.9787901043891907,
      "learning_rate": 1.669287211740042e-05,
      "loss": 0.8821,
      "num_input_tokens_seen": 837664,
      "step": 1275
    },
    {
      "epoch": 0.6708595387840671,
      "grad_norm": 1.187216877937317,
      "learning_rate": 1.6758385744234804e-05,
      "loss": 0.6507,
      "num_input_tokens_seen": 840864,
      "step": 1280
    },
    {
      "epoch": 0.6734800838574424,
      "grad_norm": 0.46330320835113525,
      "learning_rate": 1.6823899371069183e-05,
      "loss": 0.6787,
      "num_input_tokens_seen": 844576,
      "step": 1285
    },
    {
      "epoch": 0.6761006289308176,
      "grad_norm": 0.4884513318538666,
      "learning_rate": 1.6889412997903563e-05,
      "loss": 0.7121,
      "num_input_tokens_seen": 847328,
      "step": 1290
    },
    {
      "epoch": 0.6787211740041929,
      "grad_norm": 0.608089804649353,
      "learning_rate": 1.6954926624737946e-05,
      "loss": 0.6695,
      "num_input_tokens_seen": 850048,
      "step": 1295
    },
    {
      "epoch": 0.6813417190775681,
      "grad_norm": 0.5912442207336426,
      "learning_rate": 1.702044025157233e-05,
      "loss": 0.7505,
      "num_input_tokens_seen": 853568,
      "step": 1300
    },
    {
      "epoch": 0.6839622641509434,
      "grad_norm": 0.9381195306777954,
      "learning_rate": 1.708595387840671e-05,
      "loss": 0.4922,
      "num_input_tokens_seen": 857024,
      "step": 1305
    },
    {
      "epoch": 0.6865828092243187,
      "grad_norm": 0.5126638412475586,
      "learning_rate": 1.7151467505241092e-05,
      "loss": 0.6553,
      "num_input_tokens_seen": 861568,
      "step": 1310
    },
    {
      "epoch": 0.689203354297694,
      "grad_norm": 1.0476425886154175,
      "learning_rate": 1.7216981132075472e-05,
      "loss": 0.723,
      "num_input_tokens_seen": 864384,
      "step": 1315
    },
    {
      "epoch": 0.6918238993710691,
      "grad_norm": 0.5521358251571655,
      "learning_rate": 1.7282494758909855e-05,
      "loss": 0.5858,
      "num_input_tokens_seen": 867136,
      "step": 1320
    },
    {
      "epoch": 0.6944444444444444,
      "grad_norm": 0.927076518535614,
      "learning_rate": 1.7348008385744234e-05,
      "loss": 0.6873,
      "num_input_tokens_seen": 870080,
      "step": 1325
    },
    {
      "epoch": 0.6970649895178197,
      "grad_norm": 1.9468294382095337,
      "learning_rate": 1.7413522012578618e-05,
      "loss": 0.843,
      "num_input_tokens_seen": 873568,
      "step": 1330
    },
    {
      "epoch": 0.699685534591195,
      "grad_norm": 0.7613580226898193,
      "learning_rate": 1.7479035639413e-05,
      "loss": 0.7174,
      "num_input_tokens_seen": 876192,
      "step": 1335
    },
    {
      "epoch": 0.7023060796645703,
      "grad_norm": 0.8399521708488464,
      "learning_rate": 1.754454926624738e-05,
      "loss": 0.6443,
      "num_input_tokens_seen": 878784,
      "step": 1340
    },
    {
      "epoch": 0.7049266247379455,
      "grad_norm": 0.4448126554489136,
      "learning_rate": 1.761006289308176e-05,
      "loss": 0.7615,
      "num_input_tokens_seen": 882304,
      "step": 1345
    },
    {
      "epoch": 0.7075471698113207,
      "grad_norm": 0.7110517024993896,
      "learning_rate": 1.7675576519916143e-05,
      "loss": 0.5649,
      "num_input_tokens_seen": 885792,
      "step": 1350
    },
    {
      "epoch": 0.710167714884696,
      "grad_norm": 0.96793532371521,
      "learning_rate": 1.7741090146750526e-05,
      "loss": 0.6409,
      "num_input_tokens_seen": 887840,
      "step": 1355
    },
    {
      "epoch": 0.7127882599580713,
      "grad_norm": 0.613471269607544,
      "learning_rate": 1.7806603773584906e-05,
      "loss": 0.6818,
      "num_input_tokens_seen": 890848,
      "step": 1360
    },
    {
      "epoch": 0.7154088050314465,
      "grad_norm": 0.625018835067749,
      "learning_rate": 1.787211740041929e-05,
      "loss": 0.7333,
      "num_input_tokens_seen": 894304,
      "step": 1365
    },
    {
      "epoch": 0.7180293501048218,
      "grad_norm": 0.49297183752059937,
      "learning_rate": 1.7937631027253672e-05,
      "loss": 0.7262,
      "num_input_tokens_seen": 897600,
      "step": 1370
    },
    {
      "epoch": 0.7206498951781971,
      "grad_norm": 2.4186742305755615,
      "learning_rate": 1.8003144654088048e-05,
      "loss": 0.6141,
      "num_input_tokens_seen": 900928,
      "step": 1375
    },
    {
      "epoch": 0.7232704402515723,
      "grad_norm": 0.6339559555053711,
      "learning_rate": 1.806865828092243e-05,
      "loss": 0.578,
      "num_input_tokens_seen": 903872,
      "step": 1380
    },
    {
      "epoch": 0.7258909853249476,
      "grad_norm": 0.6203615665435791,
      "learning_rate": 1.8134171907756814e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 907456,
      "step": 1385
    },
    {
      "epoch": 0.7285115303983228,
      "grad_norm": 0.39690425992012024,
      "learning_rate": 1.8199685534591197e-05,
      "loss": 0.5343,
      "num_input_tokens_seen": 911072,
      "step": 1390
    },
    {
      "epoch": 0.7311320754716981,
      "grad_norm": 1.0714364051818848,
      "learning_rate": 1.8265199161425577e-05,
      "loss": 0.6462,
      "num_input_tokens_seen": 914112,
      "step": 1395
    },
    {
      "epoch": 0.7337526205450734,
      "grad_norm": 3.8465890884399414,
      "learning_rate": 1.833071278825996e-05,
      "loss": 0.6352,
      "num_input_tokens_seen": 917280,
      "step": 1400
    },
    {
      "epoch": 0.7363731656184487,
      "grad_norm": 1.6376657485961914,
      "learning_rate": 1.839622641509434e-05,
      "loss": 0.5914,
      "num_input_tokens_seen": 919936,
      "step": 1405
    },
    {
      "epoch": 0.7389937106918238,
      "grad_norm": 0.9447116851806641,
      "learning_rate": 1.846174004192872e-05,
      "loss": 0.7539,
      "num_input_tokens_seen": 922144,
      "step": 1410
    },
    {
      "epoch": 0.7416142557651991,
      "grad_norm": 0.7039919495582581,
      "learning_rate": 1.8527253668763103e-05,
      "loss": 0.6867,
      "num_input_tokens_seen": 925184,
      "step": 1415
    },
    {
      "epoch": 0.7442348008385744,
      "grad_norm": 0.6784430742263794,
      "learning_rate": 1.8592767295597486e-05,
      "loss": 0.7151,
      "num_input_tokens_seen": 928096,
      "step": 1420
    },
    {
      "epoch": 0.7468553459119497,
      "grad_norm": 0.7289260625839233,
      "learning_rate": 1.865828092243187e-05,
      "loss": 0.6563,
      "num_input_tokens_seen": 930976,
      "step": 1425
    },
    {
      "epoch": 0.749475890985325,
      "grad_norm": 1.1158171892166138,
      "learning_rate": 1.872379454926625e-05,
      "loss": 0.6259,
      "num_input_tokens_seen": 934336,
      "step": 1430
    },
    {
      "epoch": 0.7520964360587002,
      "grad_norm": 0.6251977682113647,
      "learning_rate": 1.878930817610063e-05,
      "loss": 0.655,
      "num_input_tokens_seen": 936544,
      "step": 1435
    },
    {
      "epoch": 0.7547169811320755,
      "grad_norm": 0.353899210691452,
      "learning_rate": 1.885482180293501e-05,
      "loss": 0.5205,
      "num_input_tokens_seen": 940160,
      "step": 1440
    },
    {
      "epoch": 0.7573375262054507,
      "grad_norm": 0.7253614664077759,
      "learning_rate": 1.892033542976939e-05,
      "loss": 0.718,
      "num_input_tokens_seen": 942592,
      "step": 1445
    },
    {
      "epoch": 0.759958071278826,
      "grad_norm": 0.3753323554992676,
      "learning_rate": 1.8985849056603774e-05,
      "loss": 0.5552,
      "num_input_tokens_seen": 945856,
      "step": 1450
    },
    {
      "epoch": 0.7625786163522013,
      "grad_norm": 1.0832737684249878,
      "learning_rate": 1.9051362683438157e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 948288,
      "step": 1455
    },
    {
      "epoch": 0.7651991614255765,
      "grad_norm": 1.0247654914855957,
      "learning_rate": 1.9116876310272537e-05,
      "loss": 0.758,
      "num_input_tokens_seen": 950720,
      "step": 1460
    },
    {
      "epoch": 0.7678197064989518,
      "grad_norm": 0.7053716778755188,
      "learning_rate": 1.918238993710692e-05,
      "loss": 0.5035,
      "num_input_tokens_seen": 953760,
      "step": 1465
    },
    {
      "epoch": 0.7704402515723271,
      "grad_norm": 1.0396366119384766,
      "learning_rate": 1.92479035639413e-05,
      "loss": 0.7388,
      "num_input_tokens_seen": 956864,
      "step": 1470
    },
    {
      "epoch": 0.7730607966457023,
      "grad_norm": 0.42412954568862915,
      "learning_rate": 1.9313417190775683e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 959584,
      "step": 1475
    },
    {
      "epoch": 0.7756813417190775,
      "grad_norm": 0.2897217869758606,
      "learning_rate": 1.9378930817610062e-05,
      "loss": 0.45,
      "num_input_tokens_seen": 962720,
      "step": 1480
    },
    {
      "epoch": 0.7783018867924528,
      "grad_norm": 0.7362722158432007,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 0.6231,
      "num_input_tokens_seen": 965568,
      "step": 1485
    },
    {
      "epoch": 0.7809224318658281,
      "grad_norm": 0.588396430015564,
      "learning_rate": 1.950995807127883e-05,
      "loss": 0.6091,
      "num_input_tokens_seen": 969088,
      "step": 1490
    },
    {
      "epoch": 0.7835429769392034,
      "grad_norm": 0.5124256014823914,
      "learning_rate": 1.9575471698113208e-05,
      "loss": 0.5298,
      "num_input_tokens_seen": 972960,
      "step": 1495
    },
    {
      "epoch": 0.7861635220125787,
      "grad_norm": 0.5049580931663513,
      "learning_rate": 1.9640985324947588e-05,
      "loss": 0.6023,
      "num_input_tokens_seen": 976256,
      "step": 1500
    },
    {
      "epoch": 0.7887840670859538,
      "grad_norm": 0.6012617349624634,
      "learning_rate": 1.970649895178197e-05,
      "loss": 0.5187,
      "num_input_tokens_seen": 979584,
      "step": 1505
    },
    {
      "epoch": 0.7914046121593291,
      "grad_norm": 1.4839437007904053,
      "learning_rate": 1.9772012578616354e-05,
      "loss": 0.739,
      "num_input_tokens_seen": 982208,
      "step": 1510
    },
    {
      "epoch": 0.7940251572327044,
      "grad_norm": 0.852408230304718,
      "learning_rate": 1.9837526205450734e-05,
      "loss": 0.5728,
      "num_input_tokens_seen": 986624,
      "step": 1515
    },
    {
      "epoch": 0.7966457023060797,
      "grad_norm": 0.3602868616580963,
      "learning_rate": 1.9903039832285117e-05,
      "loss": 0.6195,
      "num_input_tokens_seen": 989856,
      "step": 1520
    },
    {
      "epoch": 0.799266247379455,
      "grad_norm": 0.5933451056480408,
      "learning_rate": 1.99685534591195e-05,
      "loss": 0.6713,
      "num_input_tokens_seen": 992384,
      "step": 1525
    },
    {
      "epoch": 0.8018867924528302,
      "grad_norm": 0.41494154930114746,
      "learning_rate": 2.003406708595388e-05,
      "loss": 0.5414,
      "num_input_tokens_seen": 996096,
      "step": 1530
    },
    {
      "epoch": 0.8045073375262054,
      "grad_norm": 1.1104764938354492,
      "learning_rate": 2.009958071278826e-05,
      "loss": 0.6169,
      "num_input_tokens_seen": 998880,
      "step": 1535
    },
    {
      "epoch": 0.8071278825995807,
      "grad_norm": 0.7670848965644836,
      "learning_rate": 2.0165094339622642e-05,
      "loss": 0.667,
      "num_input_tokens_seen": 1002336,
      "step": 1540
    },
    {
      "epoch": 0.809748427672956,
      "grad_norm": 0.5282648205757141,
      "learning_rate": 2.0230607966457025e-05,
      "loss": 0.5412,
      "num_input_tokens_seen": 1005888,
      "step": 1545
    },
    {
      "epoch": 0.8123689727463312,
      "grad_norm": 0.46008962392807007,
      "learning_rate": 2.0296121593291405e-05,
      "loss": 0.4284,
      "num_input_tokens_seen": 1010496,
      "step": 1550
    },
    {
      "epoch": 0.8149895178197065,
      "grad_norm": 0.555928111076355,
      "learning_rate": 2.0361635220125788e-05,
      "loss": 0.6535,
      "num_input_tokens_seen": 1014848,
      "step": 1555
    },
    {
      "epoch": 0.8176100628930818,
      "grad_norm": 0.8677557706832886,
      "learning_rate": 2.0427148846960168e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 1018176,
      "step": 1560
    },
    {
      "epoch": 0.820230607966457,
      "grad_norm": 0.9572506546974182,
      "learning_rate": 2.049266247379455e-05,
      "loss": 0.588,
      "num_input_tokens_seen": 1021056,
      "step": 1565
    },
    {
      "epoch": 0.8228511530398323,
      "grad_norm": 0.8394912481307983,
      "learning_rate": 2.055817610062893e-05,
      "loss": 0.6529,
      "num_input_tokens_seen": 1024096,
      "step": 1570
    },
    {
      "epoch": 0.8254716981132075,
      "grad_norm": 0.7808980345726013,
      "learning_rate": 2.0623689727463314e-05,
      "loss": 0.7262,
      "num_input_tokens_seen": 1027072,
      "step": 1575
    },
    {
      "epoch": 0.8280922431865828,
      "grad_norm": 0.9330270290374756,
      "learning_rate": 2.0689203354297697e-05,
      "loss": 0.5609,
      "num_input_tokens_seen": 1029536,
      "step": 1580
    },
    {
      "epoch": 0.8307127882599581,
      "grad_norm": 0.7347900867462158,
      "learning_rate": 2.0754716981132076e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 1032320,
      "step": 1585
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 0.6196506023406982,
      "learning_rate": 2.0820230607966456e-05,
      "loss": 0.7471,
      "num_input_tokens_seen": 1035168,
      "step": 1590
    },
    {
      "epoch": 0.8359538784067087,
      "grad_norm": 0.5285455584526062,
      "learning_rate": 2.088574423480084e-05,
      "loss": 0.5767,
      "num_input_tokens_seen": 1038720,
      "step": 1595
    },
    {
      "epoch": 0.8385744234800838,
      "grad_norm": 0.9828717708587646,
      "learning_rate": 2.0951257861635222e-05,
      "loss": 0.5914,
      "num_input_tokens_seen": 1041344,
      "step": 1600
    },
    {
      "epoch": 0.8411949685534591,
      "grad_norm": 1.9562004804611206,
      "learning_rate": 2.1016771488469602e-05,
      "loss": 0.9589,
      "num_input_tokens_seen": 1043808,
      "step": 1605
    },
    {
      "epoch": 0.8438155136268344,
      "grad_norm": 0.5742682218551636,
      "learning_rate": 2.1082285115303985e-05,
      "loss": 0.5546,
      "num_input_tokens_seen": 1046752,
      "step": 1610
    },
    {
      "epoch": 0.8464360587002097,
      "grad_norm": 0.5168869495391846,
      "learning_rate": 2.1147798742138368e-05,
      "loss": 0.5959,
      "num_input_tokens_seen": 1049664,
      "step": 1615
    },
    {
      "epoch": 0.8490566037735849,
      "grad_norm": 0.6851626038551331,
      "learning_rate": 2.1213312368972748e-05,
      "loss": 0.5586,
      "num_input_tokens_seen": 1052128,
      "step": 1620
    },
    {
      "epoch": 0.8516771488469602,
      "grad_norm": 1.0467995405197144,
      "learning_rate": 2.1278825995807127e-05,
      "loss": 0.5823,
      "num_input_tokens_seen": 1055008,
      "step": 1625
    },
    {
      "epoch": 0.8542976939203354,
      "grad_norm": 0.4478730857372284,
      "learning_rate": 2.134433962264151e-05,
      "loss": 0.7042,
      "num_input_tokens_seen": 1058368,
      "step": 1630
    },
    {
      "epoch": 0.8569182389937107,
      "grad_norm": 0.7355639934539795,
      "learning_rate": 2.1409853249475894e-05,
      "loss": 0.5425,
      "num_input_tokens_seen": 1061888,
      "step": 1635
    },
    {
      "epoch": 0.859538784067086,
      "grad_norm": 1.117974042892456,
      "learning_rate": 2.1475366876310273e-05,
      "loss": 0.5245,
      "num_input_tokens_seen": 1065344,
      "step": 1640
    },
    {
      "epoch": 0.8621593291404612,
      "grad_norm": 0.551158607006073,
      "learning_rate": 2.1540880503144656e-05,
      "loss": 0.5186,
      "num_input_tokens_seen": 1068288,
      "step": 1645
    },
    {
      "epoch": 0.8647798742138365,
      "grad_norm": 0.33919891715049744,
      "learning_rate": 2.160639412997904e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 1072512,
      "step": 1650
    },
    {
      "epoch": 0.8674004192872118,
      "grad_norm": 0.4013964831829071,
      "learning_rate": 2.1671907756813416e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 1077280,
      "step": 1655
    },
    {
      "epoch": 0.870020964360587,
      "grad_norm": 0.6805403232574463,
      "learning_rate": 2.17374213836478e-05,
      "loss": 0.5949,
      "num_input_tokens_seen": 1081632,
      "step": 1660
    },
    {
      "epoch": 0.8726415094339622,
      "grad_norm": 1.06088387966156,
      "learning_rate": 2.1802935010482182e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 1084768,
      "step": 1665
    },
    {
      "epoch": 0.8752620545073375,
      "grad_norm": 0.5199851393699646,
      "learning_rate": 2.186844863731656e-05,
      "loss": 0.6568,
      "num_input_tokens_seen": 1088864,
      "step": 1670
    },
    {
      "epoch": 0.8778825995807128,
      "grad_norm": 0.48078280687332153,
      "learning_rate": 2.1933962264150945e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 1091840,
      "step": 1675
    },
    {
      "epoch": 0.8805031446540881,
      "grad_norm": 1.0831905603408813,
      "learning_rate": 2.1999475890985328e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 1095136,
      "step": 1680
    },
    {
      "epoch": 0.8831236897274634,
      "grad_norm": 0.5670017004013062,
      "learning_rate": 2.2064989517819707e-05,
      "loss": 0.6643,
      "num_input_tokens_seen": 1098848,
      "step": 1685
    },
    {
      "epoch": 0.8857442348008385,
      "grad_norm": 0.49459734559059143,
      "learning_rate": 2.2130503144654087e-05,
      "loss": 0.5881,
      "num_input_tokens_seen": 1101920,
      "step": 1690
    },
    {
      "epoch": 0.8883647798742138,
      "grad_norm": 0.6471921801567078,
      "learning_rate": 2.219601677148847e-05,
      "loss": 0.5492,
      "num_input_tokens_seen": 1105152,
      "step": 1695
    },
    {
      "epoch": 0.8909853249475891,
      "grad_norm": 0.7235283851623535,
      "learning_rate": 2.2261530398322853e-05,
      "loss": 0.5371,
      "num_input_tokens_seen": 1111328,
      "step": 1700
    },
    {
      "epoch": 0.8936058700209644,
      "grad_norm": 1.3411426544189453,
      "learning_rate": 2.2327044025157233e-05,
      "loss": 0.5846,
      "num_input_tokens_seen": 1114240,
      "step": 1705
    },
    {
      "epoch": 0.8962264150943396,
      "grad_norm": 2.1675894260406494,
      "learning_rate": 2.2392557651991616e-05,
      "loss": 0.5935,
      "num_input_tokens_seen": 1118912,
      "step": 1710
    },
    {
      "epoch": 0.8988469601677149,
      "grad_norm": 0.8841935396194458,
      "learning_rate": 2.2458071278825996e-05,
      "loss": 0.7929,
      "num_input_tokens_seen": 1122368,
      "step": 1715
    },
    {
      "epoch": 0.9014675052410901,
      "grad_norm": 0.6512711644172668,
      "learning_rate": 2.252358490566038e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 1125184,
      "step": 1720
    },
    {
      "epoch": 0.9040880503144654,
      "grad_norm": 2.2760958671569824,
      "learning_rate": 2.258909853249476e-05,
      "loss": 0.6953,
      "num_input_tokens_seen": 1128992,
      "step": 1725
    },
    {
      "epoch": 0.9067085953878407,
      "grad_norm": 0.49158087372779846,
      "learning_rate": 2.265461215932914e-05,
      "loss": 0.7204,
      "num_input_tokens_seen": 1131872,
      "step": 1730
    },
    {
      "epoch": 0.9093291404612159,
      "grad_norm": 0.5695055723190308,
      "learning_rate": 2.2720125786163524e-05,
      "loss": 0.5844,
      "num_input_tokens_seen": 1135904,
      "step": 1735
    },
    {
      "epoch": 0.9119496855345912,
      "grad_norm": 1.2628766298294067,
      "learning_rate": 2.2785639412997904e-05,
      "loss": 0.5706,
      "num_input_tokens_seen": 1139520,
      "step": 1740
    },
    {
      "epoch": 0.9145702306079665,
      "grad_norm": 1.1414315700531006,
      "learning_rate": 2.2851153039832284e-05,
      "loss": 0.5832,
      "num_input_tokens_seen": 1142304,
      "step": 1745
    },
    {
      "epoch": 0.9171907756813418,
      "grad_norm": 0.9739688634872437,
      "learning_rate": 2.2916666666666667e-05,
      "loss": 0.5513,
      "num_input_tokens_seen": 1145280,
      "step": 1750
    },
    {
      "epoch": 0.9198113207547169,
      "grad_norm": 0.9695114493370056,
      "learning_rate": 2.298218029350105e-05,
      "loss": 0.5904,
      "num_input_tokens_seen": 1147904,
      "step": 1755
    },
    {
      "epoch": 0.9224318658280922,
      "grad_norm": 1.050757646560669,
      "learning_rate": 2.304769392033543e-05,
      "loss": 0.6927,
      "num_input_tokens_seen": 1151200,
      "step": 1760
    },
    {
      "epoch": 0.9250524109014675,
      "grad_norm": 0.6474234461784363,
      "learning_rate": 2.3113207547169813e-05,
      "loss": 0.6929,
      "num_input_tokens_seen": 1154528,
      "step": 1765
    },
    {
      "epoch": 0.9276729559748428,
      "grad_norm": 0.3630145192146301,
      "learning_rate": 2.3178721174004196e-05,
      "loss": 0.6194,
      "num_input_tokens_seen": 1157632,
      "step": 1770
    },
    {
      "epoch": 0.9302935010482181,
      "grad_norm": 0.7073007225990295,
      "learning_rate": 2.3244234800838576e-05,
      "loss": 0.4724,
      "num_input_tokens_seen": 1161280,
      "step": 1775
    },
    {
      "epoch": 0.9329140461215933,
      "grad_norm": 0.5317884087562561,
      "learning_rate": 2.3309748427672955e-05,
      "loss": 0.581,
      "num_input_tokens_seen": 1164512,
      "step": 1780
    },
    {
      "epoch": 0.9355345911949685,
      "grad_norm": 0.5132631659507751,
      "learning_rate": 2.3375262054507338e-05,
      "loss": 0.5554,
      "num_input_tokens_seen": 1167808,
      "step": 1785
    },
    {
      "epoch": 0.9381551362683438,
      "grad_norm": 0.5970204472541809,
      "learning_rate": 2.344077568134172e-05,
      "loss": 0.5559,
      "num_input_tokens_seen": 1170432,
      "step": 1790
    },
    {
      "epoch": 0.9407756813417191,
      "grad_norm": 0.4327429533004761,
      "learning_rate": 2.35062893081761e-05,
      "loss": 0.4399,
      "num_input_tokens_seen": 1173088,
      "step": 1795
    },
    {
      "epoch": 0.9433962264150944,
      "grad_norm": 0.5061534643173218,
      "learning_rate": 2.3571802935010484e-05,
      "loss": 0.5142,
      "num_input_tokens_seen": 1176448,
      "step": 1800
    },
    {
      "epoch": 0.9460167714884696,
      "grad_norm": 0.7109535336494446,
      "learning_rate": 2.3637316561844867e-05,
      "loss": 0.6028,
      "num_input_tokens_seen": 1179296,
      "step": 1805
    },
    {
      "epoch": 0.9486373165618449,
      "grad_norm": 0.3227807581424713,
      "learning_rate": 2.3702830188679247e-05,
      "loss": 0.6622,
      "num_input_tokens_seen": 1186976,
      "step": 1810
    },
    {
      "epoch": 0.9512578616352201,
      "grad_norm": 0.7850125432014465,
      "learning_rate": 2.3768343815513627e-05,
      "loss": 0.6366,
      "num_input_tokens_seen": 1189248,
      "step": 1815
    },
    {
      "epoch": 0.9538784067085954,
      "grad_norm": 0.5785574316978455,
      "learning_rate": 2.383385744234801e-05,
      "loss": 0.7086,
      "num_input_tokens_seen": 1191840,
      "step": 1820
    },
    {
      "epoch": 0.9564989517819706,
      "grad_norm": 0.5041647553443909,
      "learning_rate": 2.3899371069182393e-05,
      "loss": 0.6192,
      "num_input_tokens_seen": 1194720,
      "step": 1825
    },
    {
      "epoch": 0.9591194968553459,
      "grad_norm": 0.6628458499908447,
      "learning_rate": 2.3964884696016772e-05,
      "loss": 0.5803,
      "num_input_tokens_seen": 1198592,
      "step": 1830
    },
    {
      "epoch": 0.9617400419287212,
      "grad_norm": 0.5902009010314941,
      "learning_rate": 2.4030398322851155e-05,
      "loss": 0.7007,
      "num_input_tokens_seen": 1201216,
      "step": 1835
    },
    {
      "epoch": 0.9643605870020965,
      "grad_norm": 0.6440915465354919,
      "learning_rate": 2.4095911949685535e-05,
      "loss": 0.6005,
      "num_input_tokens_seen": 1204096,
      "step": 1840
    },
    {
      "epoch": 0.9669811320754716,
      "grad_norm": 0.6704831719398499,
      "learning_rate": 2.4161425576519918e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 1206784,
      "step": 1845
    },
    {
      "epoch": 0.9696016771488469,
      "grad_norm": 1.3507215976715088,
      "learning_rate": 2.4226939203354298e-05,
      "loss": 0.5224,
      "num_input_tokens_seen": 1209632,
      "step": 1850
    },
    {
      "epoch": 0.9722222222222222,
      "grad_norm": 0.4530717432498932,
      "learning_rate": 2.429245283018868e-05,
      "loss": 0.5551,
      "num_input_tokens_seen": 1212512,
      "step": 1855
    },
    {
      "epoch": 0.9748427672955975,
      "grad_norm": 0.40448975563049316,
      "learning_rate": 2.4357966457023064e-05,
      "loss": 0.6149,
      "num_input_tokens_seen": 1216096,
      "step": 1860
    },
    {
      "epoch": 0.9774633123689728,
      "grad_norm": 0.5690193176269531,
      "learning_rate": 2.4423480083857444e-05,
      "loss": 0.5645,
      "num_input_tokens_seen": 1219264,
      "step": 1865
    },
    {
      "epoch": 0.980083857442348,
      "grad_norm": 0.5143945217132568,
      "learning_rate": 2.4488993710691823e-05,
      "loss": 0.4869,
      "num_input_tokens_seen": 1222080,
      "step": 1870
    },
    {
      "epoch": 0.9827044025157232,
      "grad_norm": 0.5912972092628479,
      "learning_rate": 2.4554507337526206e-05,
      "loss": 0.62,
      "num_input_tokens_seen": 1224288,
      "step": 1875
    },
    {
      "epoch": 0.9853249475890985,
      "grad_norm": 0.3827981650829315,
      "learning_rate": 2.4620020964360586e-05,
      "loss": 0.6072,
      "num_input_tokens_seen": 1227168,
      "step": 1880
    },
    {
      "epoch": 0.9879454926624738,
      "grad_norm": 0.6750562191009521,
      "learning_rate": 2.468553459119497e-05,
      "loss": 0.4983,
      "num_input_tokens_seen": 1230656,
      "step": 1885
    },
    {
      "epoch": 0.9905660377358491,
      "grad_norm": 0.5771728157997131,
      "learning_rate": 2.4751048218029352e-05,
      "loss": 0.6346,
      "num_input_tokens_seen": 1233312,
      "step": 1890
    },
    {
      "epoch": 0.9931865828092243,
      "grad_norm": 1.4448221921920776,
      "learning_rate": 2.4816561844863735e-05,
      "loss": 0.559,
      "num_input_tokens_seen": 1239040,
      "step": 1895
    },
    {
      "epoch": 0.9958071278825996,
      "grad_norm": 0.3123749792575836,
      "learning_rate": 2.488207547169811e-05,
      "loss": 0.5876,
      "num_input_tokens_seen": 1242720,
      "step": 1900
    },
    {
      "epoch": 0.9984276729559748,
      "grad_norm": 0.6665271520614624,
      "learning_rate": 2.4947589098532495e-05,
      "loss": 0.6003,
      "num_input_tokens_seen": 1245664,
      "step": 1905
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.5419943332672119,
      "eval_runtime": 9.7433,
      "eval_samples_per_second": 87.035,
      "eval_steps_per_second": 21.759,
      "num_input_tokens_seen": 1246832,
      "step": 1908
    },
    {
      "epoch": 1.00104821802935,
      "grad_norm": 0.9669973254203796,
      "learning_rate": 2.5013102725366878e-05,
      "loss": 1.0124,
      "num_input_tokens_seen": 1248016,
      "step": 1910
    },
    {
      "epoch": 1.0036687631027255,
      "grad_norm": 0.8142218589782715,
      "learning_rate": 2.507861635220126e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 1251600,
      "step": 1915
    },
    {
      "epoch": 1.0062893081761006,
      "grad_norm": 0.4535997211933136,
      "learning_rate": 2.514412997903564e-05,
      "loss": 0.7159,
      "num_input_tokens_seen": 1254640,
      "step": 1920
    },
    {
      "epoch": 1.0089098532494758,
      "grad_norm": 0.46130847930908203,
      "learning_rate": 2.520964360587002e-05,
      "loss": 0.4721,
      "num_input_tokens_seen": 1258576,
      "step": 1925
    },
    {
      "epoch": 1.0115303983228512,
      "grad_norm": 0.700411856174469,
      "learning_rate": 2.5275157232704407e-05,
      "loss": 0.6069,
      "num_input_tokens_seen": 1261488,
      "step": 1930
    },
    {
      "epoch": 1.0141509433962264,
      "grad_norm": 0.5395225286483765,
      "learning_rate": 2.5340670859538783e-05,
      "loss": 0.614,
      "num_input_tokens_seen": 1264688,
      "step": 1935
    },
    {
      "epoch": 1.0167714884696017,
      "grad_norm": 0.5008525848388672,
      "learning_rate": 2.540618448637317e-05,
      "loss": 0.508,
      "num_input_tokens_seen": 1267216,
      "step": 1940
    },
    {
      "epoch": 1.019392033542977,
      "grad_norm": 0.8372783064842224,
      "learning_rate": 2.547169811320755e-05,
      "loss": 0.667,
      "num_input_tokens_seen": 1270000,
      "step": 1945
    },
    {
      "epoch": 1.0220125786163523,
      "grad_norm": 0.43964841961860657,
      "learning_rate": 2.553721174004193e-05,
      "loss": 0.5972,
      "num_input_tokens_seen": 1275920,
      "step": 1950
    },
    {
      "epoch": 1.0246331236897275,
      "grad_norm": 0.38094860315322876,
      "learning_rate": 2.5602725366876312e-05,
      "loss": 0.352,
      "num_input_tokens_seen": 1279280,
      "step": 1955
    },
    {
      "epoch": 1.0272536687631026,
      "grad_norm": 0.5493337512016296,
      "learning_rate": 2.566823899371069e-05,
      "loss": 0.7275,
      "num_input_tokens_seen": 1282608,
      "step": 1960
    },
    {
      "epoch": 1.029874213836478,
      "grad_norm": 0.7752300500869751,
      "learning_rate": 2.5733752620545075e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 1285136,
      "step": 1965
    },
    {
      "epoch": 1.0324947589098532,
      "grad_norm": 0.4463907480239868,
      "learning_rate": 2.5799266247379454e-05,
      "loss": 0.6776,
      "num_input_tokens_seen": 1288336,
      "step": 1970
    },
    {
      "epoch": 1.0351153039832286,
      "grad_norm": 0.5139288306236267,
      "learning_rate": 2.586477987421384e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 1290928,
      "step": 1975
    },
    {
      "epoch": 1.0377358490566038,
      "grad_norm": 1.2172062397003174,
      "learning_rate": 2.593029350104822e-05,
      "loss": 0.539,
      "num_input_tokens_seen": 1294960,
      "step": 1980
    },
    {
      "epoch": 1.040356394129979,
      "grad_norm": 0.5657357573509216,
      "learning_rate": 2.59958071278826e-05,
      "loss": 0.6211,
      "num_input_tokens_seen": 1298960,
      "step": 1985
    },
    {
      "epoch": 1.0429769392033543,
      "grad_norm": 0.8074899315834045,
      "learning_rate": 2.6061320754716983e-05,
      "loss": 0.5895,
      "num_input_tokens_seen": 1302288,
      "step": 1990
    },
    {
      "epoch": 1.0455974842767295,
      "grad_norm": 0.4673888087272644,
      "learning_rate": 2.6126834381551363e-05,
      "loss": 0.508,
      "num_input_tokens_seen": 1305872,
      "step": 1995
    },
    {
      "epoch": 1.0482180293501049,
      "grad_norm": 0.5775157809257507,
      "learning_rate": 2.6192348008385746e-05,
      "loss": 0.4897,
      "num_input_tokens_seen": 1309808,
      "step": 2000
    },
    {
      "epoch": 1.05083857442348,
      "grad_norm": 0.5566932559013367,
      "learning_rate": 2.6257861635220126e-05,
      "loss": 0.554,
      "num_input_tokens_seen": 1312400,
      "step": 2005
    },
    {
      "epoch": 1.0534591194968554,
      "grad_norm": 0.5367104411125183,
      "learning_rate": 2.6323375262054505e-05,
      "loss": 0.585,
      "num_input_tokens_seen": 1314896,
      "step": 2010
    },
    {
      "epoch": 1.0560796645702306,
      "grad_norm": 0.6933735013008118,
      "learning_rate": 2.6388888888888892e-05,
      "loss": 0.5471,
      "num_input_tokens_seen": 1317392,
      "step": 2015
    },
    {
      "epoch": 1.0587002096436058,
      "grad_norm": 0.5866854786872864,
      "learning_rate": 2.645440251572327e-05,
      "loss": 0.7726,
      "num_input_tokens_seen": 1319920,
      "step": 2020
    },
    {
      "epoch": 1.0613207547169812,
      "grad_norm": 0.4545484781265259,
      "learning_rate": 2.6519916142557655e-05,
      "loss": 0.5992,
      "num_input_tokens_seen": 1323408,
      "step": 2025
    },
    {
      "epoch": 1.0639412997903563,
      "grad_norm": 0.8238288760185242,
      "learning_rate": 2.6585429769392034e-05,
      "loss": 0.6219,
      "num_input_tokens_seen": 1326192,
      "step": 2030
    },
    {
      "epoch": 1.0665618448637317,
      "grad_norm": 0.4457896053791046,
      "learning_rate": 2.6650943396226417e-05,
      "loss": 0.5479,
      "num_input_tokens_seen": 1329648,
      "step": 2035
    },
    {
      "epoch": 1.069182389937107,
      "grad_norm": 0.5743188858032227,
      "learning_rate": 2.6716457023060797e-05,
      "loss": 0.675,
      "num_input_tokens_seen": 1332720,
      "step": 2040
    },
    {
      "epoch": 1.0718029350104823,
      "grad_norm": 0.8634965419769287,
      "learning_rate": 2.6781970649895177e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 1338288,
      "step": 2045
    },
    {
      "epoch": 1.0744234800838575,
      "grad_norm": 0.683789849281311,
      "learning_rate": 2.6847484276729563e-05,
      "loss": 0.6349,
      "num_input_tokens_seen": 1341040,
      "step": 2050
    },
    {
      "epoch": 1.0770440251572326,
      "grad_norm": 0.9188680052757263,
      "learning_rate": 2.6912997903563943e-05,
      "loss": 0.6111,
      "num_input_tokens_seen": 1343920,
      "step": 2055
    },
    {
      "epoch": 1.079664570230608,
      "grad_norm": 0.4920220375061035,
      "learning_rate": 2.6978511530398326e-05,
      "loss": 0.4993,
      "num_input_tokens_seen": 1346832,
      "step": 2060
    },
    {
      "epoch": 1.0822851153039832,
      "grad_norm": 0.7459357976913452,
      "learning_rate": 2.7044025157232706e-05,
      "loss": 0.4884,
      "num_input_tokens_seen": 1350160,
      "step": 2065
    },
    {
      "epoch": 1.0849056603773586,
      "grad_norm": 0.38920122385025024,
      "learning_rate": 2.710953878406709e-05,
      "loss": 0.6199,
      "num_input_tokens_seen": 1354128,
      "step": 2070
    },
    {
      "epoch": 1.0875262054507338,
      "grad_norm": 0.4801965057849884,
      "learning_rate": 2.717505241090147e-05,
      "loss": 0.5197,
      "num_input_tokens_seen": 1357840,
      "step": 2075
    },
    {
      "epoch": 1.090146750524109,
      "grad_norm": 0.6005801558494568,
      "learning_rate": 2.7240566037735848e-05,
      "loss": 0.6679,
      "num_input_tokens_seen": 1360560,
      "step": 2080
    },
    {
      "epoch": 1.0927672955974843,
      "grad_norm": 0.3399917781352997,
      "learning_rate": 2.7306079664570235e-05,
      "loss": 0.4728,
      "num_input_tokens_seen": 1364176,
      "step": 2085
    },
    {
      "epoch": 1.0953878406708595,
      "grad_norm": 0.5016943216323853,
      "learning_rate": 2.737159329140461e-05,
      "loss": 0.6223,
      "num_input_tokens_seen": 1367120,
      "step": 2090
    },
    {
      "epoch": 1.0980083857442349,
      "grad_norm": 0.4248996078968048,
      "learning_rate": 2.7437106918238997e-05,
      "loss": 0.7685,
      "num_input_tokens_seen": 1370800,
      "step": 2095
    },
    {
      "epoch": 1.10062893081761,
      "grad_norm": 0.5017946362495422,
      "learning_rate": 2.7502620545073377e-05,
      "loss": 0.5002,
      "num_input_tokens_seen": 1374640,
      "step": 2100
    },
    {
      "epoch": 1.1032494758909852,
      "grad_norm": 0.482124000787735,
      "learning_rate": 2.7568134171907757e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 1378000,
      "step": 2105
    },
    {
      "epoch": 1.1058700209643606,
      "grad_norm": 0.40818867087364197,
      "learning_rate": 2.763364779874214e-05,
      "loss": 0.5472,
      "num_input_tokens_seen": 1381328,
      "step": 2110
    },
    {
      "epoch": 1.1084905660377358,
      "grad_norm": 0.5725263953208923,
      "learning_rate": 2.769916142557652e-05,
      "loss": 0.5335,
      "num_input_tokens_seen": 1383600,
      "step": 2115
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 0.3514525890350342,
      "learning_rate": 2.7764675052410906e-05,
      "loss": 0.5602,
      "num_input_tokens_seen": 1387024,
      "step": 2120
    },
    {
      "epoch": 1.1137316561844863,
      "grad_norm": 0.34395840764045715,
      "learning_rate": 2.7830188679245282e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 1390352,
      "step": 2125
    },
    {
      "epoch": 1.1163522012578617,
      "grad_norm": 0.6196674704551697,
      "learning_rate": 2.789570230607967e-05,
      "loss": 0.5455,
      "num_input_tokens_seen": 1393328,
      "step": 2130
    },
    {
      "epoch": 1.118972746331237,
      "grad_norm": 0.4625190794467926,
      "learning_rate": 2.796121593291405e-05,
      "loss": 0.4907,
      "num_input_tokens_seen": 1396688,
      "step": 2135
    },
    {
      "epoch": 1.121593291404612,
      "grad_norm": 0.6455276012420654,
      "learning_rate": 2.8026729559748428e-05,
      "loss": 0.6433,
      "num_input_tokens_seen": 1400400,
      "step": 2140
    },
    {
      "epoch": 1.1242138364779874,
      "grad_norm": 0.6182273626327515,
      "learning_rate": 2.809224318658281e-05,
      "loss": 0.5063,
      "num_input_tokens_seen": 1404144,
      "step": 2145
    },
    {
      "epoch": 1.1268343815513626,
      "grad_norm": 0.7649286985397339,
      "learning_rate": 2.815775681341719e-05,
      "loss": 0.6432,
      "num_input_tokens_seen": 1406960,
      "step": 2150
    },
    {
      "epoch": 1.129454926624738,
      "grad_norm": 0.5346161723136902,
      "learning_rate": 2.8223270440251577e-05,
      "loss": 0.4947,
      "num_input_tokens_seen": 1409840,
      "step": 2155
    },
    {
      "epoch": 1.1320754716981132,
      "grad_norm": 0.44779592752456665,
      "learning_rate": 2.8288784067085954e-05,
      "loss": 0.7216,
      "num_input_tokens_seen": 1412944,
      "step": 2160
    },
    {
      "epoch": 1.1346960167714886,
      "grad_norm": 0.5656867027282715,
      "learning_rate": 2.8354297693920333e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 1416272,
      "step": 2165
    },
    {
      "epoch": 1.1373165618448637,
      "grad_norm": 0.4887397885322571,
      "learning_rate": 2.841981132075472e-05,
      "loss": 0.7838,
      "num_input_tokens_seen": 1419248,
      "step": 2170
    },
    {
      "epoch": 1.139937106918239,
      "grad_norm": 0.4569319486618042,
      "learning_rate": 2.84853249475891e-05,
      "loss": 0.5723,
      "num_input_tokens_seen": 1422352,
      "step": 2175
    },
    {
      "epoch": 1.1425576519916143,
      "grad_norm": 0.3695921301841736,
      "learning_rate": 2.8550838574423482e-05,
      "loss": 0.4539,
      "num_input_tokens_seen": 1425808,
      "step": 2180
    },
    {
      "epoch": 1.1451781970649895,
      "grad_norm": 0.5759515762329102,
      "learning_rate": 2.8616352201257862e-05,
      "loss": 0.7159,
      "num_input_tokens_seen": 1429392,
      "step": 2185
    },
    {
      "epoch": 1.1477987421383649,
      "grad_norm": 1.0965392589569092,
      "learning_rate": 2.868186582809225e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 1432208,
      "step": 2190
    },
    {
      "epoch": 1.15041928721174,
      "grad_norm": 0.6222060322761536,
      "learning_rate": 2.8747379454926625e-05,
      "loss": 0.4977,
      "num_input_tokens_seen": 1434928,
      "step": 2195
    },
    {
      "epoch": 1.1530398322851152,
      "grad_norm": 0.6482958793640137,
      "learning_rate": 2.8812893081761005e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 1438192,
      "step": 2200
    },
    {
      "epoch": 1.1556603773584906,
      "grad_norm": 0.5301708579063416,
      "learning_rate": 2.887840670859539e-05,
      "loss": 0.5532,
      "num_input_tokens_seen": 1441136,
      "step": 2205
    },
    {
      "epoch": 1.1582809224318658,
      "grad_norm": 0.5119888186454773,
      "learning_rate": 2.894392033542977e-05,
      "loss": 0.6213,
      "num_input_tokens_seen": 1444528,
      "step": 2210
    },
    {
      "epoch": 1.1609014675052411,
      "grad_norm": 0.41877511143684387,
      "learning_rate": 2.9009433962264154e-05,
      "loss": 0.7423,
      "num_input_tokens_seen": 1448240,
      "step": 2215
    },
    {
      "epoch": 1.1635220125786163,
      "grad_norm": 0.8839535713195801,
      "learning_rate": 2.9074947589098533e-05,
      "loss": 0.8548,
      "num_input_tokens_seen": 1451792,
      "step": 2220
    },
    {
      "epoch": 1.1661425576519917,
      "grad_norm": 0.4072844386100769,
      "learning_rate": 2.9140461215932913e-05,
      "loss": 0.5567,
      "num_input_tokens_seen": 1453904,
      "step": 2225
    },
    {
      "epoch": 1.1687631027253669,
      "grad_norm": 0.6948492527008057,
      "learning_rate": 2.9205974842767296e-05,
      "loss": 0.571,
      "num_input_tokens_seen": 1457200,
      "step": 2230
    },
    {
      "epoch": 1.171383647798742,
      "grad_norm": 0.466958224773407,
      "learning_rate": 2.9271488469601676e-05,
      "loss": 0.4757,
      "num_input_tokens_seen": 1461488,
      "step": 2235
    },
    {
      "epoch": 1.1740041928721174,
      "grad_norm": 0.6431574821472168,
      "learning_rate": 2.9337002096436062e-05,
      "loss": 0.481,
      "num_input_tokens_seen": 1464080,
      "step": 2240
    },
    {
      "epoch": 1.1766247379454926,
      "grad_norm": 0.6345336437225342,
      "learning_rate": 2.9402515723270442e-05,
      "loss": 0.5943,
      "num_input_tokens_seen": 1466832,
      "step": 2245
    },
    {
      "epoch": 1.179245283018868,
      "grad_norm": 0.8432369232177734,
      "learning_rate": 2.9468029350104825e-05,
      "loss": 0.467,
      "num_input_tokens_seen": 1469520,
      "step": 2250
    },
    {
      "epoch": 1.1818658280922432,
      "grad_norm": 0.5343722105026245,
      "learning_rate": 2.9533542976939205e-05,
      "loss": 0.6521,
      "num_input_tokens_seen": 1476336,
      "step": 2255
    },
    {
      "epoch": 1.1844863731656186,
      "grad_norm": 0.39733821153640747,
      "learning_rate": 2.9599056603773584e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 1479984,
      "step": 2260
    },
    {
      "epoch": 1.1871069182389937,
      "grad_norm": 0.45649802684783936,
      "learning_rate": 2.9664570230607968e-05,
      "loss": 0.51,
      "num_input_tokens_seen": 1483056,
      "step": 2265
    },
    {
      "epoch": 1.189727463312369,
      "grad_norm": 1.8473836183547974,
      "learning_rate": 2.9730083857442347e-05,
      "loss": 0.6181,
      "num_input_tokens_seen": 1485936,
      "step": 2270
    },
    {
      "epoch": 1.1923480083857443,
      "grad_norm": 0.38081392645835876,
      "learning_rate": 2.9795597484276734e-05,
      "loss": 0.5518,
      "num_input_tokens_seen": 1488688,
      "step": 2275
    },
    {
      "epoch": 1.1949685534591195,
      "grad_norm": 0.3757662773132324,
      "learning_rate": 2.9861111111111113e-05,
      "loss": 0.6245,
      "num_input_tokens_seen": 1492272,
      "step": 2280
    },
    {
      "epoch": 1.1975890985324948,
      "grad_norm": 0.4594447910785675,
      "learning_rate": 2.9926624737945496e-05,
      "loss": 0.3964,
      "num_input_tokens_seen": 1495504,
      "step": 2285
    },
    {
      "epoch": 1.20020964360587,
      "grad_norm": 0.4345969259738922,
      "learning_rate": 2.9992138364779876e-05,
      "loss": 0.6083,
      "num_input_tokens_seen": 1498736,
      "step": 2290
    },
    {
      "epoch": 1.2028301886792452,
      "grad_norm": 0.37107089161872864,
      "learning_rate": 3.0057651991614256e-05,
      "loss": 0.5215,
      "num_input_tokens_seen": 1501616,
      "step": 2295
    },
    {
      "epoch": 1.2054507337526206,
      "grad_norm": 0.4245457649230957,
      "learning_rate": 3.012316561844864e-05,
      "loss": 0.6646,
      "num_input_tokens_seen": 1504560,
      "step": 2300
    },
    {
      "epoch": 1.2080712788259957,
      "grad_norm": 0.8836343288421631,
      "learning_rate": 3.018867924528302e-05,
      "loss": 0.5638,
      "num_input_tokens_seen": 1508080,
      "step": 2305
    },
    {
      "epoch": 1.2106918238993711,
      "grad_norm": 0.32655468583106995,
      "learning_rate": 3.0254192872117405e-05,
      "loss": 0.4703,
      "num_input_tokens_seen": 1511824,
      "step": 2310
    },
    {
      "epoch": 1.2133123689727463,
      "grad_norm": 0.5428155660629272,
      "learning_rate": 3.0319706498951785e-05,
      "loss": 0.4144,
      "num_input_tokens_seen": 1515088,
      "step": 2315
    },
    {
      "epoch": 1.2159329140461215,
      "grad_norm": 0.9561067223548889,
      "learning_rate": 3.038522012578616e-05,
      "loss": 0.5199,
      "num_input_tokens_seen": 1517744,
      "step": 2320
    },
    {
      "epoch": 1.2185534591194969,
      "grad_norm": 0.7228496074676514,
      "learning_rate": 3.0450733752620547e-05,
      "loss": 0.5882,
      "num_input_tokens_seen": 1521808,
      "step": 2325
    },
    {
      "epoch": 1.221174004192872,
      "grad_norm": 0.28171175718307495,
      "learning_rate": 3.0516247379454927e-05,
      "loss": 0.5142,
      "num_input_tokens_seen": 1525296,
      "step": 2330
    },
    {
      "epoch": 1.2237945492662474,
      "grad_norm": 0.6692594289779663,
      "learning_rate": 3.0581761006289314e-05,
      "loss": 0.6254,
      "num_input_tokens_seen": 1527856,
      "step": 2335
    },
    {
      "epoch": 1.2264150943396226,
      "grad_norm": 1.4081171751022339,
      "learning_rate": 3.064727463312369e-05,
      "loss": 0.5158,
      "num_input_tokens_seen": 1530384,
      "step": 2340
    },
    {
      "epoch": 1.229035639412998,
      "grad_norm": 0.4367225766181946,
      "learning_rate": 3.071278825995807e-05,
      "loss": 0.3488,
      "num_input_tokens_seen": 1534224,
      "step": 2345
    },
    {
      "epoch": 1.2316561844863732,
      "grad_norm": 0.5359001159667969,
      "learning_rate": 3.0778301886792456e-05,
      "loss": 0.5141,
      "num_input_tokens_seen": 1536656,
      "step": 2350
    },
    {
      "epoch": 1.2342767295597485,
      "grad_norm": 0.5708819627761841,
      "learning_rate": 3.084381551362683e-05,
      "loss": 0.5644,
      "num_input_tokens_seen": 1541168,
      "step": 2355
    },
    {
      "epoch": 1.2368972746331237,
      "grad_norm": 0.8100712895393372,
      "learning_rate": 3.0909329140461215e-05,
      "loss": 0.5656,
      "num_input_tokens_seen": 1543632,
      "step": 2360
    },
    {
      "epoch": 1.2395178197064989,
      "grad_norm": 0.7040530443191528,
      "learning_rate": 3.09748427672956e-05,
      "loss": 0.5087,
      "num_input_tokens_seen": 1548816,
      "step": 2365
    },
    {
      "epoch": 1.2421383647798743,
      "grad_norm": 0.7124499082565308,
      "learning_rate": 3.104035639412998e-05,
      "loss": 0.5204,
      "num_input_tokens_seen": 1552368,
      "step": 2370
    },
    {
      "epoch": 1.2447589098532494,
      "grad_norm": 0.9657700061798096,
      "learning_rate": 3.1105870020964365e-05,
      "loss": 0.6363,
      "num_input_tokens_seen": 1555504,
      "step": 2375
    },
    {
      "epoch": 1.2473794549266248,
      "grad_norm": 0.7678791880607605,
      "learning_rate": 3.117138364779874e-05,
      "loss": 0.5679,
      "num_input_tokens_seen": 1558096,
      "step": 2380
    },
    {
      "epoch": 1.25,
      "grad_norm": 2.1716511249542236,
      "learning_rate": 3.1236897274633124e-05,
      "loss": 0.6488,
      "num_input_tokens_seen": 1561008,
      "step": 2385
    },
    {
      "epoch": 1.2526205450733752,
      "grad_norm": 0.45344114303588867,
      "learning_rate": 3.130241090146751e-05,
      "loss": 0.629,
      "num_input_tokens_seen": 1564176,
      "step": 2390
    },
    {
      "epoch": 1.2552410901467506,
      "grad_norm": 0.4007768929004669,
      "learning_rate": 3.136792452830189e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 1566992,
      "step": 2395
    },
    {
      "epoch": 1.2578616352201257,
      "grad_norm": 0.4348301887512207,
      "learning_rate": 3.1433438155136266e-05,
      "loss": 0.5108,
      "num_input_tokens_seen": 1570352,
      "step": 2400
    },
    {
      "epoch": 1.2604821802935011,
      "grad_norm": 0.4226798117160797,
      "learning_rate": 3.1498951781970656e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 1573040,
      "step": 2405
    },
    {
      "epoch": 1.2631027253668763,
      "grad_norm": 0.8444429039955139,
      "learning_rate": 3.156446540880503e-05,
      "loss": 0.5161,
      "num_input_tokens_seen": 1576944,
      "step": 2410
    },
    {
      "epoch": 1.2657232704402515,
      "grad_norm": 0.5381986498832703,
      "learning_rate": 3.162997903563941e-05,
      "loss": 0.504,
      "num_input_tokens_seen": 1580080,
      "step": 2415
    },
    {
      "epoch": 1.2683438155136268,
      "grad_norm": 0.6422607898712158,
      "learning_rate": 3.16954926624738e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 1582512,
      "step": 2420
    },
    {
      "epoch": 1.270964360587002,
      "grad_norm": 0.6969612240791321,
      "learning_rate": 3.1761006289308175e-05,
      "loss": 0.5542,
      "num_input_tokens_seen": 1586512,
      "step": 2425
    },
    {
      "epoch": 1.2735849056603774,
      "grad_norm": 0.4626805782318115,
      "learning_rate": 3.182651991614256e-05,
      "loss": 0.6062,
      "num_input_tokens_seen": 1589296,
      "step": 2430
    },
    {
      "epoch": 1.2762054507337526,
      "grad_norm": 2.8754355907440186,
      "learning_rate": 3.189203354297694e-05,
      "loss": 0.5976,
      "num_input_tokens_seen": 1592656,
      "step": 2435
    },
    {
      "epoch": 1.2788259958071277,
      "grad_norm": 0.4509221315383911,
      "learning_rate": 3.1957547169811324e-05,
      "loss": 0.6428,
      "num_input_tokens_seen": 1596368,
      "step": 2440
    },
    {
      "epoch": 1.2814465408805031,
      "grad_norm": 0.5743530988693237,
      "learning_rate": 3.202306079664571e-05,
      "loss": 0.5216,
      "num_input_tokens_seen": 1598832,
      "step": 2445
    },
    {
      "epoch": 1.2840670859538785,
      "grad_norm": 0.4863353967666626,
      "learning_rate": 3.2088574423480084e-05,
      "loss": 0.4792,
      "num_input_tokens_seen": 1602992,
      "step": 2450
    },
    {
      "epoch": 1.2866876310272537,
      "grad_norm": 0.9305756688117981,
      "learning_rate": 3.215408805031447e-05,
      "loss": 0.5942,
      "num_input_tokens_seen": 1607344,
      "step": 2455
    },
    {
      "epoch": 1.2893081761006289,
      "grad_norm": 0.6911294460296631,
      "learning_rate": 3.221960167714885e-05,
      "loss": 0.6624,
      "num_input_tokens_seen": 1610992,
      "step": 2460
    },
    {
      "epoch": 1.2919287211740043,
      "grad_norm": 0.5909953713417053,
      "learning_rate": 3.228511530398323e-05,
      "loss": 0.5553,
      "num_input_tokens_seen": 1614000,
      "step": 2465
    },
    {
      "epoch": 1.2945492662473794,
      "grad_norm": 0.48902344703674316,
      "learning_rate": 3.235062893081761e-05,
      "loss": 0.5765,
      "num_input_tokens_seen": 1616752,
      "step": 2470
    },
    {
      "epoch": 1.2971698113207548,
      "grad_norm": 0.40715569257736206,
      "learning_rate": 3.241614255765199e-05,
      "loss": 0.8527,
      "num_input_tokens_seen": 1620144,
      "step": 2475
    },
    {
      "epoch": 1.29979035639413,
      "grad_norm": 0.4562304615974426,
      "learning_rate": 3.2481656184486375e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 1624880,
      "step": 2480
    },
    {
      "epoch": 1.3024109014675052,
      "grad_norm": 0.3565390408039093,
      "learning_rate": 3.254716981132075e-05,
      "loss": 0.5995,
      "num_input_tokens_seen": 1627984,
      "step": 2485
    },
    {
      "epoch": 1.3050314465408805,
      "grad_norm": 0.36138787865638733,
      "learning_rate": 3.261268343815514e-05,
      "loss": 0.7129,
      "num_input_tokens_seen": 1631408,
      "step": 2490
    },
    {
      "epoch": 1.3076519916142557,
      "grad_norm": 0.4887508749961853,
      "learning_rate": 3.267819706498952e-05,
      "loss": 0.7051,
      "num_input_tokens_seen": 1635248,
      "step": 2495
    },
    {
      "epoch": 1.310272536687631,
      "grad_norm": 0.4441046416759491,
      "learning_rate": 3.27437106918239e-05,
      "loss": 0.5593,
      "num_input_tokens_seen": 1637904,
      "step": 2500
    },
    {
      "epoch": 1.3128930817610063,
      "grad_norm": 0.32313206791877747,
      "learning_rate": 3.2809224318658284e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 1643216,
      "step": 2505
    },
    {
      "epoch": 1.3155136268343814,
      "grad_norm": 0.36352646350860596,
      "learning_rate": 3.287473794549266e-05,
      "loss": 0.6031,
      "num_input_tokens_seen": 1646320,
      "step": 2510
    },
    {
      "epoch": 1.3181341719077568,
      "grad_norm": 0.5347557067871094,
      "learning_rate": 3.294025157232704e-05,
      "loss": 0.5014,
      "num_input_tokens_seen": 1650192,
      "step": 2515
    },
    {
      "epoch": 1.320754716981132,
      "grad_norm": 0.6605188846588135,
      "learning_rate": 3.3005765199161426e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 1653840,
      "step": 2520
    },
    {
      "epoch": 1.3233752620545074,
      "grad_norm": 0.31519246101379395,
      "learning_rate": 3.307127882599581e-05,
      "loss": 0.523,
      "num_input_tokens_seen": 1656944,
      "step": 2525
    },
    {
      "epoch": 1.3259958071278826,
      "grad_norm": 0.5268510580062866,
      "learning_rate": 3.313679245283019e-05,
      "loss": 0.611,
      "num_input_tokens_seen": 1661424,
      "step": 2530
    },
    {
      "epoch": 1.3286163522012577,
      "grad_norm": 0.3817065954208374,
      "learning_rate": 3.320230607966457e-05,
      "loss": 0.5033,
      "num_input_tokens_seen": 1664976,
      "step": 2535
    },
    {
      "epoch": 1.3312368972746331,
      "grad_norm": 0.824608325958252,
      "learning_rate": 3.326781970649895e-05,
      "loss": 0.5498,
      "num_input_tokens_seen": 1668528,
      "step": 2540
    },
    {
      "epoch": 1.3338574423480085,
      "grad_norm": 0.4202815890312195,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.484,
      "num_input_tokens_seen": 1671888,
      "step": 2545
    },
    {
      "epoch": 1.3364779874213837,
      "grad_norm": 0.4363122582435608,
      "learning_rate": 3.339884696016772e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 1675280,
      "step": 2550
    },
    {
      "epoch": 1.3390985324947589,
      "grad_norm": 0.5919983983039856,
      "learning_rate": 3.3464360587002094e-05,
      "loss": 0.5434,
      "num_input_tokens_seen": 1678256,
      "step": 2555
    },
    {
      "epoch": 1.3417190775681342,
      "grad_norm": 0.8708918690681458,
      "learning_rate": 3.3529874213836484e-05,
      "loss": 0.652,
      "num_input_tokens_seen": 1681360,
      "step": 2560
    },
    {
      "epoch": 1.3443396226415094,
      "grad_norm": 0.4857345223426819,
      "learning_rate": 3.359538784067086e-05,
      "loss": 0.6348,
      "num_input_tokens_seen": 1684752,
      "step": 2565
    },
    {
      "epoch": 1.3469601677148848,
      "grad_norm": 0.423127144575119,
      "learning_rate": 3.3660901467505244e-05,
      "loss": 0.4713,
      "num_input_tokens_seen": 1689072,
      "step": 2570
    },
    {
      "epoch": 1.34958071278826,
      "grad_norm": 0.5766263008117676,
      "learning_rate": 3.3726415094339627e-05,
      "loss": 0.5385,
      "num_input_tokens_seen": 1692048,
      "step": 2575
    },
    {
      "epoch": 1.3522012578616351,
      "grad_norm": 0.827324628829956,
      "learning_rate": 3.3791928721174e-05,
      "loss": 0.5191,
      "num_input_tokens_seen": 1694800,
      "step": 2580
    },
    {
      "epoch": 1.3548218029350105,
      "grad_norm": 0.4341335892677307,
      "learning_rate": 3.3857442348008386e-05,
      "loss": 0.5444,
      "num_input_tokens_seen": 1698768,
      "step": 2585
    },
    {
      "epoch": 1.3574423480083857,
      "grad_norm": 0.9878074526786804,
      "learning_rate": 3.392295597484277e-05,
      "loss": 0.5949,
      "num_input_tokens_seen": 1701616,
      "step": 2590
    },
    {
      "epoch": 1.360062893081761,
      "grad_norm": 0.4043055474758148,
      "learning_rate": 3.398846960167715e-05,
      "loss": 0.5686,
      "num_input_tokens_seen": 1704528,
      "step": 2595
    },
    {
      "epoch": 1.3626834381551363,
      "grad_norm": 0.5483990907669067,
      "learning_rate": 3.4053983228511535e-05,
      "loss": 0.5278,
      "num_input_tokens_seen": 1708112,
      "step": 2600
    },
    {
      "epoch": 1.3653039832285114,
      "grad_norm": 0.7645253539085388,
      "learning_rate": 3.411949685534591e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 1710864,
      "step": 2605
    },
    {
      "epoch": 1.3679245283018868,
      "grad_norm": 0.590736448764801,
      "learning_rate": 3.4185010482180295e-05,
      "loss": 0.3878,
      "num_input_tokens_seen": 1713648,
      "step": 2610
    },
    {
      "epoch": 1.370545073375262,
      "grad_norm": 0.6757999658584595,
      "learning_rate": 3.425052410901468e-05,
      "loss": 0.6299,
      "num_input_tokens_seen": 1716176,
      "step": 2615
    },
    {
      "epoch": 1.3731656184486374,
      "grad_norm": 0.4901481866836548,
      "learning_rate": 3.431603773584906e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 1719344,
      "step": 2620
    },
    {
      "epoch": 1.3757861635220126,
      "grad_norm": 0.681650698184967,
      "learning_rate": 3.438155136268344e-05,
      "loss": 0.5318,
      "num_input_tokens_seen": 1722800,
      "step": 2625
    },
    {
      "epoch": 1.3784067085953877,
      "grad_norm": 0.946635365486145,
      "learning_rate": 3.444706498951782e-05,
      "loss": 0.6815,
      "num_input_tokens_seen": 1725840,
      "step": 2630
    },
    {
      "epoch": 1.381027253668763,
      "grad_norm": 0.5060742497444153,
      "learning_rate": 3.45125786163522e-05,
      "loss": 0.5335,
      "num_input_tokens_seen": 1729424,
      "step": 2635
    },
    {
      "epoch": 1.3836477987421385,
      "grad_norm": 0.47502216696739197,
      "learning_rate": 3.457809224318658e-05,
      "loss": 0.5824,
      "num_input_tokens_seen": 1734128,
      "step": 2640
    },
    {
      "epoch": 1.3862683438155137,
      "grad_norm": 1.9975090026855469,
      "learning_rate": 3.464360587002097e-05,
      "loss": 0.5851,
      "num_input_tokens_seen": 1737136,
      "step": 2645
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 0.7648647427558899,
      "learning_rate": 3.4709119496855346e-05,
      "loss": 0.6331,
      "num_input_tokens_seen": 1739696,
      "step": 2650
    },
    {
      "epoch": 1.3915094339622642,
      "grad_norm": 0.37267398834228516,
      "learning_rate": 3.477463312368973e-05,
      "loss": 0.54,
      "num_input_tokens_seen": 1742608,
      "step": 2655
    },
    {
      "epoch": 1.3941299790356394,
      "grad_norm": 0.36457619071006775,
      "learning_rate": 3.484014675052411e-05,
      "loss": 0.5335,
      "num_input_tokens_seen": 1745584,
      "step": 2660
    },
    {
      "epoch": 1.3967505241090148,
      "grad_norm": 0.38573476672172546,
      "learning_rate": 3.490566037735849e-05,
      "loss": 0.5523,
      "num_input_tokens_seen": 1748976,
      "step": 2665
    },
    {
      "epoch": 1.39937106918239,
      "grad_norm": 0.590144157409668,
      "learning_rate": 3.497117400419288e-05,
      "loss": 0.8451,
      "num_input_tokens_seen": 1751696,
      "step": 2670
    },
    {
      "epoch": 1.4019916142557651,
      "grad_norm": 0.7047613859176636,
      "learning_rate": 3.5036687631027254e-05,
      "loss": 0.6353,
      "num_input_tokens_seen": 1754640,
      "step": 2675
    },
    {
      "epoch": 1.4046121593291405,
      "grad_norm": 1.2212213277816772,
      "learning_rate": 3.510220125786164e-05,
      "loss": 0.6584,
      "num_input_tokens_seen": 1757488,
      "step": 2680
    },
    {
      "epoch": 1.4072327044025157,
      "grad_norm": 0.797192394733429,
      "learning_rate": 3.516771488469602e-05,
      "loss": 0.5907,
      "num_input_tokens_seen": 1760368,
      "step": 2685
    },
    {
      "epoch": 1.409853249475891,
      "grad_norm": 0.277334988117218,
      "learning_rate": 3.5233228511530397e-05,
      "loss": 0.4851,
      "num_input_tokens_seen": 1763184,
      "step": 2690
    },
    {
      "epoch": 1.4124737945492662,
      "grad_norm": 0.9548255205154419,
      "learning_rate": 3.529874213836478e-05,
      "loss": 0.4671,
      "num_input_tokens_seen": 1766352,
      "step": 2695
    },
    {
      "epoch": 1.4150943396226414,
      "grad_norm": 0.41342073678970337,
      "learning_rate": 3.536425576519916e-05,
      "loss": 0.6415,
      "num_input_tokens_seen": 1769936,
      "step": 2700
    },
    {
      "epoch": 1.4177148846960168,
      "grad_norm": 0.5568838119506836,
      "learning_rate": 3.5429769392033546e-05,
      "loss": 0.6735,
      "num_input_tokens_seen": 1773040,
      "step": 2705
    },
    {
      "epoch": 1.420335429769392,
      "grad_norm": 0.426290899515152,
      "learning_rate": 3.549528301886792e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 1776688,
      "step": 2710
    },
    {
      "epoch": 1.4229559748427674,
      "grad_norm": 0.522803783416748,
      "learning_rate": 3.556079664570231e-05,
      "loss": 0.5554,
      "num_input_tokens_seen": 1779728,
      "step": 2715
    },
    {
      "epoch": 1.4255765199161425,
      "grad_norm": 0.49343329668045044,
      "learning_rate": 3.562631027253669e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 1783056,
      "step": 2720
    },
    {
      "epoch": 1.4281970649895177,
      "grad_norm": 0.5351893305778503,
      "learning_rate": 3.569182389937107e-05,
      "loss": 0.6119,
      "num_input_tokens_seen": 1785840,
      "step": 2725
    },
    {
      "epoch": 1.430817610062893,
      "grad_norm": 0.7463110685348511,
      "learning_rate": 3.5757337526205454e-05,
      "loss": 0.6007,
      "num_input_tokens_seen": 1788080,
      "step": 2730
    },
    {
      "epoch": 1.4334381551362683,
      "grad_norm": 0.537499725818634,
      "learning_rate": 3.582285115303983e-05,
      "loss": 0.6922,
      "num_input_tokens_seen": 1791600,
      "step": 2735
    },
    {
      "epoch": 1.4360587002096437,
      "grad_norm": 0.5690001845359802,
      "learning_rate": 3.588836477987422e-05,
      "loss": 0.5936,
      "num_input_tokens_seen": 1794832,
      "step": 2740
    },
    {
      "epoch": 1.4386792452830188,
      "grad_norm": 0.6088069081306458,
      "learning_rate": 3.59538784067086e-05,
      "loss": 0.6276,
      "num_input_tokens_seen": 1798256,
      "step": 2745
    },
    {
      "epoch": 1.441299790356394,
      "grad_norm": 0.4835801422595978,
      "learning_rate": 3.601939203354298e-05,
      "loss": 0.6595,
      "num_input_tokens_seen": 1802224,
      "step": 2750
    },
    {
      "epoch": 1.4439203354297694,
      "grad_norm": 0.42416658997535706,
      "learning_rate": 3.608490566037736e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 1805648,
      "step": 2755
    },
    {
      "epoch": 1.4465408805031448,
      "grad_norm": 0.5441771149635315,
      "learning_rate": 3.615041928721174e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 1808528,
      "step": 2760
    },
    {
      "epoch": 1.44916142557652,
      "grad_norm": 0.41396135091781616,
      "learning_rate": 3.621593291404612e-05,
      "loss": 0.5828,
      "num_input_tokens_seen": 1811888,
      "step": 2765
    },
    {
      "epoch": 1.4517819706498951,
      "grad_norm": 0.6425541639328003,
      "learning_rate": 3.6281446540880505e-05,
      "loss": 0.5609,
      "num_input_tokens_seen": 1814672,
      "step": 2770
    },
    {
      "epoch": 1.4544025157232705,
      "grad_norm": 0.723282516002655,
      "learning_rate": 3.634696016771489e-05,
      "loss": 0.5405,
      "num_input_tokens_seen": 1817552,
      "step": 2775
    },
    {
      "epoch": 1.4570230607966457,
      "grad_norm": 0.41498780250549316,
      "learning_rate": 3.6412473794549265e-05,
      "loss": 0.5378,
      "num_input_tokens_seen": 1820464,
      "step": 2780
    },
    {
      "epoch": 1.459643605870021,
      "grad_norm": 0.7002676725387573,
      "learning_rate": 3.647798742138365e-05,
      "loss": 0.5033,
      "num_input_tokens_seen": 1823152,
      "step": 2785
    },
    {
      "epoch": 1.4622641509433962,
      "grad_norm": 0.5561481714248657,
      "learning_rate": 3.654350104821803e-05,
      "loss": 0.52,
      "num_input_tokens_seen": 1826320,
      "step": 2790
    },
    {
      "epoch": 1.4648846960167714,
      "grad_norm": 0.6150663495063782,
      "learning_rate": 3.6609014675052414e-05,
      "loss": 0.7036,
      "num_input_tokens_seen": 1830640,
      "step": 2795
    },
    {
      "epoch": 1.4675052410901468,
      "grad_norm": 0.32511404156684875,
      "learning_rate": 3.66745283018868e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 1833712,
      "step": 2800
    },
    {
      "epoch": 1.470125786163522,
      "grad_norm": 0.49239736795425415,
      "learning_rate": 3.6740041928721173e-05,
      "loss": 0.622,
      "num_input_tokens_seen": 1837616,
      "step": 2805
    },
    {
      "epoch": 1.4727463312368974,
      "grad_norm": 0.5866422057151794,
      "learning_rate": 3.6805555555555556e-05,
      "loss": 0.7905,
      "num_input_tokens_seen": 1841264,
      "step": 2810
    },
    {
      "epoch": 1.4753668763102725,
      "grad_norm": 0.4465859830379486,
      "learning_rate": 3.687106918238994e-05,
      "loss": 0.4998,
      "num_input_tokens_seen": 1843888,
      "step": 2815
    },
    {
      "epoch": 1.4779874213836477,
      "grad_norm": 0.43349117040634155,
      "learning_rate": 3.6936582809224316e-05,
      "loss": 0.5958,
      "num_input_tokens_seen": 1847440,
      "step": 2820
    },
    {
      "epoch": 1.480607966457023,
      "grad_norm": 1.6231694221496582,
      "learning_rate": 3.7002096436058706e-05,
      "loss": 0.6872,
      "num_input_tokens_seen": 1850576,
      "step": 2825
    },
    {
      "epoch": 1.4832285115303983,
      "grad_norm": 0.5407583713531494,
      "learning_rate": 3.706761006289308e-05,
      "loss": 0.469,
      "num_input_tokens_seen": 1853200,
      "step": 2830
    },
    {
      "epoch": 1.4858490566037736,
      "grad_norm": 19.98455810546875,
      "learning_rate": 3.7133123689727465e-05,
      "loss": 0.7544,
      "num_input_tokens_seen": 1856784,
      "step": 2835
    },
    {
      "epoch": 1.4884696016771488,
      "grad_norm": 0.4914146065711975,
      "learning_rate": 3.719863731656185e-05,
      "loss": 0.6346,
      "num_input_tokens_seen": 1859984,
      "step": 2840
    },
    {
      "epoch": 1.491090146750524,
      "grad_norm": 0.36156773567199707,
      "learning_rate": 3.7264150943396224e-05,
      "loss": 0.5329,
      "num_input_tokens_seen": 1863920,
      "step": 2845
    },
    {
      "epoch": 1.4937106918238994,
      "grad_norm": 0.4188242256641388,
      "learning_rate": 3.732966457023061e-05,
      "loss": 0.6604,
      "num_input_tokens_seen": 1867664,
      "step": 2850
    },
    {
      "epoch": 1.4963312368972748,
      "grad_norm": 0.5262622237205505,
      "learning_rate": 3.739517819706499e-05,
      "loss": 0.5424,
      "num_input_tokens_seen": 1870352,
      "step": 2855
    },
    {
      "epoch": 1.49895178197065,
      "grad_norm": 0.4682943820953369,
      "learning_rate": 3.7460691823899374e-05,
      "loss": 0.5947,
      "num_input_tokens_seen": 1873488,
      "step": 2860
    },
    {
      "epoch": 1.501572327044025,
      "grad_norm": 0.8866236209869385,
      "learning_rate": 3.752620545073376e-05,
      "loss": 0.5085,
      "num_input_tokens_seen": 1876496,
      "step": 2865
    },
    {
      "epoch": 1.5041928721174003,
      "grad_norm": 0.5718808770179749,
      "learning_rate": 3.759171907756814e-05,
      "loss": 0.6639,
      "num_input_tokens_seen": 1880304,
      "step": 2870
    },
    {
      "epoch": 1.5068134171907757,
      "grad_norm": 0.4782927930355072,
      "learning_rate": 3.7657232704402516e-05,
      "loss": 0.4595,
      "num_input_tokens_seen": 1882960,
      "step": 2875
    },
    {
      "epoch": 1.509433962264151,
      "grad_norm": 0.3172502815723419,
      "learning_rate": 3.77227463312369e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 1886288,
      "step": 2880
    },
    {
      "epoch": 1.5120545073375262,
      "grad_norm": 0.6434236168861389,
      "learning_rate": 3.778825995807128e-05,
      "loss": 0.4593,
      "num_input_tokens_seen": 1890416,
      "step": 2885
    },
    {
      "epoch": 1.5146750524109014,
      "grad_norm": 0.4197690784931183,
      "learning_rate": 3.785377358490566e-05,
      "loss": 0.5688,
      "num_input_tokens_seen": 1893648,
      "step": 2890
    },
    {
      "epoch": 1.5172955974842768,
      "grad_norm": 0.29512012004852295,
      "learning_rate": 3.791928721174005e-05,
      "loss": 0.4488,
      "num_input_tokens_seen": 1897296,
      "step": 2895
    },
    {
      "epoch": 1.519916142557652,
      "grad_norm": 0.4427507221698761,
      "learning_rate": 3.7984800838574425e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 1899856,
      "step": 2900
    },
    {
      "epoch": 1.5225366876310273,
      "grad_norm": 0.5204561352729797,
      "learning_rate": 3.805031446540881e-05,
      "loss": 0.4053,
      "num_input_tokens_seen": 1906544,
      "step": 2905
    },
    {
      "epoch": 1.5251572327044025,
      "grad_norm": 0.45569100975990295,
      "learning_rate": 3.811582809224319e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 1909680,
      "step": 2910
    },
    {
      "epoch": 1.5277777777777777,
      "grad_norm": 0.3465849757194519,
      "learning_rate": 3.818134171907757e-05,
      "loss": 0.5096,
      "num_input_tokens_seen": 1913008,
      "step": 2915
    },
    {
      "epoch": 1.530398322851153,
      "grad_norm": 0.4926144778728485,
      "learning_rate": 3.824685534591195e-05,
      "loss": 0.6225,
      "num_input_tokens_seen": 1916080,
      "step": 2920
    },
    {
      "epoch": 1.5330188679245285,
      "grad_norm": 0.7985512018203735,
      "learning_rate": 3.831236897274633e-05,
      "loss": 0.606,
      "num_input_tokens_seen": 1919504,
      "step": 2925
    },
    {
      "epoch": 1.5356394129979036,
      "grad_norm": 0.5850768089294434,
      "learning_rate": 3.8377882599580716e-05,
      "loss": 0.554,
      "num_input_tokens_seen": 1922736,
      "step": 2930
    },
    {
      "epoch": 1.5382599580712788,
      "grad_norm": 0.42282262444496155,
      "learning_rate": 3.844339622641509e-05,
      "loss": 0.3989,
      "num_input_tokens_seen": 1925456,
      "step": 2935
    },
    {
      "epoch": 1.540880503144654,
      "grad_norm": 0.36374926567077637,
      "learning_rate": 3.8508909853249476e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 1929936,
      "step": 2940
    },
    {
      "epoch": 1.5435010482180294,
      "grad_norm": 0.5302799344062805,
      "learning_rate": 3.857442348008386e-05,
      "loss": 0.5018,
      "num_input_tokens_seen": 1933232,
      "step": 2945
    },
    {
      "epoch": 1.5461215932914047,
      "grad_norm": 0.3198789954185486,
      "learning_rate": 3.863993710691824e-05,
      "loss": 0.5894,
      "num_input_tokens_seen": 1938288,
      "step": 2950
    },
    {
      "epoch": 1.54874213836478,
      "grad_norm": 0.5279403328895569,
      "learning_rate": 3.8705450733752625e-05,
      "loss": 0.5278,
      "num_input_tokens_seen": 1941104,
      "step": 2955
    },
    {
      "epoch": 1.551362683438155,
      "grad_norm": 0.4753953814506531,
      "learning_rate": 3.8770964360587e-05,
      "loss": 0.408,
      "num_input_tokens_seen": 1943632,
      "step": 2960
    },
    {
      "epoch": 1.5539832285115303,
      "grad_norm": 0.5522209405899048,
      "learning_rate": 3.883647798742139e-05,
      "loss": 0.6598,
      "num_input_tokens_seen": 1946288,
      "step": 2965
    },
    {
      "epoch": 1.5566037735849056,
      "grad_norm": 0.4942642152309418,
      "learning_rate": 3.890199161425577e-05,
      "loss": 0.5305,
      "num_input_tokens_seen": 1949264,
      "step": 2970
    },
    {
      "epoch": 1.559224318658281,
      "grad_norm": 0.546662449836731,
      "learning_rate": 3.8967505241090144e-05,
      "loss": 0.6853,
      "num_input_tokens_seen": 1952016,
      "step": 2975
    },
    {
      "epoch": 1.5618448637316562,
      "grad_norm": 0.39907363057136536,
      "learning_rate": 3.9033018867924534e-05,
      "loss": 0.6567,
      "num_input_tokens_seen": 1956112,
      "step": 2980
    },
    {
      "epoch": 1.5644654088050314,
      "grad_norm": 0.7312511205673218,
      "learning_rate": 3.909853249475891e-05,
      "loss": 0.6155,
      "num_input_tokens_seen": 1959248,
      "step": 2985
    },
    {
      "epoch": 1.5670859538784065,
      "grad_norm": 0.5816545486450195,
      "learning_rate": 3.916404612159329e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 1961616,
      "step": 2990
    },
    {
      "epoch": 1.569706498951782,
      "grad_norm": 0.4327949583530426,
      "learning_rate": 3.9229559748427676e-05,
      "loss": 0.5396,
      "num_input_tokens_seen": 1964720,
      "step": 2995
    },
    {
      "epoch": 1.5723270440251573,
      "grad_norm": 0.4961060583591461,
      "learning_rate": 3.929507337526205e-05,
      "loss": 0.4718,
      "num_input_tokens_seen": 1967280,
      "step": 3000
    },
    {
      "epoch": 1.5749475890985325,
      "grad_norm": 0.3823797404766083,
      "learning_rate": 3.9360587002096435e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 1970992,
      "step": 3005
    },
    {
      "epoch": 1.5775681341719077,
      "grad_norm": 0.31847888231277466,
      "learning_rate": 3.942610062893082e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 1974320,
      "step": 3010
    },
    {
      "epoch": 1.580188679245283,
      "grad_norm": 0.8810333013534546,
      "learning_rate": 3.94916142557652e-05,
      "loss": 0.4382,
      "num_input_tokens_seen": 1977680,
      "step": 3015
    },
    {
      "epoch": 1.5828092243186582,
      "grad_norm": 0.6474024057388306,
      "learning_rate": 3.9557127882599585e-05,
      "loss": 0.5308,
      "num_input_tokens_seen": 1980688,
      "step": 3020
    },
    {
      "epoch": 1.5854297693920336,
      "grad_norm": 0.7594582438468933,
      "learning_rate": 3.962264150943397e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 1983728,
      "step": 3025
    },
    {
      "epoch": 1.5880503144654088,
      "grad_norm": 0.4525432884693146,
      "learning_rate": 3.9688155136268344e-05,
      "loss": 0.5404,
      "num_input_tokens_seen": 1986992,
      "step": 3030
    },
    {
      "epoch": 1.590670859538784,
      "grad_norm": 0.7856127023696899,
      "learning_rate": 3.975366876310273e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 1990032,
      "step": 3035
    },
    {
      "epoch": 1.5932914046121593,
      "grad_norm": 0.4027352035045624,
      "learning_rate": 3.981918238993711e-05,
      "loss": 0.4923,
      "num_input_tokens_seen": 1992912,
      "step": 3040
    },
    {
      "epoch": 1.5959119496855347,
      "grad_norm": 0.7030415534973145,
      "learning_rate": 3.9884696016771486e-05,
      "loss": 0.4914,
      "num_input_tokens_seen": 1995664,
      "step": 3045
    },
    {
      "epoch": 1.59853249475891,
      "grad_norm": 0.3181837201118469,
      "learning_rate": 3.9950209643605876e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 1999696,
      "step": 3050
    },
    {
      "epoch": 1.601153039832285,
      "grad_norm": 0.6153348684310913,
      "learning_rate": 4.001572327044025e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 2002128,
      "step": 3055
    },
    {
      "epoch": 1.6037735849056602,
      "grad_norm": 0.5664045810699463,
      "learning_rate": 4.0081236897274636e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 2005712,
      "step": 3060
    },
    {
      "epoch": 1.6063941299790356,
      "grad_norm": 0.5213569402694702,
      "learning_rate": 4.014675052410902e-05,
      "loss": 0.487,
      "num_input_tokens_seen": 2008720,
      "step": 3065
    },
    {
      "epoch": 1.609014675052411,
      "grad_norm": 0.48021915555000305,
      "learning_rate": 4.0212264150943395e-05,
      "loss": 0.5827,
      "num_input_tokens_seen": 2012176,
      "step": 3070
    },
    {
      "epoch": 1.6116352201257862,
      "grad_norm": 0.6010363698005676,
      "learning_rate": 4.027777777777778e-05,
      "loss": 0.5302,
      "num_input_tokens_seen": 2014448,
      "step": 3075
    },
    {
      "epoch": 1.6142557651991614,
      "grad_norm": 0.27420973777770996,
      "learning_rate": 4.034329140461216e-05,
      "loss": 0.5209,
      "num_input_tokens_seen": 2017488,
      "step": 3080
    },
    {
      "epoch": 1.6168763102725365,
      "grad_norm": 0.8087742328643799,
      "learning_rate": 4.0408805031446544e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 2020400,
      "step": 3085
    },
    {
      "epoch": 1.619496855345912,
      "grad_norm": 0.578599750995636,
      "learning_rate": 4.047431865828093e-05,
      "loss": 0.4966,
      "num_input_tokens_seen": 2023216,
      "step": 3090
    },
    {
      "epoch": 1.6221174004192873,
      "grad_norm": 0.4981738030910492,
      "learning_rate": 4.0539832285115304e-05,
      "loss": 0.5596,
      "num_input_tokens_seen": 2028368,
      "step": 3095
    },
    {
      "epoch": 1.6247379454926625,
      "grad_norm": 0.8507988452911377,
      "learning_rate": 4.0605345911949687e-05,
      "loss": 0.5751,
      "num_input_tokens_seen": 2030896,
      "step": 3100
    },
    {
      "epoch": 1.6273584905660377,
      "grad_norm": 0.5051009654998779,
      "learning_rate": 4.067085953878407e-05,
      "loss": 0.5387,
      "num_input_tokens_seen": 2034128,
      "step": 3105
    },
    {
      "epoch": 1.629979035639413,
      "grad_norm": 0.4888501465320587,
      "learning_rate": 4.073637316561845e-05,
      "loss": 0.5761,
      "num_input_tokens_seen": 2037584,
      "step": 3110
    },
    {
      "epoch": 1.6325995807127882,
      "grad_norm": 0.5876232981681824,
      "learning_rate": 4.080188679245283e-05,
      "loss": 0.4981,
      "num_input_tokens_seen": 2040368,
      "step": 3115
    },
    {
      "epoch": 1.6352201257861636,
      "grad_norm": 0.5532404184341431,
      "learning_rate": 4.086740041928722e-05,
      "loss": 0.5354,
      "num_input_tokens_seen": 2043440,
      "step": 3120
    },
    {
      "epoch": 1.6378406708595388,
      "grad_norm": 0.7356247305870056,
      "learning_rate": 4.0932914046121595e-05,
      "loss": 0.6388,
      "num_input_tokens_seen": 2046224,
      "step": 3125
    },
    {
      "epoch": 1.640461215932914,
      "grad_norm": 0.3903687596321106,
      "learning_rate": 4.099842767295597e-05,
      "loss": 0.4789,
      "num_input_tokens_seen": 2049264,
      "step": 3130
    },
    {
      "epoch": 1.6430817610062893,
      "grad_norm": 0.49304285645484924,
      "learning_rate": 4.106394129979036e-05,
      "loss": 0.4873,
      "num_input_tokens_seen": 2053232,
      "step": 3135
    },
    {
      "epoch": 1.6457023060796647,
      "grad_norm": 0.4448840916156769,
      "learning_rate": 4.112945492662474e-05,
      "loss": 0.4758,
      "num_input_tokens_seen": 2056592,
      "step": 3140
    },
    {
      "epoch": 1.64832285115304,
      "grad_norm": 0.42423123121261597,
      "learning_rate": 4.119496855345912e-05,
      "loss": 0.5539,
      "num_input_tokens_seen": 2060688,
      "step": 3145
    },
    {
      "epoch": 1.650943396226415,
      "grad_norm": 0.3329792618751526,
      "learning_rate": 4.1260482180293504e-05,
      "loss": 0.4948,
      "num_input_tokens_seen": 2064752,
      "step": 3150
    },
    {
      "epoch": 1.6535639412997902,
      "grad_norm": 0.5124645233154297,
      "learning_rate": 4.132599580712788e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 2067280,
      "step": 3155
    },
    {
      "epoch": 1.6561844863731656,
      "grad_norm": 0.8192888498306274,
      "learning_rate": 4.139150943396227e-05,
      "loss": 0.5719,
      "num_input_tokens_seen": 2070608,
      "step": 3160
    },
    {
      "epoch": 1.658805031446541,
      "grad_norm": 0.4918099045753479,
      "learning_rate": 4.1457023060796646e-05,
      "loss": 0.5132,
      "num_input_tokens_seen": 2074160,
      "step": 3165
    },
    {
      "epoch": 1.6614255765199162,
      "grad_norm": 0.5540370345115662,
      "learning_rate": 4.152253668763103e-05,
      "loss": 0.4602,
      "num_input_tokens_seen": 2077168,
      "step": 3170
    },
    {
      "epoch": 1.6640461215932913,
      "grad_norm": 0.6161043047904968,
      "learning_rate": 4.158805031446541e-05,
      "loss": 0.4992,
      "num_input_tokens_seen": 2080784,
      "step": 3175
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 0.7462447881698608,
      "learning_rate": 4.1653563941299795e-05,
      "loss": 0.5667,
      "num_input_tokens_seen": 2084400,
      "step": 3180
    },
    {
      "epoch": 1.669287211740042,
      "grad_norm": 0.44070225954055786,
      "learning_rate": 4.171907756813417e-05,
      "loss": 0.611,
      "num_input_tokens_seen": 2087632,
      "step": 3185
    },
    {
      "epoch": 1.6719077568134173,
      "grad_norm": 0.5273495316505432,
      "learning_rate": 4.1784591194968555e-05,
      "loss": 0.5803,
      "num_input_tokens_seen": 2091280,
      "step": 3190
    },
    {
      "epoch": 1.6745283018867925,
      "grad_norm": 0.34433281421661377,
      "learning_rate": 4.185010482180294e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 2094640,
      "step": 3195
    },
    {
      "epoch": 1.6771488469601676,
      "grad_norm": 0.5687239766120911,
      "learning_rate": 4.1915618448637314e-05,
      "loss": 0.7111,
      "num_input_tokens_seen": 2098064,
      "step": 3200
    },
    {
      "epoch": 1.679769392033543,
      "grad_norm": 0.49156779050827026,
      "learning_rate": 4.1981132075471704e-05,
      "loss": 0.5339,
      "num_input_tokens_seen": 2102064,
      "step": 3205
    },
    {
      "epoch": 1.6823899371069182,
      "grad_norm": 0.499965637922287,
      "learning_rate": 4.204664570230608e-05,
      "loss": 0.6894,
      "num_input_tokens_seen": 2105296,
      "step": 3210
    },
    {
      "epoch": 1.6850104821802936,
      "grad_norm": 0.4212835133075714,
      "learning_rate": 4.2112159329140463e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 2108592,
      "step": 3215
    },
    {
      "epoch": 1.6876310272536688,
      "grad_norm": 0.45257049798965454,
      "learning_rate": 4.2177672955974846e-05,
      "loss": 0.3554,
      "num_input_tokens_seen": 2111632,
      "step": 3220
    },
    {
      "epoch": 1.690251572327044,
      "grad_norm": 0.41597113013267517,
      "learning_rate": 4.224318658280922e-05,
      "loss": 0.7347,
      "num_input_tokens_seen": 2115792,
      "step": 3225
    },
    {
      "epoch": 1.6928721174004193,
      "grad_norm": 0.4650072157382965,
      "learning_rate": 4.2308700209643606e-05,
      "loss": 0.4275,
      "num_input_tokens_seen": 2118832,
      "step": 3230
    },
    {
      "epoch": 1.6954926624737947,
      "grad_norm": 0.3390311896800995,
      "learning_rate": 4.237421383647799e-05,
      "loss": 0.5126,
      "num_input_tokens_seen": 2122032,
      "step": 3235
    },
    {
      "epoch": 1.6981132075471699,
      "grad_norm": 0.6570661067962646,
      "learning_rate": 4.243972746331237e-05,
      "loss": 0.5226,
      "num_input_tokens_seen": 2124400,
      "step": 3240
    },
    {
      "epoch": 1.700733752620545,
      "grad_norm": 0.5206285715103149,
      "learning_rate": 4.2505241090146755e-05,
      "loss": 0.5801,
      "num_input_tokens_seen": 2126928,
      "step": 3245
    },
    {
      "epoch": 1.7033542976939202,
      "grad_norm": 0.37016725540161133,
      "learning_rate": 4.257075471698113e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 2130128,
      "step": 3250
    },
    {
      "epoch": 1.7059748427672956,
      "grad_norm": 0.3089825212955475,
      "learning_rate": 4.2636268343815514e-05,
      "loss": 0.5707,
      "num_input_tokens_seen": 2133360,
      "step": 3255
    },
    {
      "epoch": 1.708595387840671,
      "grad_norm": 0.4995526969432831,
      "learning_rate": 4.27017819706499e-05,
      "loss": 0.6574,
      "num_input_tokens_seen": 2136464,
      "step": 3260
    },
    {
      "epoch": 1.7112159329140462,
      "grad_norm": 0.35094261169433594,
      "learning_rate": 4.276729559748428e-05,
      "loss": 0.6964,
      "num_input_tokens_seen": 2140080,
      "step": 3265
    },
    {
      "epoch": 1.7138364779874213,
      "grad_norm": 1.382200837135315,
      "learning_rate": 4.283280922431866e-05,
      "loss": 0.6109,
      "num_input_tokens_seen": 2142960,
      "step": 3270
    },
    {
      "epoch": 1.7164570230607965,
      "grad_norm": 0.4647575318813324,
      "learning_rate": 4.289832285115305e-05,
      "loss": 0.5616,
      "num_input_tokens_seen": 2146224,
      "step": 3275
    },
    {
      "epoch": 1.719077568134172,
      "grad_norm": 0.3445565700531006,
      "learning_rate": 4.296383647798742e-05,
      "loss": 0.492,
      "num_input_tokens_seen": 2150992,
      "step": 3280
    },
    {
      "epoch": 1.7216981132075473,
      "grad_norm": 0.4999692440032959,
      "learning_rate": 4.3029350104821806e-05,
      "loss": 0.6595,
      "num_input_tokens_seen": 2154448,
      "step": 3285
    },
    {
      "epoch": 1.7243186582809225,
      "grad_norm": 0.43481698632240295,
      "learning_rate": 4.309486373165619e-05,
      "loss": 0.372,
      "num_input_tokens_seen": 2156752,
      "step": 3290
    },
    {
      "epoch": 1.7269392033542976,
      "grad_norm": 0.6187880039215088,
      "learning_rate": 4.3160377358490565e-05,
      "loss": 0.6095,
      "num_input_tokens_seen": 2159568,
      "step": 3295
    },
    {
      "epoch": 1.7295597484276728,
      "grad_norm": 0.4759276509284973,
      "learning_rate": 4.322589098532495e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 2162736,
      "step": 3300
    },
    {
      "epoch": 1.7321802935010482,
      "grad_norm": 0.2933844029903412,
      "learning_rate": 4.329140461215933e-05,
      "loss": 0.6193,
      "num_input_tokens_seen": 2166288,
      "step": 3305
    },
    {
      "epoch": 1.7348008385744236,
      "grad_norm": 0.6290852427482605,
      "learning_rate": 4.335691823899371e-05,
      "loss": 0.5575,
      "num_input_tokens_seen": 2169584,
      "step": 3310
    },
    {
      "epoch": 1.7374213836477987,
      "grad_norm": 0.3897894620895386,
      "learning_rate": 4.34224318658281e-05,
      "loss": 0.539,
      "num_input_tokens_seen": 2172560,
      "step": 3315
    },
    {
      "epoch": 1.740041928721174,
      "grad_norm": 1.5937798023223877,
      "learning_rate": 4.3487945492662474e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 2175600,
      "step": 3320
    },
    {
      "epoch": 1.7426624737945493,
      "grad_norm": 0.34403344988822937,
      "learning_rate": 4.355345911949686e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 2178800,
      "step": 3325
    },
    {
      "epoch": 1.7452830188679245,
      "grad_norm": 0.3311350345611572,
      "learning_rate": 4.361897274633124e-05,
      "loss": 0.504,
      "num_input_tokens_seen": 2181744,
      "step": 3330
    },
    {
      "epoch": 1.7479035639412999,
      "grad_norm": 0.4531150460243225,
      "learning_rate": 4.368448637316562e-05,
      "loss": 0.5018,
      "num_input_tokens_seen": 2184720,
      "step": 3335
    },
    {
      "epoch": 1.750524109014675,
      "grad_norm": 1.00504469871521,
      "learning_rate": 4.375e-05,
      "loss": 0.5294,
      "num_input_tokens_seen": 2186992,
      "step": 3340
    },
    {
      "epoch": 1.7531446540880502,
      "grad_norm": 0.3402861952781677,
      "learning_rate": 4.381551362683438e-05,
      "loss": 0.5044,
      "num_input_tokens_seen": 2190192,
      "step": 3345
    },
    {
      "epoch": 1.7557651991614256,
      "grad_norm": 0.4238360822200775,
      "learning_rate": 4.3881027253668766e-05,
      "loss": 0.4499,
      "num_input_tokens_seen": 2193488,
      "step": 3350
    },
    {
      "epoch": 1.758385744234801,
      "grad_norm": 0.6114832758903503,
      "learning_rate": 4.394654088050314e-05,
      "loss": 0.487,
      "num_input_tokens_seen": 2196528,
      "step": 3355
    },
    {
      "epoch": 1.7610062893081762,
      "grad_norm": 0.39197707176208496,
      "learning_rate": 4.401205450733753e-05,
      "loss": 0.3858,
      "num_input_tokens_seen": 2202224,
      "step": 3360
    },
    {
      "epoch": 1.7636268343815513,
      "grad_norm": 0.20794476568698883,
      "learning_rate": 4.407756813417191e-05,
      "loss": 0.4521,
      "num_input_tokens_seen": 2206160,
      "step": 3365
    },
    {
      "epoch": 1.7662473794549265,
      "grad_norm": 0.4788846969604492,
      "learning_rate": 4.414308176100629e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 2209296,
      "step": 3370
    },
    {
      "epoch": 1.7688679245283019,
      "grad_norm": 0.7072663307189941,
      "learning_rate": 4.4208595387840674e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 2212368,
      "step": 3375
    },
    {
      "epoch": 1.7714884696016773,
      "grad_norm": 0.5586097240447998,
      "learning_rate": 4.427410901467505e-05,
      "loss": 0.5425,
      "num_input_tokens_seen": 2215760,
      "step": 3380
    },
    {
      "epoch": 1.7741090146750524,
      "grad_norm": 0.4002189040184021,
      "learning_rate": 4.433962264150944e-05,
      "loss": 0.6104,
      "num_input_tokens_seen": 2219312,
      "step": 3385
    },
    {
      "epoch": 1.7767295597484276,
      "grad_norm": 0.5494151711463928,
      "learning_rate": 4.440513626834382e-05,
      "loss": 0.3252,
      "num_input_tokens_seen": 2221712,
      "step": 3390
    },
    {
      "epoch": 1.7793501048218028,
      "grad_norm": 0.49495822191238403,
      "learning_rate": 4.44706498951782e-05,
      "loss": 0.5962,
      "num_input_tokens_seen": 2224816,
      "step": 3395
    },
    {
      "epoch": 1.7819706498951782,
      "grad_norm": 0.41036465764045715,
      "learning_rate": 4.453616352201258e-05,
      "loss": 0.4246,
      "num_input_tokens_seen": 2228464,
      "step": 3400
    },
    {
      "epoch": 1.7845911949685536,
      "grad_norm": 0.6541240215301514,
      "learning_rate": 4.460167714884696e-05,
      "loss": 0.5504,
      "num_input_tokens_seen": 2232592,
      "step": 3405
    },
    {
      "epoch": 1.7872117400419287,
      "grad_norm": 0.7459917068481445,
      "learning_rate": 4.466719077568134e-05,
      "loss": 0.4914,
      "num_input_tokens_seen": 2235728,
      "step": 3410
    },
    {
      "epoch": 1.789832285115304,
      "grad_norm": 0.389504611492157,
      "learning_rate": 4.4732704402515725e-05,
      "loss": 0.3619,
      "num_input_tokens_seen": 2239216,
      "step": 3415
    },
    {
      "epoch": 1.7924528301886793,
      "grad_norm": 0.5933860540390015,
      "learning_rate": 4.479821802935011e-05,
      "loss": 0.5028,
      "num_input_tokens_seen": 2242256,
      "step": 3420
    },
    {
      "epoch": 1.7950733752620545,
      "grad_norm": 0.5577006340026855,
      "learning_rate": 4.4863731656184485e-05,
      "loss": 0.5079,
      "num_input_tokens_seen": 2245744,
      "step": 3425
    },
    {
      "epoch": 1.7976939203354299,
      "grad_norm": 0.5507862567901611,
      "learning_rate": 4.4929245283018875e-05,
      "loss": 0.548,
      "num_input_tokens_seen": 2249808,
      "step": 3430
    },
    {
      "epoch": 1.800314465408805,
      "grad_norm": 0.5086677074432373,
      "learning_rate": 4.499475890985325e-05,
      "loss": 0.5033,
      "num_input_tokens_seen": 2252976,
      "step": 3435
    },
    {
      "epoch": 1.8029350104821802,
      "grad_norm": 0.6095401644706726,
      "learning_rate": 4.5060272536687634e-05,
      "loss": 0.5346,
      "num_input_tokens_seen": 2256368,
      "step": 3440
    },
    {
      "epoch": 1.8055555555555556,
      "grad_norm": 0.8001255989074707,
      "learning_rate": 4.512578616352202e-05,
      "loss": 0.5231,
      "num_input_tokens_seen": 2259536,
      "step": 3445
    },
    {
      "epoch": 1.808176100628931,
      "grad_norm": 0.5274432897567749,
      "learning_rate": 4.519129979035639e-05,
      "loss": 0.5092,
      "num_input_tokens_seen": 2262928,
      "step": 3450
    },
    {
      "epoch": 1.8107966457023061,
      "grad_norm": 1.7363172769546509,
      "learning_rate": 4.525681341719078e-05,
      "loss": 0.8839,
      "num_input_tokens_seen": 2265520,
      "step": 3455
    },
    {
      "epoch": 1.8134171907756813,
      "grad_norm": 0.5483840107917786,
      "learning_rate": 4.532232704402516e-05,
      "loss": 0.7213,
      "num_input_tokens_seen": 2268464,
      "step": 3460
    },
    {
      "epoch": 1.8160377358490565,
      "grad_norm": 0.4060637950897217,
      "learning_rate": 4.5387840670859536e-05,
      "loss": 0.5482,
      "num_input_tokens_seen": 2271504,
      "step": 3465
    },
    {
      "epoch": 1.8186582809224319,
      "grad_norm": 0.6333862543106079,
      "learning_rate": 4.5453354297693926e-05,
      "loss": 0.4939,
      "num_input_tokens_seen": 2274384,
      "step": 3470
    },
    {
      "epoch": 1.8212788259958073,
      "grad_norm": 0.2920350730419159,
      "learning_rate": 4.55188679245283e-05,
      "loss": 0.4011,
      "num_input_tokens_seen": 2277648,
      "step": 3475
    },
    {
      "epoch": 1.8238993710691824,
      "grad_norm": 0.8254259824752808,
      "learning_rate": 4.5584381551362685e-05,
      "loss": 0.6413,
      "num_input_tokens_seen": 2280528,
      "step": 3480
    },
    {
      "epoch": 1.8265199161425576,
      "grad_norm": 0.25148746371269226,
      "learning_rate": 4.564989517819707e-05,
      "loss": 0.5403,
      "num_input_tokens_seen": 2283792,
      "step": 3485
    },
    {
      "epoch": 1.8291404612159328,
      "grad_norm": 0.47780412435531616,
      "learning_rate": 4.571540880503145e-05,
      "loss": 0.646,
      "num_input_tokens_seen": 2286832,
      "step": 3490
    },
    {
      "epoch": 1.8317610062893082,
      "grad_norm": 0.437588095664978,
      "learning_rate": 4.578092243186583e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 2291312,
      "step": 3495
    },
    {
      "epoch": 1.8343815513626835,
      "grad_norm": 0.5692160725593567,
      "learning_rate": 4.584643605870021e-05,
      "loss": 0.4819,
      "num_input_tokens_seen": 2294384,
      "step": 3500
    },
    {
      "epoch": 1.8370020964360587,
      "grad_norm": 0.4526466131210327,
      "learning_rate": 4.5911949685534594e-05,
      "loss": 0.6166,
      "num_input_tokens_seen": 2297232,
      "step": 3505
    },
    {
      "epoch": 1.8396226415094339,
      "grad_norm": 0.3562130033969879,
      "learning_rate": 4.5977463312368977e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 2300560,
      "step": 3510
    },
    {
      "epoch": 1.8422431865828093,
      "grad_norm": 0.6137692332267761,
      "learning_rate": 4.604297693920336e-05,
      "loss": 0.3532,
      "num_input_tokens_seen": 2303312,
      "step": 3515
    },
    {
      "epoch": 1.8448637316561844,
      "grad_norm": 0.45196324586868286,
      "learning_rate": 4.6108490566037736e-05,
      "loss": 0.5671,
      "num_input_tokens_seen": 2306096,
      "step": 3520
    },
    {
      "epoch": 1.8474842767295598,
      "grad_norm": 0.4311434328556061,
      "learning_rate": 4.617400419287212e-05,
      "loss": 0.5564,
      "num_input_tokens_seen": 2310096,
      "step": 3525
    },
    {
      "epoch": 1.850104821802935,
      "grad_norm": 0.4174090325832367,
      "learning_rate": 4.62395178197065e-05,
      "loss": 0.4992,
      "num_input_tokens_seen": 2313072,
      "step": 3530
    },
    {
      "epoch": 1.8527253668763102,
      "grad_norm": 0.3088938891887665,
      "learning_rate": 4.630503144654088e-05,
      "loss": 0.5029,
      "num_input_tokens_seen": 2316304,
      "step": 3535
    },
    {
      "epoch": 1.8553459119496856,
      "grad_norm": 0.42637163400650024,
      "learning_rate": 4.637054507337527e-05,
      "loss": 0.5532,
      "num_input_tokens_seen": 2318832,
      "step": 3540
    },
    {
      "epoch": 1.857966457023061,
      "grad_norm": 0.5204555988311768,
      "learning_rate": 4.6436058700209645e-05,
      "loss": 0.5244,
      "num_input_tokens_seen": 2321840,
      "step": 3545
    },
    {
      "epoch": 1.8605870020964361,
      "grad_norm": 0.48170772194862366,
      "learning_rate": 4.650157232704403e-05,
      "loss": 0.6137,
      "num_input_tokens_seen": 2324400,
      "step": 3550
    },
    {
      "epoch": 1.8632075471698113,
      "grad_norm": 0.4245053827762604,
      "learning_rate": 4.656708595387841e-05,
      "loss": 0.468,
      "num_input_tokens_seen": 2327376,
      "step": 3555
    },
    {
      "epoch": 1.8658280922431865,
      "grad_norm": 0.21125857532024384,
      "learning_rate": 4.663259958071279e-05,
      "loss": 0.4556,
      "num_input_tokens_seen": 2330832,
      "step": 3560
    },
    {
      "epoch": 1.8684486373165619,
      "grad_norm": 0.5338155031204224,
      "learning_rate": 4.669811320754717e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 2335376,
      "step": 3565
    },
    {
      "epoch": 1.8710691823899372,
      "grad_norm": 0.46195515990257263,
      "learning_rate": 4.676362683438155e-05,
      "loss": 0.5594,
      "num_input_tokens_seen": 2338128,
      "step": 3570
    },
    {
      "epoch": 1.8736897274633124,
      "grad_norm": 0.32860124111175537,
      "learning_rate": 4.6829140461215936e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 2341552,
      "step": 3575
    },
    {
      "epoch": 1.8763102725366876,
      "grad_norm": 0.763809323310852,
      "learning_rate": 4.689465408805032e-05,
      "loss": 0.557,
      "num_input_tokens_seen": 2344304,
      "step": 3580
    },
    {
      "epoch": 1.8789308176100628,
      "grad_norm": 0.40254029631614685,
      "learning_rate": 4.69601677148847e-05,
      "loss": 0.3427,
      "num_input_tokens_seen": 2347216,
      "step": 3585
    },
    {
      "epoch": 1.8815513626834381,
      "grad_norm": 0.31355738639831543,
      "learning_rate": 4.702568134171908e-05,
      "loss": 0.565,
      "num_input_tokens_seen": 2350640,
      "step": 3590
    },
    {
      "epoch": 1.8841719077568135,
      "grad_norm": 0.44015800952911377,
      "learning_rate": 4.709119496855346e-05,
      "loss": 0.5252,
      "num_input_tokens_seen": 2354224,
      "step": 3595
    },
    {
      "epoch": 1.8867924528301887,
      "grad_norm": 0.44539132714271545,
      "learning_rate": 4.7156708595387845e-05,
      "loss": 0.5446,
      "num_input_tokens_seen": 2356688,
      "step": 3600
    },
    {
      "epoch": 1.8894129979035639,
      "grad_norm": 0.7855734825134277,
      "learning_rate": 4.722222222222222e-05,
      "loss": 0.5261,
      "num_input_tokens_seen": 2359568,
      "step": 3605
    },
    {
      "epoch": 1.892033542976939,
      "grad_norm": 0.3332217037677765,
      "learning_rate": 4.728773584905661e-05,
      "loss": 0.6019,
      "num_input_tokens_seen": 2363280,
      "step": 3610
    },
    {
      "epoch": 1.8946540880503144,
      "grad_norm": 0.5584601759910583,
      "learning_rate": 4.735324947589099e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 2366480,
      "step": 3615
    },
    {
      "epoch": 1.8972746331236898,
      "grad_norm": 0.41489744186401367,
      "learning_rate": 4.7418763102725364e-05,
      "loss": 0.6715,
      "num_input_tokens_seen": 2369200,
      "step": 3620
    },
    {
      "epoch": 1.899895178197065,
      "grad_norm": 0.28779441118240356,
      "learning_rate": 4.7484276729559753e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 2373296,
      "step": 3625
    },
    {
      "epoch": 1.9025157232704402,
      "grad_norm": 0.3000185787677765,
      "learning_rate": 4.754979035639413e-05,
      "loss": 0.5133,
      "num_input_tokens_seen": 2376432,
      "step": 3630
    },
    {
      "epoch": 1.9051362683438156,
      "grad_norm": 0.4513719081878662,
      "learning_rate": 4.761530398322851e-05,
      "loss": 0.4789,
      "num_input_tokens_seen": 2379344,
      "step": 3635
    },
    {
      "epoch": 1.9077568134171907,
      "grad_norm": 0.34539419412612915,
      "learning_rate": 4.7680817610062896e-05,
      "loss": 0.6101,
      "num_input_tokens_seen": 2382960,
      "step": 3640
    },
    {
      "epoch": 1.9103773584905661,
      "grad_norm": 0.5721227526664734,
      "learning_rate": 4.774633123689728e-05,
      "loss": 0.4958,
      "num_input_tokens_seen": 2386064,
      "step": 3645
    },
    {
      "epoch": 1.9129979035639413,
      "grad_norm": 0.4363231360912323,
      "learning_rate": 4.7811844863731655e-05,
      "loss": 0.5159,
      "num_input_tokens_seen": 2390384,
      "step": 3650
    },
    {
      "epoch": 1.9156184486373165,
      "grad_norm": 0.41266536712646484,
      "learning_rate": 4.787735849056604e-05,
      "loss": 0.6017,
      "num_input_tokens_seen": 2394384,
      "step": 3655
    },
    {
      "epoch": 1.9182389937106918,
      "grad_norm": 1.2100493907928467,
      "learning_rate": 4.794287211740042e-05,
      "loss": 0.4877,
      "num_input_tokens_seen": 2397584,
      "step": 3660
    },
    {
      "epoch": 1.9208595387840672,
      "grad_norm": 0.6407088041305542,
      "learning_rate": 4.8008385744234804e-05,
      "loss": 0.5707,
      "num_input_tokens_seen": 2400560,
      "step": 3665
    },
    {
      "epoch": 1.9234800838574424,
      "grad_norm": 0.7161715030670166,
      "learning_rate": 4.807389937106919e-05,
      "loss": 0.5269,
      "num_input_tokens_seen": 2403280,
      "step": 3670
    },
    {
      "epoch": 1.9261006289308176,
      "grad_norm": 0.810691773891449,
      "learning_rate": 4.8139412997903564e-05,
      "loss": 0.6601,
      "num_input_tokens_seen": 2406192,
      "step": 3675
    },
    {
      "epoch": 1.9287211740041927,
      "grad_norm": 0.9129345417022705,
      "learning_rate": 4.8204926624737954e-05,
      "loss": 0.4934,
      "num_input_tokens_seen": 2409584,
      "step": 3680
    },
    {
      "epoch": 1.9313417190775681,
      "grad_norm": 0.5142127275466919,
      "learning_rate": 4.827044025157233e-05,
      "loss": 0.4993,
      "num_input_tokens_seen": 2412400,
      "step": 3685
    },
    {
      "epoch": 1.9339622641509435,
      "grad_norm": 1.0649254322052002,
      "learning_rate": 4.8335953878406706e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 2415504,
      "step": 3690
    },
    {
      "epoch": 1.9365828092243187,
      "grad_norm": 0.3486470878124237,
      "learning_rate": 4.8401467505241096e-05,
      "loss": 0.5534,
      "num_input_tokens_seen": 2418256,
      "step": 3695
    },
    {
      "epoch": 1.9392033542976939,
      "grad_norm": 0.5005247592926025,
      "learning_rate": 4.846698113207547e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 2421776,
      "step": 3700
    },
    {
      "epoch": 1.941823899371069,
      "grad_norm": 0.4407968819141388,
      "learning_rate": 4.8532494758909855e-05,
      "loss": 0.6574,
      "num_input_tokens_seen": 2425808,
      "step": 3705
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 0.36269301176071167,
      "learning_rate": 4.859800838574424e-05,
      "loss": 0.4014,
      "num_input_tokens_seen": 2428880,
      "step": 3710
    },
    {
      "epoch": 1.9470649895178198,
      "grad_norm": 0.5756610035896301,
      "learning_rate": 4.8663522012578615e-05,
      "loss": 0.534,
      "num_input_tokens_seen": 2432112,
      "step": 3715
    },
    {
      "epoch": 1.949685534591195,
      "grad_norm": 0.6445869207382202,
      "learning_rate": 4.8729035639413e-05,
      "loss": 0.4097,
      "num_input_tokens_seen": 2435120,
      "step": 3720
    },
    {
      "epoch": 1.9523060796645701,
      "grad_norm": 0.6140881180763245,
      "learning_rate": 4.879454926624738e-05,
      "loss": 0.6057,
      "num_input_tokens_seen": 2437392,
      "step": 3725
    },
    {
      "epoch": 1.9549266247379455,
      "grad_norm": 0.456304669380188,
      "learning_rate": 4.8860062893081764e-05,
      "loss": 0.5089,
      "num_input_tokens_seen": 2439920,
      "step": 3730
    },
    {
      "epoch": 1.9575471698113207,
      "grad_norm": 0.32572707533836365,
      "learning_rate": 4.892557651991615e-05,
      "loss": 0.414,
      "num_input_tokens_seen": 2443536,
      "step": 3735
    },
    {
      "epoch": 1.960167714884696,
      "grad_norm": 0.3502039313316345,
      "learning_rate": 4.899109014675053e-05,
      "loss": 0.5459,
      "num_input_tokens_seen": 2446096,
      "step": 3740
    },
    {
      "epoch": 1.9627882599580713,
      "grad_norm": 0.687382161617279,
      "learning_rate": 4.9056603773584906e-05,
      "loss": 0.605,
      "num_input_tokens_seen": 2448976,
      "step": 3745
    },
    {
      "epoch": 1.9654088050314464,
      "grad_norm": 0.38137558102607727,
      "learning_rate": 4.912211740041929e-05,
      "loss": 0.615,
      "num_input_tokens_seen": 2451632,
      "step": 3750
    },
    {
      "epoch": 1.9680293501048218,
      "grad_norm": 0.46966156363487244,
      "learning_rate": 4.918763102725367e-05,
      "loss": 0.6593,
      "num_input_tokens_seen": 2454928,
      "step": 3755
    },
    {
      "epoch": 1.9706498951781972,
      "grad_norm": 0.6439046859741211,
      "learning_rate": 4.925314465408805e-05,
      "loss": 0.5229,
      "num_input_tokens_seen": 2457232,
      "step": 3760
    },
    {
      "epoch": 1.9732704402515724,
      "grad_norm": 0.6238644123077393,
      "learning_rate": 4.931865828092244e-05,
      "loss": 0.6242,
      "num_input_tokens_seen": 2460176,
      "step": 3765
    },
    {
      "epoch": 1.9758909853249476,
      "grad_norm": 0.5397049784660339,
      "learning_rate": 4.9384171907756815e-05,
      "loss": 0.6487,
      "num_input_tokens_seen": 2464432,
      "step": 3770
    },
    {
      "epoch": 1.9785115303983227,
      "grad_norm": 0.504950761795044,
      "learning_rate": 4.944968553459119e-05,
      "loss": 0.5755,
      "num_input_tokens_seen": 2469136,
      "step": 3775
    },
    {
      "epoch": 1.9811320754716981,
      "grad_norm": 0.3264515697956085,
      "learning_rate": 4.951519916142558e-05,
      "loss": 0.4697,
      "num_input_tokens_seen": 2472400,
      "step": 3780
    },
    {
      "epoch": 1.9837526205450735,
      "grad_norm": 0.4945368766784668,
      "learning_rate": 4.958071278825996e-05,
      "loss": 0.4588,
      "num_input_tokens_seen": 2477328,
      "step": 3785
    },
    {
      "epoch": 1.9863731656184487,
      "grad_norm": 0.501618504524231,
      "learning_rate": 4.964622641509434e-05,
      "loss": 0.5454,
      "num_input_tokens_seen": 2480656,
      "step": 3790
    },
    {
      "epoch": 1.9889937106918238,
      "grad_norm": 0.6138590574264526,
      "learning_rate": 4.9711740041928724e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 2484816,
      "step": 3795
    },
    {
      "epoch": 1.991614255765199,
      "grad_norm": 0.30146610736846924,
      "learning_rate": 4.977725366876311e-05,
      "loss": 0.5629,
      "num_input_tokens_seen": 2488752,
      "step": 3800
    },
    {
      "epoch": 1.9942348008385744,
      "grad_norm": 0.3862920105457306,
      "learning_rate": 4.984276729559749e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 2491952,
      "step": 3805
    },
    {
      "epoch": 1.9968553459119498,
      "grad_norm": 0.40544769167900085,
      "learning_rate": 4.9908280922431866e-05,
      "loss": 0.5833,
      "num_input_tokens_seen": 2495248,
      "step": 3810
    },
    {
      "epoch": 1.999475890985325,
      "grad_norm": 0.33703407645225525,
      "learning_rate": 4.997379454926625e-05,
      "loss": 0.562,
      "num_input_tokens_seen": 2497840,
      "step": 3815
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.4900906980037689,
      "eval_runtime": 9.77,
      "eval_samples_per_second": 86.796,
      "eval_steps_per_second": 21.699,
      "num_input_tokens_seen": 2497936,
      "step": 3816
    },
    {
      "epoch": 2.0020964360587,
      "grad_norm": 0.3275356590747833,
      "learning_rate": 4.999999905865133e-05,
      "loss": 0.4158,
      "num_input_tokens_seen": 2501104,
      "step": 3820
    },
    {
      "epoch": 2.0047169811320753,
      "grad_norm": 0.6528013944625854,
      "learning_rate": 4.999999330596526e-05,
      "loss": 0.5785,
      "num_input_tokens_seen": 2504624,
      "step": 3825
    },
    {
      "epoch": 2.007337526205451,
      "grad_norm": 0.6280234456062317,
      "learning_rate": 4.9999982323565806e-05,
      "loss": 0.5264,
      "num_input_tokens_seen": 2507024,
      "step": 3830
    },
    {
      "epoch": 2.009958071278826,
      "grad_norm": 0.7375000715255737,
      "learning_rate": 4.999996611145527e-05,
      "loss": 0.5053,
      "num_input_tokens_seen": 2509808,
      "step": 3835
    },
    {
      "epoch": 2.0125786163522013,
      "grad_norm": 0.42413797974586487,
      "learning_rate": 4.9999944669637024e-05,
      "loss": 0.4686,
      "num_input_tokens_seen": 2512944,
      "step": 3840
    },
    {
      "epoch": 2.0151991614255764,
      "grad_norm": 0.6098459959030151,
      "learning_rate": 4.999991799811557e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 2516496,
      "step": 3845
    },
    {
      "epoch": 2.0178197064989516,
      "grad_norm": 0.43358421325683594,
      "learning_rate": 4.99998860968965e-05,
      "loss": 0.3468,
      "num_input_tokens_seen": 2519664,
      "step": 3850
    },
    {
      "epoch": 2.020440251572327,
      "grad_norm": 0.4804440438747406,
      "learning_rate": 4.999984896598646e-05,
      "loss": 0.4257,
      "num_input_tokens_seen": 2522992,
      "step": 3855
    },
    {
      "epoch": 2.0230607966457024,
      "grad_norm": 0.3866780400276184,
      "learning_rate": 4.999980660539323e-05,
      "loss": 0.5688,
      "num_input_tokens_seen": 2526416,
      "step": 3860
    },
    {
      "epoch": 2.0256813417190775,
      "grad_norm": 0.5517774820327759,
      "learning_rate": 4.999975901512567e-05,
      "loss": 0.4525,
      "num_input_tokens_seen": 2529200,
      "step": 3865
    },
    {
      "epoch": 2.0283018867924527,
      "grad_norm": 0.5885216593742371,
      "learning_rate": 4.999970619519374e-05,
      "loss": 0.5456,
      "num_input_tokens_seen": 2532336,
      "step": 3870
    },
    {
      "epoch": 2.030922431865828,
      "grad_norm": 0.3615633547306061,
      "learning_rate": 4.999964814560849e-05,
      "loss": 0.5534,
      "num_input_tokens_seen": 2536784,
      "step": 3875
    },
    {
      "epoch": 2.0335429769392035,
      "grad_norm": 0.4563370645046234,
      "learning_rate": 4.999958486638205e-05,
      "loss": 0.5034,
      "num_input_tokens_seen": 2539824,
      "step": 3880
    },
    {
      "epoch": 2.0361635220125787,
      "grad_norm": 0.30791929364204407,
      "learning_rate": 4.999951635752767e-05,
      "loss": 0.5876,
      "num_input_tokens_seen": 2543664,
      "step": 3885
    },
    {
      "epoch": 2.038784067085954,
      "grad_norm": 0.3270447850227356,
      "learning_rate": 4.999944261905968e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 2547632,
      "step": 3890
    },
    {
      "epoch": 2.041404612159329,
      "grad_norm": 0.33457112312316895,
      "learning_rate": 4.99993636509935e-05,
      "loss": 0.561,
      "num_input_tokens_seen": 2551312,
      "step": 3895
    },
    {
      "epoch": 2.0440251572327046,
      "grad_norm": 1.6710660457611084,
      "learning_rate": 4.999927945334565e-05,
      "loss": 0.6204,
      "num_input_tokens_seen": 2554640,
      "step": 3900
    },
    {
      "epoch": 2.04664570230608,
      "grad_norm": 0.5960668325424194,
      "learning_rate": 4.9999190026133745e-05,
      "loss": 0.3819,
      "num_input_tokens_seen": 2557744,
      "step": 3905
    },
    {
      "epoch": 2.049266247379455,
      "grad_norm": 0.4199926555156708,
      "learning_rate": 4.9999095369376506e-05,
      "loss": 0.509,
      "num_input_tokens_seen": 2560816,
      "step": 3910
    },
    {
      "epoch": 2.05188679245283,
      "grad_norm": 0.6274471879005432,
      "learning_rate": 4.99989954830937e-05,
      "loss": 0.6206,
      "num_input_tokens_seen": 2563376,
      "step": 3915
    },
    {
      "epoch": 2.0545073375262053,
      "grad_norm": 0.4323531687259674,
      "learning_rate": 4.9998890367306255e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 2566640,
      "step": 3920
    },
    {
      "epoch": 2.057127882599581,
      "grad_norm": 0.3950580954551697,
      "learning_rate": 4.999878002203614e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 2570000,
      "step": 3925
    },
    {
      "epoch": 2.059748427672956,
      "grad_norm": 0.3405066132545471,
      "learning_rate": 4.999866444730645e-05,
      "loss": 0.5447,
      "num_input_tokens_seen": 2572912,
      "step": 3930
    },
    {
      "epoch": 2.0623689727463312,
      "grad_norm": 0.5552766919136047,
      "learning_rate": 4.999854364314136e-05,
      "loss": 0.5167,
      "num_input_tokens_seen": 2576208,
      "step": 3935
    },
    {
      "epoch": 2.0649895178197064,
      "grad_norm": 0.5057927370071411,
      "learning_rate": 4.999841760956614e-05,
      "loss": 0.648,
      "num_input_tokens_seen": 2580176,
      "step": 3940
    },
    {
      "epoch": 2.0676100628930816,
      "grad_norm": 0.6546316742897034,
      "learning_rate": 4.999828634660715e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 2583504,
      "step": 3945
    },
    {
      "epoch": 2.070230607966457,
      "grad_norm": 0.9485448598861694,
      "learning_rate": 4.9998149854291856e-05,
      "loss": 0.5925,
      "num_input_tokens_seen": 2586928,
      "step": 3950
    },
    {
      "epoch": 2.0728511530398324,
      "grad_norm": 0.3281436860561371,
      "learning_rate": 4.99980081326488e-05,
      "loss": 0.5447,
      "num_input_tokens_seen": 2589808,
      "step": 3955
    },
    {
      "epoch": 2.0754716981132075,
      "grad_norm": 0.4288594424724579,
      "learning_rate": 4.999786118170763e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 2592720,
      "step": 3960
    },
    {
      "epoch": 2.0780922431865827,
      "grad_norm": 0.47140851616859436,
      "learning_rate": 4.99977090014991e-05,
      "loss": 0.5792,
      "num_input_tokens_seen": 2595056,
      "step": 3965
    },
    {
      "epoch": 2.080712788259958,
      "grad_norm": 1.0509934425354004,
      "learning_rate": 4.9997551592055034e-05,
      "loss": 0.4018,
      "num_input_tokens_seen": 2598256,
      "step": 3970
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 0.477696031332016,
      "learning_rate": 4.999738895340836e-05,
      "loss": 0.5934,
      "num_input_tokens_seen": 2602064,
      "step": 3975
    },
    {
      "epoch": 2.0859538784067087,
      "grad_norm": 0.34219276905059814,
      "learning_rate": 4.999722108559311e-05,
      "loss": 0.5609,
      "num_input_tokens_seen": 2605232,
      "step": 3980
    },
    {
      "epoch": 2.088574423480084,
      "grad_norm": 0.33968740701675415,
      "learning_rate": 4.9997047988644385e-05,
      "loss": 0.5877,
      "num_input_tokens_seen": 2608336,
      "step": 3985
    },
    {
      "epoch": 2.091194968553459,
      "grad_norm": 0.2721796929836273,
      "learning_rate": 4.99968696625984e-05,
      "loss": 0.4606,
      "num_input_tokens_seen": 2611472,
      "step": 3990
    },
    {
      "epoch": 2.0938155136268346,
      "grad_norm": 0.2881348133087158,
      "learning_rate": 4.999668610749246e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 2614704,
      "step": 3995
    },
    {
      "epoch": 2.0964360587002098,
      "grad_norm": 0.4017444849014282,
      "learning_rate": 4.999649732336498e-05,
      "loss": 0.5012,
      "num_input_tokens_seen": 2617328,
      "step": 4000
    },
    {
      "epoch": 2.099056603773585,
      "grad_norm": 0.34844622015953064,
      "learning_rate": 4.999630331025542e-05,
      "loss": 0.3981,
      "num_input_tokens_seen": 2620656,
      "step": 4005
    },
    {
      "epoch": 2.10167714884696,
      "grad_norm": 0.6767883896827698,
      "learning_rate": 4.9996104068204385e-05,
      "loss": 0.6054,
      "num_input_tokens_seen": 2624464,
      "step": 4010
    },
    {
      "epoch": 2.1042976939203353,
      "grad_norm": 0.48706933856010437,
      "learning_rate": 4.999589959725355e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 2629296,
      "step": 4015
    },
    {
      "epoch": 2.106918238993711,
      "grad_norm": 0.58553147315979,
      "learning_rate": 4.999568989744568e-05,
      "loss": 0.4731,
      "num_input_tokens_seen": 2631792,
      "step": 4020
    },
    {
      "epoch": 2.109538784067086,
      "grad_norm": 0.3970984220504761,
      "learning_rate": 4.999547496882466e-05,
      "loss": 0.4155,
      "num_input_tokens_seen": 2635024,
      "step": 4025
    },
    {
      "epoch": 2.1121593291404612,
      "grad_norm": 0.30777376890182495,
      "learning_rate": 4.999525481143543e-05,
      "loss": 0.4447,
      "num_input_tokens_seen": 2638800,
      "step": 4030
    },
    {
      "epoch": 2.1147798742138364,
      "grad_norm": 0.2306014597415924,
      "learning_rate": 4.999502942532406e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 2642800,
      "step": 4035
    },
    {
      "epoch": 2.1174004192872116,
      "grad_norm": 0.4418477714061737,
      "learning_rate": 4.999479881053769e-05,
      "loss": 0.4622,
      "num_input_tokens_seen": 2645104,
      "step": 4040
    },
    {
      "epoch": 2.120020964360587,
      "grad_norm": 0.4893420338630676,
      "learning_rate": 4.999456296712457e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 2648144,
      "step": 4045
    },
    {
      "epoch": 2.1226415094339623,
      "grad_norm": 0.44544801115989685,
      "learning_rate": 4.9994321895134025e-05,
      "loss": 0.5638,
      "num_input_tokens_seen": 2651152,
      "step": 4050
    },
    {
      "epoch": 2.1252620545073375,
      "grad_norm": 0.38847440481185913,
      "learning_rate": 4.999407559461649e-05,
      "loss": 0.4595,
      "num_input_tokens_seen": 2657232,
      "step": 4055
    },
    {
      "epoch": 2.1278825995807127,
      "grad_norm": 0.3894541263580322,
      "learning_rate": 4.999382406562349e-05,
      "loss": 0.3742,
      "num_input_tokens_seen": 2660112,
      "step": 4060
    },
    {
      "epoch": 2.130503144654088,
      "grad_norm": 0.6211386322975159,
      "learning_rate": 4.999356730820765e-05,
      "loss": 0.759,
      "num_input_tokens_seen": 2664400,
      "step": 4065
    },
    {
      "epoch": 2.1331236897274635,
      "grad_norm": 0.5371118783950806,
      "learning_rate": 4.999330532242266e-05,
      "loss": 0.5377,
      "num_input_tokens_seen": 2667632,
      "step": 4070
    },
    {
      "epoch": 2.1357442348008386,
      "grad_norm": 0.4019675850868225,
      "learning_rate": 4.999303810832333e-05,
      "loss": 0.4745,
      "num_input_tokens_seen": 2671504,
      "step": 4075
    },
    {
      "epoch": 2.138364779874214,
      "grad_norm": 0.32578957080841064,
      "learning_rate": 4.999276566596557e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 2674384,
      "step": 4080
    },
    {
      "epoch": 2.140985324947589,
      "grad_norm": 0.3133445680141449,
      "learning_rate": 4.9992487995406364e-05,
      "loss": 0.501,
      "num_input_tokens_seen": 2677392,
      "step": 4085
    },
    {
      "epoch": 2.1436058700209646,
      "grad_norm": 0.5110812783241272,
      "learning_rate": 4.9992205096703806e-05,
      "loss": 0.6302,
      "num_input_tokens_seen": 2680240,
      "step": 4090
    },
    {
      "epoch": 2.1462264150943398,
      "grad_norm": 0.38739919662475586,
      "learning_rate": 4.9991916969917054e-05,
      "loss": 0.5331,
      "num_input_tokens_seen": 2683888,
      "step": 4095
    },
    {
      "epoch": 2.148846960167715,
      "grad_norm": 1.0954504013061523,
      "learning_rate": 4.99916236151064e-05,
      "loss": 0.569,
      "num_input_tokens_seen": 2686352,
      "step": 4100
    },
    {
      "epoch": 2.15146750524109,
      "grad_norm": 0.4263034164905548,
      "learning_rate": 4.999132503233321e-05,
      "loss": 0.5259,
      "num_input_tokens_seen": 2688560,
      "step": 4105
    },
    {
      "epoch": 2.1540880503144653,
      "grad_norm": 0.4237642288208008,
      "learning_rate": 4.999102122165994e-05,
      "loss": 0.435,
      "num_input_tokens_seen": 2691056,
      "step": 4110
    },
    {
      "epoch": 2.156708595387841,
      "grad_norm": 0.3437797427177429,
      "learning_rate": 4.9990712183150135e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 2694032,
      "step": 4115
    },
    {
      "epoch": 2.159329140461216,
      "grad_norm": 1.1931719779968262,
      "learning_rate": 4.9990397916868456e-05,
      "loss": 0.5668,
      "num_input_tokens_seen": 2696944,
      "step": 4120
    },
    {
      "epoch": 2.161949685534591,
      "grad_norm": 0.5498374700546265,
      "learning_rate": 4.999007842288064e-05,
      "loss": 0.5855,
      "num_input_tokens_seen": 2699632,
      "step": 4125
    },
    {
      "epoch": 2.1645702306079664,
      "grad_norm": 0.6017036437988281,
      "learning_rate": 4.9989753701253514e-05,
      "loss": 0.4728,
      "num_input_tokens_seen": 2703312,
      "step": 4130
    },
    {
      "epoch": 2.1671907756813416,
      "grad_norm": 0.5072382688522339,
      "learning_rate": 4.998942375205502e-05,
      "loss": 0.5365,
      "num_input_tokens_seen": 2706928,
      "step": 4135
    },
    {
      "epoch": 2.169811320754717,
      "grad_norm": 0.5449848175048828,
      "learning_rate": 4.998908857535416e-05,
      "loss": 0.6242,
      "num_input_tokens_seen": 2710992,
      "step": 4140
    },
    {
      "epoch": 2.1724318658280923,
      "grad_norm": 0.3982486128807068,
      "learning_rate": 4.9988748171221065e-05,
      "loss": 0.4804,
      "num_input_tokens_seen": 2714544,
      "step": 4145
    },
    {
      "epoch": 2.1750524109014675,
      "grad_norm": 0.6768431067466736,
      "learning_rate": 4.998840253972695e-05,
      "loss": 0.6404,
      "num_input_tokens_seen": 2718064,
      "step": 4150
    },
    {
      "epoch": 2.1776729559748427,
      "grad_norm": 0.9070594310760498,
      "learning_rate": 4.9988051680944094e-05,
      "loss": 0.6543,
      "num_input_tokens_seen": 2721648,
      "step": 4155
    },
    {
      "epoch": 2.180293501048218,
      "grad_norm": 0.6221420168876648,
      "learning_rate": 4.998769559494591e-05,
      "loss": 0.4979,
      "num_input_tokens_seen": 2725200,
      "step": 4160
    },
    {
      "epoch": 2.1829140461215935,
      "grad_norm": 0.39443185925483704,
      "learning_rate": 4.998733428180689e-05,
      "loss": 0.514,
      "num_input_tokens_seen": 2728432,
      "step": 4165
    },
    {
      "epoch": 2.1855345911949686,
      "grad_norm": 0.44083961844444275,
      "learning_rate": 4.998696774160259e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 2732688,
      "step": 4170
    },
    {
      "epoch": 2.188155136268344,
      "grad_norm": 0.4505149722099304,
      "learning_rate": 4.9986595974409714e-05,
      "loss": 0.5581,
      "num_input_tokens_seen": 2738448,
      "step": 4175
    },
    {
      "epoch": 2.190775681341719,
      "grad_norm": 0.47586289048194885,
      "learning_rate": 4.998621898030603e-05,
      "loss": 0.4775,
      "num_input_tokens_seen": 2741552,
      "step": 4180
    },
    {
      "epoch": 2.1933962264150946,
      "grad_norm": 0.6734265089035034,
      "learning_rate": 4.998583675937039e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 2744816,
      "step": 4185
    },
    {
      "epoch": 2.1960167714884697,
      "grad_norm": 0.535045325756073,
      "learning_rate": 4.998544931168274e-05,
      "loss": 0.6026,
      "num_input_tokens_seen": 2748912,
      "step": 4190
    },
    {
      "epoch": 2.198637316561845,
      "grad_norm": 0.5120397210121155,
      "learning_rate": 4.9985056637324154e-05,
      "loss": 0.5484,
      "num_input_tokens_seen": 2752240,
      "step": 4195
    },
    {
      "epoch": 2.20125786163522,
      "grad_norm": 0.3621827960014343,
      "learning_rate": 4.998465873637677e-05,
      "loss": 0.5443,
      "num_input_tokens_seen": 2755216,
      "step": 4200
    },
    {
      "epoch": 2.2038784067085953,
      "grad_norm": 0.7478585243225098,
      "learning_rate": 4.9984255608923806e-05,
      "loss": 0.6149,
      "num_input_tokens_seen": 2758224,
      "step": 4205
    },
    {
      "epoch": 2.2064989517819704,
      "grad_norm": 0.31545475125312805,
      "learning_rate": 4.998384725504961e-05,
      "loss": 0.6434,
      "num_input_tokens_seen": 2761680,
      "step": 4210
    },
    {
      "epoch": 2.209119496855346,
      "grad_norm": 0.31697115302085876,
      "learning_rate": 4.9983433674839596e-05,
      "loss": 0.5122,
      "num_input_tokens_seen": 2764464,
      "step": 4215
    },
    {
      "epoch": 2.211740041928721,
      "grad_norm": 0.37922021746635437,
      "learning_rate": 4.998301486838028e-05,
      "loss": 0.4995,
      "num_input_tokens_seen": 2767600,
      "step": 4220
    },
    {
      "epoch": 2.2143605870020964,
      "grad_norm": 0.41518062353134155,
      "learning_rate": 4.998259083575929e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 2770928,
      "step": 4225
    },
    {
      "epoch": 2.2169811320754715,
      "grad_norm": 0.3189833462238312,
      "learning_rate": 4.99821615770653e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 2774544,
      "step": 4230
    },
    {
      "epoch": 2.219601677148847,
      "grad_norm": 0.6175532937049866,
      "learning_rate": 4.998172709238813e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 2777392,
      "step": 4235
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 0.364234983921051,
      "learning_rate": 4.9981287381818654e-05,
      "loss": 0.569,
      "num_input_tokens_seen": 2780272,
      "step": 4240
    },
    {
      "epoch": 2.2248427672955975,
      "grad_norm": 0.6558355689048767,
      "learning_rate": 4.9980842445448856e-05,
      "loss": 0.5455,
      "num_input_tokens_seen": 2783824,
      "step": 4245
    },
    {
      "epoch": 2.2274633123689727,
      "grad_norm": 0.5286229252815247,
      "learning_rate": 4.9980392283371815e-05,
      "loss": 0.5027,
      "num_input_tokens_seen": 2786992,
      "step": 4250
    },
    {
      "epoch": 2.230083857442348,
      "grad_norm": 0.27726009488105774,
      "learning_rate": 4.99799368956817e-05,
      "loss": 0.3915,
      "num_input_tokens_seen": 2789840,
      "step": 4255
    },
    {
      "epoch": 2.2327044025157234,
      "grad_norm": 0.6377810835838318,
      "learning_rate": 4.997947628247378e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 2792432,
      "step": 4260
    },
    {
      "epoch": 2.2353249475890986,
      "grad_norm": 0.9526160955429077,
      "learning_rate": 4.9979010443844395e-05,
      "loss": 0.6024,
      "num_input_tokens_seen": 2795056,
      "step": 4265
    },
    {
      "epoch": 2.237945492662474,
      "grad_norm": 0.3813949227333069,
      "learning_rate": 4.9978539379891e-05,
      "loss": 0.532,
      "num_input_tokens_seen": 2798416,
      "step": 4270
    },
    {
      "epoch": 2.240566037735849,
      "grad_norm": 0.441756933927536,
      "learning_rate": 4.997806309071214e-05,
      "loss": 0.4183,
      "num_input_tokens_seen": 2801040,
      "step": 4275
    },
    {
      "epoch": 2.243186582809224,
      "grad_norm": 0.38625243306159973,
      "learning_rate": 4.997758157640745e-05,
      "loss": 0.496,
      "num_input_tokens_seen": 2804880,
      "step": 4280
    },
    {
      "epoch": 2.2458071278825997,
      "grad_norm": 0.6709590554237366,
      "learning_rate": 4.997709483707765e-05,
      "loss": 0.4478,
      "num_input_tokens_seen": 2810864,
      "step": 4285
    },
    {
      "epoch": 2.248427672955975,
      "grad_norm": 0.7588622570037842,
      "learning_rate": 4.997660287282456e-05,
      "loss": 0.5116,
      "num_input_tokens_seen": 2813328,
      "step": 4290
    },
    {
      "epoch": 2.25104821802935,
      "grad_norm": 0.5403189063072205,
      "learning_rate": 4.99761056837511e-05,
      "loss": 0.5318,
      "num_input_tokens_seen": 2816016,
      "step": 4295
    },
    {
      "epoch": 2.2536687631027252,
      "grad_norm": 0.4533413052558899,
      "learning_rate": 4.997560326996128e-05,
      "loss": 0.6279,
      "num_input_tokens_seen": 2818512,
      "step": 4300
    },
    {
      "epoch": 2.2562893081761004,
      "grad_norm": 0.31407779455184937,
      "learning_rate": 4.997509563156018e-05,
      "loss": 0.611,
      "num_input_tokens_seen": 2821776,
      "step": 4305
    },
    {
      "epoch": 2.258909853249476,
      "grad_norm": 0.47030770778656006,
      "learning_rate": 4.997458276865401e-05,
      "loss": 0.4747,
      "num_input_tokens_seen": 2825232,
      "step": 4310
    },
    {
      "epoch": 2.261530398322851,
      "grad_norm": 0.5923716425895691,
      "learning_rate": 4.997406468135005e-05,
      "loss": 0.5625,
      "num_input_tokens_seen": 2828592,
      "step": 4315
    },
    {
      "epoch": 2.2641509433962264,
      "grad_norm": 0.7022002935409546,
      "learning_rate": 4.9973541369756686e-05,
      "loss": 0.5163,
      "num_input_tokens_seen": 2831472,
      "step": 4320
    },
    {
      "epoch": 2.2667714884696015,
      "grad_norm": 0.3170570433139801,
      "learning_rate": 4.9973012833983365e-05,
      "loss": 0.3806,
      "num_input_tokens_seen": 2834640,
      "step": 4325
    },
    {
      "epoch": 2.269392033542977,
      "grad_norm": 0.510639488697052,
      "learning_rate": 4.9972479074140684e-05,
      "loss": 0.3875,
      "num_input_tokens_seen": 2838160,
      "step": 4330
    },
    {
      "epoch": 2.2720125786163523,
      "grad_norm": 0.4294596016407013,
      "learning_rate": 4.9971940090340276e-05,
      "loss": 0.3465,
      "num_input_tokens_seen": 2842160,
      "step": 4335
    },
    {
      "epoch": 2.2746331236897275,
      "grad_norm": 0.41547757387161255,
      "learning_rate": 4.997139588269489e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 2845712,
      "step": 4340
    },
    {
      "epoch": 2.2772536687631026,
      "grad_norm": 0.4104653000831604,
      "learning_rate": 4.997084645131838e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 2848976,
      "step": 4345
    },
    {
      "epoch": 2.279874213836478,
      "grad_norm": 0.6132417917251587,
      "learning_rate": 4.9970291796325666e-05,
      "loss": 0.4254,
      "num_input_tokens_seen": 2852304,
      "step": 4350
    },
    {
      "epoch": 2.2824947589098534,
      "grad_norm": 0.31981325149536133,
      "learning_rate": 4.996973191783279e-05,
      "loss": 0.7048,
      "num_input_tokens_seen": 2855152,
      "step": 4355
    },
    {
      "epoch": 2.2851153039832286,
      "grad_norm": 0.6403321623802185,
      "learning_rate": 4.996916681595687e-05,
      "loss": 0.5309,
      "num_input_tokens_seen": 2857968,
      "step": 4360
    },
    {
      "epoch": 2.2877358490566038,
      "grad_norm": 0.32801949977874756,
      "learning_rate": 4.996859649081611e-05,
      "loss": 0.4213,
      "num_input_tokens_seen": 2862032,
      "step": 4365
    },
    {
      "epoch": 2.290356394129979,
      "grad_norm": 0.46767231822013855,
      "learning_rate": 4.996802094252983e-05,
      "loss": 0.5253,
      "num_input_tokens_seen": 2864624,
      "step": 4370
    },
    {
      "epoch": 2.2929769392033545,
      "grad_norm": 0.45213377475738525,
      "learning_rate": 4.996744017121841e-05,
      "loss": 0.5472,
      "num_input_tokens_seen": 2868208,
      "step": 4375
    },
    {
      "epoch": 2.2955974842767297,
      "grad_norm": 0.3259497284889221,
      "learning_rate": 4.9966854177003355e-05,
      "loss": 0.7022,
      "num_input_tokens_seen": 2871056,
      "step": 4380
    },
    {
      "epoch": 2.298218029350105,
      "grad_norm": 0.5604590773582458,
      "learning_rate": 4.996626296000724e-05,
      "loss": 0.5223,
      "num_input_tokens_seen": 2873904,
      "step": 4385
    },
    {
      "epoch": 2.30083857442348,
      "grad_norm": 0.43816709518432617,
      "learning_rate": 4.996566652035374e-05,
      "loss": 0.5124,
      "num_input_tokens_seen": 2877040,
      "step": 4390
    },
    {
      "epoch": 2.3034591194968552,
      "grad_norm": 0.49217358231544495,
      "learning_rate": 4.996506485816763e-05,
      "loss": 0.5709,
      "num_input_tokens_seen": 2880560,
      "step": 4395
    },
    {
      "epoch": 2.3060796645702304,
      "grad_norm": 0.6156191229820251,
      "learning_rate": 4.996445797357477e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 2883920,
      "step": 4400
    },
    {
      "epoch": 2.308700209643606,
      "grad_norm": 0.5675243735313416,
      "learning_rate": 4.996384586670211e-05,
      "loss": 0.4984,
      "num_input_tokens_seen": 2886448,
      "step": 4405
    },
    {
      "epoch": 2.311320754716981,
      "grad_norm": 0.6121154427528381,
      "learning_rate": 4.99632285376777e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 2889520,
      "step": 4410
    },
    {
      "epoch": 2.3139412997903563,
      "grad_norm": 0.24169592559337616,
      "learning_rate": 4.9962605986630674e-05,
      "loss": 0.5086,
      "num_input_tokens_seen": 2893040,
      "step": 4415
    },
    {
      "epoch": 2.3165618448637315,
      "grad_norm": 0.26065966486930847,
      "learning_rate": 4.996197821369126e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 2896880,
      "step": 4420
    },
    {
      "epoch": 2.319182389937107,
      "grad_norm": 0.45667481422424316,
      "learning_rate": 4.9961345218990796e-05,
      "loss": 0.4183,
      "num_input_tokens_seen": 2899600,
      "step": 4425
    },
    {
      "epoch": 2.3218029350104823,
      "grad_norm": 0.33234259486198425,
      "learning_rate": 4.996070700266168e-05,
      "loss": 0.4381,
      "num_input_tokens_seen": 2902608,
      "step": 4430
    },
    {
      "epoch": 2.3244234800838575,
      "grad_norm": 0.24436025321483612,
      "learning_rate": 4.9960063564837424e-05,
      "loss": 0.5422,
      "num_input_tokens_seen": 2906224,
      "step": 4435
    },
    {
      "epoch": 2.3270440251572326,
      "grad_norm": 0.3902607262134552,
      "learning_rate": 4.995941490565262e-05,
      "loss": 0.4146,
      "num_input_tokens_seen": 2909232,
      "step": 4440
    },
    {
      "epoch": 2.329664570230608,
      "grad_norm": 0.943528413772583,
      "learning_rate": 4.995876102524299e-05,
      "loss": 0.4835,
      "num_input_tokens_seen": 2916848,
      "step": 4445
    },
    {
      "epoch": 2.3322851153039834,
      "grad_norm": 0.21343442797660828,
      "learning_rate": 4.995810192374529e-05,
      "loss": 0.5725,
      "num_input_tokens_seen": 2920560,
      "step": 4450
    },
    {
      "epoch": 2.3349056603773586,
      "grad_norm": 0.5326602458953857,
      "learning_rate": 4.99574376012974e-05,
      "loss": 0.5712,
      "num_input_tokens_seen": 2923408,
      "step": 4455
    },
    {
      "epoch": 2.3375262054507338,
      "grad_norm": 0.35463377833366394,
      "learning_rate": 4.995676805803829e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 2927280,
      "step": 4460
    },
    {
      "epoch": 2.340146750524109,
      "grad_norm": 0.5178223252296448,
      "learning_rate": 4.995609329410804e-05,
      "loss": 0.6948,
      "num_input_tokens_seen": 2930800,
      "step": 4465
    },
    {
      "epoch": 2.342767295597484,
      "grad_norm": 0.5247440934181213,
      "learning_rate": 4.9955413309647776e-05,
      "loss": 0.4618,
      "num_input_tokens_seen": 2934448,
      "step": 4470
    },
    {
      "epoch": 2.3453878406708597,
      "grad_norm": 0.6142507195472717,
      "learning_rate": 4.9954728104799756e-05,
      "loss": 0.6994,
      "num_input_tokens_seen": 2937104,
      "step": 4475
    },
    {
      "epoch": 2.348008385744235,
      "grad_norm": 0.33563748002052307,
      "learning_rate": 4.995403767970731e-05,
      "loss": 0.4953,
      "num_input_tokens_seen": 2940368,
      "step": 4480
    },
    {
      "epoch": 2.35062893081761,
      "grad_norm": 0.5940802693367004,
      "learning_rate": 4.995334203451488e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 2942960,
      "step": 4485
    },
    {
      "epoch": 2.353249475890985,
      "grad_norm": 1.0549418926239014,
      "learning_rate": 4.995264116936798e-05,
      "loss": 0.5483,
      "num_input_tokens_seen": 2946704,
      "step": 4490
    },
    {
      "epoch": 2.3558700209643604,
      "grad_norm": 0.8892385959625244,
      "learning_rate": 4.995193508441322e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 2950608,
      "step": 4495
    },
    {
      "epoch": 2.358490566037736,
      "grad_norm": 0.30656734108924866,
      "learning_rate": 4.99512237797983e-05,
      "loss": 0.4799,
      "num_input_tokens_seen": 2954416,
      "step": 4500
    },
    {
      "epoch": 2.361111111111111,
      "grad_norm": 0.30260396003723145,
      "learning_rate": 4.995050725567203e-05,
      "loss": 0.5509,
      "num_input_tokens_seen": 2957424,
      "step": 4505
    },
    {
      "epoch": 2.3637316561844863,
      "grad_norm": 0.6975590586662292,
      "learning_rate": 4.994978551218429e-05,
      "loss": 0.5291,
      "num_input_tokens_seen": 2960400,
      "step": 4510
    },
    {
      "epoch": 2.3663522012578615,
      "grad_norm": 0.44081875681877136,
      "learning_rate": 4.9949058549486066e-05,
      "loss": 0.5633,
      "num_input_tokens_seen": 2963568,
      "step": 4515
    },
    {
      "epoch": 2.368972746331237,
      "grad_norm": 0.5280702710151672,
      "learning_rate": 4.994832636772943e-05,
      "loss": 0.694,
      "num_input_tokens_seen": 2966384,
      "step": 4520
    },
    {
      "epoch": 2.3715932914046123,
      "grad_norm": 0.3112006187438965,
      "learning_rate": 4.994758896706754e-05,
      "loss": 0.356,
      "num_input_tokens_seen": 2969360,
      "step": 4525
    },
    {
      "epoch": 2.3742138364779874,
      "grad_norm": 0.4762044847011566,
      "learning_rate": 4.994684634765465e-05,
      "loss": 0.67,
      "num_input_tokens_seen": 2972560,
      "step": 4530
    },
    {
      "epoch": 2.3768343815513626,
      "grad_norm": 0.6561261415481567,
      "learning_rate": 4.9946098509646125e-05,
      "loss": 0.6645,
      "num_input_tokens_seen": 2975472,
      "step": 4535
    },
    {
      "epoch": 2.379454926624738,
      "grad_norm": 0.46299126744270325,
      "learning_rate": 4.994534545319838e-05,
      "loss": 0.5611,
      "num_input_tokens_seen": 2978288,
      "step": 4540
    },
    {
      "epoch": 2.3820754716981134,
      "grad_norm": 0.3966447710990906,
      "learning_rate": 4.994458717846896e-05,
      "loss": 0.592,
      "num_input_tokens_seen": 2981840,
      "step": 4545
    },
    {
      "epoch": 2.3846960167714886,
      "grad_norm": 0.5258379578590393,
      "learning_rate": 4.9943823685616485e-05,
      "loss": 0.5288,
      "num_input_tokens_seen": 2985104,
      "step": 4550
    },
    {
      "epoch": 2.3873165618448637,
      "grad_norm": 0.3625716269016266,
      "learning_rate": 4.994305497480067e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 2989360,
      "step": 4555
    },
    {
      "epoch": 2.389937106918239,
      "grad_norm": 0.4372296631336212,
      "learning_rate": 4.994228104618233e-05,
      "loss": 0.5156,
      "num_input_tokens_seen": 2993136,
      "step": 4560
    },
    {
      "epoch": 2.392557651991614,
      "grad_norm": 0.4090736508369446,
      "learning_rate": 4.994150189992334e-05,
      "loss": 0.5544,
      "num_input_tokens_seen": 2996848,
      "step": 4565
    },
    {
      "epoch": 2.3951781970649897,
      "grad_norm": 0.5166860818862915,
      "learning_rate": 4.99407175361867e-05,
      "loss": 0.5419,
      "num_input_tokens_seen": 2999952,
      "step": 4570
    },
    {
      "epoch": 2.397798742138365,
      "grad_norm": 0.41518786549568176,
      "learning_rate": 4.99399279551365e-05,
      "loss": 0.5871,
      "num_input_tokens_seen": 3002896,
      "step": 4575
    },
    {
      "epoch": 2.40041928721174,
      "grad_norm": 0.38364964723587036,
      "learning_rate": 4.9939133156937886e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 3005936,
      "step": 4580
    },
    {
      "epoch": 2.403039832285115,
      "grad_norm": 0.5733475089073181,
      "learning_rate": 4.993833314175716e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 3009488,
      "step": 4585
    },
    {
      "epoch": 2.4056603773584904,
      "grad_norm": 0.5358508825302124,
      "learning_rate": 4.993752790976163e-05,
      "loss": 0.4611,
      "num_input_tokens_seen": 3012560,
      "step": 4590
    },
    {
      "epoch": 2.408280922431866,
      "grad_norm": 0.42176559567451477,
      "learning_rate": 4.993671746111977e-05,
      "loss": 0.3643,
      "num_input_tokens_seen": 3015440,
      "step": 4595
    },
    {
      "epoch": 2.410901467505241,
      "grad_norm": 0.3691204786300659,
      "learning_rate": 4.993590179600111e-05,
      "loss": 0.6271,
      "num_input_tokens_seen": 3019088,
      "step": 4600
    },
    {
      "epoch": 2.4135220125786163,
      "grad_norm": 0.7808572053909302,
      "learning_rate": 4.9935080914576284e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 3022384,
      "step": 4605
    },
    {
      "epoch": 2.4161425576519915,
      "grad_norm": 0.2967089116573334,
      "learning_rate": 4.9934254817016995e-05,
      "loss": 0.4002,
      "num_input_tokens_seen": 3026672,
      "step": 4610
    },
    {
      "epoch": 2.418763102725367,
      "grad_norm": 0.3209041357040405,
      "learning_rate": 4.993342350349607e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 3029936,
      "step": 4615
    },
    {
      "epoch": 2.4213836477987423,
      "grad_norm": 0.3865405321121216,
      "learning_rate": 4.9932586974187404e-05,
      "loss": 0.6691,
      "num_input_tokens_seen": 3032752,
      "step": 4620
    },
    {
      "epoch": 2.4240041928721174,
      "grad_norm": 0.5562931299209595,
      "learning_rate": 4.9931745229265986e-05,
      "loss": 0.5628,
      "num_input_tokens_seen": 3036400,
      "step": 4625
    },
    {
      "epoch": 2.4266247379454926,
      "grad_norm": 0.5005999207496643,
      "learning_rate": 4.99308982689079e-05,
      "loss": 0.5205,
      "num_input_tokens_seen": 3039152,
      "step": 4630
    },
    {
      "epoch": 2.4292452830188678,
      "grad_norm": 0.2597678601741791,
      "learning_rate": 4.993004609329033e-05,
      "loss": 0.4033,
      "num_input_tokens_seen": 3042096,
      "step": 4635
    },
    {
      "epoch": 2.431865828092243,
      "grad_norm": 0.4639030694961548,
      "learning_rate": 4.992918870259153e-05,
      "loss": 0.3414,
      "num_input_tokens_seen": 3045136,
      "step": 4640
    },
    {
      "epoch": 2.4344863731656186,
      "grad_norm": 0.7038489580154419,
      "learning_rate": 4.992832609699086e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 3047344,
      "step": 4645
    },
    {
      "epoch": 2.4371069182389937,
      "grad_norm": 0.4306497871875763,
      "learning_rate": 4.992745827666877e-05,
      "loss": 0.5267,
      "num_input_tokens_seen": 3049968,
      "step": 4650
    },
    {
      "epoch": 2.439727463312369,
      "grad_norm": 0.3370714783668518,
      "learning_rate": 4.99265852418068e-05,
      "loss": 0.5817,
      "num_input_tokens_seen": 3052784,
      "step": 4655
    },
    {
      "epoch": 2.442348008385744,
      "grad_norm": 0.2806496322154999,
      "learning_rate": 4.992570699258757e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 3056752,
      "step": 4660
    },
    {
      "epoch": 2.4449685534591197,
      "grad_norm": 0.3123174011707306,
      "learning_rate": 4.99248235291948e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 3059696,
      "step": 4665
    },
    {
      "epoch": 2.447589098532495,
      "grad_norm": 0.5698472261428833,
      "learning_rate": 4.9923934851813314e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 3062608,
      "step": 4670
    },
    {
      "epoch": 2.45020964360587,
      "grad_norm": 0.34024864435195923,
      "learning_rate": 4.9923040960628996e-05,
      "loss": 0.4864,
      "num_input_tokens_seen": 3065712,
      "step": 4675
    },
    {
      "epoch": 2.452830188679245,
      "grad_norm": 0.2419264167547226,
      "learning_rate": 4.9922141855828856e-05,
      "loss": 0.6347,
      "num_input_tokens_seen": 3069936,
      "step": 4680
    },
    {
      "epoch": 2.4554507337526204,
      "grad_norm": 0.43009430170059204,
      "learning_rate": 4.992123753760096e-05,
      "loss": 0.4591,
      "num_input_tokens_seen": 3073072,
      "step": 4685
    },
    {
      "epoch": 2.458071278825996,
      "grad_norm": 0.2653846740722656,
      "learning_rate": 4.9920328006134484e-05,
      "loss": 0.509,
      "num_input_tokens_seen": 3076368,
      "step": 4690
    },
    {
      "epoch": 2.460691823899371,
      "grad_norm": 0.8600221276283264,
      "learning_rate": 4.991941326161971e-05,
      "loss": 0.5063,
      "num_input_tokens_seen": 3079024,
      "step": 4695
    },
    {
      "epoch": 2.4633123689727463,
      "grad_norm": 0.2477535754442215,
      "learning_rate": 4.991849330424796e-05,
      "loss": 0.4816,
      "num_input_tokens_seen": 3085616,
      "step": 4700
    },
    {
      "epoch": 2.4659329140461215,
      "grad_norm": 0.6084620356559753,
      "learning_rate": 4.9917568134211703e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 3088720,
      "step": 4705
    },
    {
      "epoch": 2.468553459119497,
      "grad_norm": 0.41588473320007324,
      "learning_rate": 4.991663775170447e-05,
      "loss": 0.645,
      "num_input_tokens_seen": 3091632,
      "step": 4710
    },
    {
      "epoch": 2.4711740041928723,
      "grad_norm": 0.33066320419311523,
      "learning_rate": 4.991570215692088e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 3094960,
      "step": 4715
    },
    {
      "epoch": 2.4737945492662474,
      "grad_norm": 0.7084963321685791,
      "learning_rate": 4.991476135005665e-05,
      "loss": 0.5499,
      "num_input_tokens_seen": 3097744,
      "step": 4720
    },
    {
      "epoch": 2.4764150943396226,
      "grad_norm": 0.38208815455436707,
      "learning_rate": 4.991381533130859e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 3100656,
      "step": 4725
    },
    {
      "epoch": 2.4790356394129978,
      "grad_norm": 0.5314891338348389,
      "learning_rate": 4.99128641008746e-05,
      "loss": 0.5653,
      "num_input_tokens_seen": 3103632,
      "step": 4730
    },
    {
      "epoch": 2.481656184486373,
      "grad_norm": 0.33032968640327454,
      "learning_rate": 4.9911907658953664e-05,
      "loss": 0.5962,
      "num_input_tokens_seen": 3107216,
      "step": 4735
    },
    {
      "epoch": 2.4842767295597485,
      "grad_norm": 0.8113180994987488,
      "learning_rate": 4.991094600574584e-05,
      "loss": 0.6722,
      "num_input_tokens_seen": 3109680,
      "step": 4740
    },
    {
      "epoch": 2.4868972746331237,
      "grad_norm": 0.3397659957408905,
      "learning_rate": 4.9909979141452324e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 3112624,
      "step": 4745
    },
    {
      "epoch": 2.489517819706499,
      "grad_norm": 0.4305604100227356,
      "learning_rate": 4.9909007066275346e-05,
      "loss": 0.5808,
      "num_input_tokens_seen": 3116048,
      "step": 4750
    },
    {
      "epoch": 2.492138364779874,
      "grad_norm": 0.5336838960647583,
      "learning_rate": 4.990802978041828e-05,
      "loss": 0.573,
      "num_input_tokens_seen": 3118864,
      "step": 4755
    },
    {
      "epoch": 2.4947589098532497,
      "grad_norm": 0.4417584538459778,
      "learning_rate": 4.990704728408555e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 3122192,
      "step": 4760
    },
    {
      "epoch": 2.497379454926625,
      "grad_norm": 0.3388408124446869,
      "learning_rate": 4.990605957748268e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 3128016,
      "step": 4765
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.6262451410293579,
      "learning_rate": 4.990506666081628e-05,
      "loss": 0.4491,
      "num_input_tokens_seen": 3130736,
      "step": 4770
    },
    {
      "epoch": 2.502620545073375,
      "grad_norm": 0.5051946640014648,
      "learning_rate": 4.990406853429408e-05,
      "loss": 0.5426,
      "num_input_tokens_seen": 3133616,
      "step": 4775
    },
    {
      "epoch": 2.5052410901467503,
      "grad_norm": 0.9539276957511902,
      "learning_rate": 4.990306519812485e-05,
      "loss": 0.6133,
      "num_input_tokens_seen": 3136848,
      "step": 4780
    },
    {
      "epoch": 2.507861635220126,
      "grad_norm": 0.4823147654533386,
      "learning_rate": 4.99020566525185e-05,
      "loss": 0.5288,
      "num_input_tokens_seen": 3139600,
      "step": 4785
    },
    {
      "epoch": 2.510482180293501,
      "grad_norm": 0.39631807804107666,
      "learning_rate": 4.9901042897685996e-05,
      "loss": 0.5143,
      "num_input_tokens_seen": 3142896,
      "step": 4790
    },
    {
      "epoch": 2.5131027253668763,
      "grad_norm": 0.3393003046512604,
      "learning_rate": 4.99000239338394e-05,
      "loss": 0.4315,
      "num_input_tokens_seen": 3147248,
      "step": 4795
    },
    {
      "epoch": 2.5157232704402515,
      "grad_norm": 0.5131325125694275,
      "learning_rate": 4.989899976119187e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 3150288,
      "step": 4800
    },
    {
      "epoch": 2.518343815513627,
      "grad_norm": 0.22959139943122864,
      "learning_rate": 4.989797037995766e-05,
      "loss": 0.4834,
      "num_input_tokens_seen": 3153808,
      "step": 4805
    },
    {
      "epoch": 2.5209643605870022,
      "grad_norm": 0.4151860177516937,
      "learning_rate": 4.989693579035209e-05,
      "loss": 0.6027,
      "num_input_tokens_seen": 3156784,
      "step": 4810
    },
    {
      "epoch": 2.5235849056603774,
      "grad_norm": 0.6952019929885864,
      "learning_rate": 4.989589599259159e-05,
      "loss": 0.5938,
      "num_input_tokens_seen": 3160176,
      "step": 4815
    },
    {
      "epoch": 2.5262054507337526,
      "grad_norm": 1.096691608428955,
      "learning_rate": 4.9894850986893685e-05,
      "loss": 0.6189,
      "num_input_tokens_seen": 3162608,
      "step": 4820
    },
    {
      "epoch": 2.5288259958071277,
      "grad_norm": 0.4078587591648102,
      "learning_rate": 4.9893800773476965e-05,
      "loss": 0.5563,
      "num_input_tokens_seen": 3166576,
      "step": 4825
    },
    {
      "epoch": 2.531446540880503,
      "grad_norm": 0.40279796719551086,
      "learning_rate": 4.9892745352561134e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 3169552,
      "step": 4830
    },
    {
      "epoch": 2.5340670859538785,
      "grad_norm": 0.26333844661712646,
      "learning_rate": 4.9891684724366955e-05,
      "loss": 0.4993,
      "num_input_tokens_seen": 3173616,
      "step": 4835
    },
    {
      "epoch": 2.5366876310272537,
      "grad_norm": 0.398740291595459,
      "learning_rate": 4.989061888911631e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 3176816,
      "step": 4840
    },
    {
      "epoch": 2.539308176100629,
      "grad_norm": 0.3426074683666229,
      "learning_rate": 4.988954784703217e-05,
      "loss": 0.476,
      "num_input_tokens_seen": 3179824,
      "step": 4845
    },
    {
      "epoch": 2.541928721174004,
      "grad_norm": 0.41728678345680237,
      "learning_rate": 4.9888471598338574e-05,
      "loss": 0.3806,
      "num_input_tokens_seen": 3183056,
      "step": 4850
    },
    {
      "epoch": 2.5445492662473796,
      "grad_norm": 0.5351296067237854,
      "learning_rate": 4.988739014326066e-05,
      "loss": 0.4593,
      "num_input_tokens_seen": 3187216,
      "step": 4855
    },
    {
      "epoch": 2.547169811320755,
      "grad_norm": 0.36407920718193054,
      "learning_rate": 4.9886303482024656e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 3190576,
      "step": 4860
    },
    {
      "epoch": 2.54979035639413,
      "grad_norm": 0.2594355642795563,
      "learning_rate": 4.988521161485789e-05,
      "loss": 0.582,
      "num_input_tokens_seen": 3194000,
      "step": 4865
    },
    {
      "epoch": 2.552410901467505,
      "grad_norm": 0.23808975517749786,
      "learning_rate": 4.9884114541988747e-05,
      "loss": 0.4389,
      "num_input_tokens_seen": 3197232,
      "step": 4870
    },
    {
      "epoch": 2.5550314465408803,
      "grad_norm": 0.4145663380622864,
      "learning_rate": 4.988301226364674e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 3200592,
      "step": 4875
    },
    {
      "epoch": 2.5576519916142555,
      "grad_norm": 0.38026493787765503,
      "learning_rate": 4.988190478006245e-05,
      "loss": 0.63,
      "num_input_tokens_seen": 3204240,
      "step": 4880
    },
    {
      "epoch": 2.560272536687631,
      "grad_norm": 0.37682077288627625,
      "learning_rate": 4.988079209146754e-05,
      "loss": 0.4497,
      "num_input_tokens_seen": 3207216,
      "step": 4885
    },
    {
      "epoch": 2.5628930817610063,
      "grad_norm": 0.38509365916252136,
      "learning_rate": 4.9879674198094794e-05,
      "loss": 0.5316,
      "num_input_tokens_seen": 3210096,
      "step": 4890
    },
    {
      "epoch": 2.5655136268343814,
      "grad_norm": 0.9665071964263916,
      "learning_rate": 4.987855110017804e-05,
      "loss": 0.682,
      "num_input_tokens_seen": 3213104,
      "step": 4895
    },
    {
      "epoch": 2.568134171907757,
      "grad_norm": 0.30184730887413025,
      "learning_rate": 4.9877422797952224e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 3216432,
      "step": 4900
    },
    {
      "epoch": 2.5707547169811322,
      "grad_norm": 0.26569876074790955,
      "learning_rate": 4.987628929165338e-05,
      "loss": 0.5609,
      "num_input_tokens_seen": 3219696,
      "step": 4905
    },
    {
      "epoch": 2.5733752620545074,
      "grad_norm": 0.9278947710990906,
      "learning_rate": 4.9875150581518615e-05,
      "loss": 0.5407,
      "num_input_tokens_seen": 3222448,
      "step": 4910
    },
    {
      "epoch": 2.5759958071278826,
      "grad_norm": 0.2309046983718872,
      "learning_rate": 4.987400666778614e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 3226128,
      "step": 4915
    },
    {
      "epoch": 2.5786163522012577,
      "grad_norm": 1.0854445695877075,
      "learning_rate": 4.987285755069524e-05,
      "loss": 0.7414,
      "num_input_tokens_seen": 3229232,
      "step": 4920
    },
    {
      "epoch": 2.581236897274633,
      "grad_norm": 0.3355877101421356,
      "learning_rate": 4.987170323048632e-05,
      "loss": 0.5322,
      "num_input_tokens_seen": 3232048,
      "step": 4925
    },
    {
      "epoch": 2.5838574423480085,
      "grad_norm": 0.6738042235374451,
      "learning_rate": 4.987054370740083e-05,
      "loss": 0.6527,
      "num_input_tokens_seen": 3235120,
      "step": 4930
    },
    {
      "epoch": 2.5864779874213837,
      "grad_norm": 0.923831582069397,
      "learning_rate": 4.986937898168134e-05,
      "loss": 0.4935,
      "num_input_tokens_seen": 3238224,
      "step": 4935
    },
    {
      "epoch": 2.589098532494759,
      "grad_norm": 0.6806281208992004,
      "learning_rate": 4.986820905357148e-05,
      "loss": 0.5452,
      "num_input_tokens_seen": 3241744,
      "step": 4940
    },
    {
      "epoch": 2.591719077568134,
      "grad_norm": 0.9399994015693665,
      "learning_rate": 4.9867033923315995e-05,
      "loss": 0.5686,
      "num_input_tokens_seen": 3244528,
      "step": 4945
    },
    {
      "epoch": 2.5943396226415096,
      "grad_norm": 0.5232332348823547,
      "learning_rate": 4.9865853591160724e-05,
      "loss": 0.5752,
      "num_input_tokens_seen": 3247952,
      "step": 4950
    },
    {
      "epoch": 2.596960167714885,
      "grad_norm": 0.3049243092536926,
      "learning_rate": 4.9864668057352556e-05,
      "loss": 0.5971,
      "num_input_tokens_seen": 3251120,
      "step": 4955
    },
    {
      "epoch": 2.59958071278826,
      "grad_norm": 0.7418012022972107,
      "learning_rate": 4.9863477322139517e-05,
      "loss": 0.6623,
      "num_input_tokens_seen": 3255408,
      "step": 4960
    },
    {
      "epoch": 2.602201257861635,
      "grad_norm": 0.34134426712989807,
      "learning_rate": 4.986228138577067e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 3258832,
      "step": 4965
    },
    {
      "epoch": 2.6048218029350103,
      "grad_norm": 0.8299354910850525,
      "learning_rate": 4.986108024849619e-05,
      "loss": 0.5218,
      "num_input_tokens_seen": 3261072,
      "step": 4970
    },
    {
      "epoch": 2.6074423480083855,
      "grad_norm": 0.7661663889884949,
      "learning_rate": 4.985987391056737e-05,
      "loss": 0.6159,
      "num_input_tokens_seen": 3263760,
      "step": 4975
    },
    {
      "epoch": 2.610062893081761,
      "grad_norm": 0.5416978001594543,
      "learning_rate": 4.9858662372236544e-05,
      "loss": 0.5822,
      "num_input_tokens_seen": 3266192,
      "step": 4980
    },
    {
      "epoch": 2.6126834381551363,
      "grad_norm": 0.7956660389900208,
      "learning_rate": 4.985744563375714e-05,
      "loss": 0.5009,
      "num_input_tokens_seen": 3268656,
      "step": 4985
    },
    {
      "epoch": 2.6153039832285114,
      "grad_norm": 0.4792504608631134,
      "learning_rate": 4.985622369538371e-05,
      "loss": 0.532,
      "num_input_tokens_seen": 3272112,
      "step": 4990
    },
    {
      "epoch": 2.617924528301887,
      "grad_norm": 0.52021324634552,
      "learning_rate": 4.985499655737185e-05,
      "loss": 0.5739,
      "num_input_tokens_seen": 3275056,
      "step": 4995
    },
    {
      "epoch": 2.620545073375262,
      "grad_norm": 0.5916310548782349,
      "learning_rate": 4.985376421997827e-05,
      "loss": 0.4341,
      "num_input_tokens_seen": 3278416,
      "step": 5000
    },
    {
      "epoch": 2.6231656184486374,
      "grad_norm": 0.8033234477043152,
      "learning_rate": 4.985252668346076e-05,
      "loss": 0.5075,
      "num_input_tokens_seen": 3281616,
      "step": 5005
    },
    {
      "epoch": 2.6257861635220126,
      "grad_norm": 0.3850017488002777,
      "learning_rate": 4.9851283948078206e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 3285008,
      "step": 5010
    },
    {
      "epoch": 2.6284067085953877,
      "grad_norm": 0.6302533149719238,
      "learning_rate": 4.985003601409057e-05,
      "loss": 0.4964,
      "num_input_tokens_seen": 3288112,
      "step": 5015
    },
    {
      "epoch": 2.631027253668763,
      "grad_norm": 1.6596513986587524,
      "learning_rate": 4.9848782881758906e-05,
      "loss": 0.5574,
      "num_input_tokens_seen": 3291408,
      "step": 5020
    },
    {
      "epoch": 2.6336477987421385,
      "grad_norm": 0.4511530101299286,
      "learning_rate": 4.9847524551345345e-05,
      "loss": 0.5229,
      "num_input_tokens_seen": 3294128,
      "step": 5025
    },
    {
      "epoch": 2.6362683438155137,
      "grad_norm": 0.595307469367981,
      "learning_rate": 4.984626102311313e-05,
      "loss": 0.4448,
      "num_input_tokens_seen": 3296688,
      "step": 5030
    },
    {
      "epoch": 2.638888888888889,
      "grad_norm": 0.941825270652771,
      "learning_rate": 4.984499229732656e-05,
      "loss": 0.3264,
      "num_input_tokens_seen": 3300528,
      "step": 5035
    },
    {
      "epoch": 2.641509433962264,
      "grad_norm": 0.6187476515769958,
      "learning_rate": 4.984371837425106e-05,
      "loss": 0.5009,
      "num_input_tokens_seen": 3304368,
      "step": 5040
    },
    {
      "epoch": 2.6441299790356396,
      "grad_norm": 0.6872362494468689,
      "learning_rate": 4.9842439254153106e-05,
      "loss": 0.7796,
      "num_input_tokens_seen": 3307504,
      "step": 5045
    },
    {
      "epoch": 2.646750524109015,
      "grad_norm": 0.4942760765552521,
      "learning_rate": 4.984115493730027e-05,
      "loss": 0.6043,
      "num_input_tokens_seen": 3309840,
      "step": 5050
    },
    {
      "epoch": 2.64937106918239,
      "grad_norm": 0.5607582926750183,
      "learning_rate": 4.983986542396123e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 3313264,
      "step": 5055
    },
    {
      "epoch": 2.651991614255765,
      "grad_norm": 0.6398875713348389,
      "learning_rate": 4.983857071440573e-05,
      "loss": 0.5117,
      "num_input_tokens_seen": 3316368,
      "step": 5060
    },
    {
      "epoch": 2.6546121593291403,
      "grad_norm": 0.6971040964126587,
      "learning_rate": 4.9837270808904616e-05,
      "loss": 0.5599,
      "num_input_tokens_seen": 3319504,
      "step": 5065
    },
    {
      "epoch": 2.6572327044025155,
      "grad_norm": 0.6190070509910583,
      "learning_rate": 4.9835965707729807e-05,
      "loss": 0.5141,
      "num_input_tokens_seen": 3322096,
      "step": 5070
    },
    {
      "epoch": 2.659853249475891,
      "grad_norm": 0.3372151255607605,
      "learning_rate": 4.983465541115431e-05,
      "loss": 0.557,
      "num_input_tokens_seen": 3325072,
      "step": 5075
    },
    {
      "epoch": 2.6624737945492662,
      "grad_norm": 0.3492213189601898,
      "learning_rate": 4.983333991945223e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 3329456,
      "step": 5080
    },
    {
      "epoch": 2.6650943396226414,
      "grad_norm": 0.551700234413147,
      "learning_rate": 4.983201923289875e-05,
      "loss": 0.5536,
      "num_input_tokens_seen": 3332592,
      "step": 5085
    },
    {
      "epoch": 2.667714884696017,
      "grad_norm": 0.3345338702201843,
      "learning_rate": 4.983069335177014e-05,
      "loss": 0.6581,
      "num_input_tokens_seen": 3335152,
      "step": 5090
    },
    {
      "epoch": 2.670335429769392,
      "grad_norm": 0.32373344898223877,
      "learning_rate": 4.982936227634378e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 3338384,
      "step": 5095
    },
    {
      "epoch": 2.6729559748427674,
      "grad_norm": 0.6042876243591309,
      "learning_rate": 4.982802600689809e-05,
      "loss": 0.5168,
      "num_input_tokens_seen": 3341840,
      "step": 5100
    },
    {
      "epoch": 2.6755765199161425,
      "grad_norm": 0.44970089197158813,
      "learning_rate": 4.982668454371261e-05,
      "loss": 0.6465,
      "num_input_tokens_seen": 3345776,
      "step": 5105
    },
    {
      "epoch": 2.6781970649895177,
      "grad_norm": 0.5957599878311157,
      "learning_rate": 4.9825337887067966e-05,
      "loss": 0.412,
      "num_input_tokens_seen": 3348656,
      "step": 5110
    },
    {
      "epoch": 2.680817610062893,
      "grad_norm": 0.4465002417564392,
      "learning_rate": 4.982398603724586e-05,
      "loss": 0.4905,
      "num_input_tokens_seen": 3351344,
      "step": 5115
    },
    {
      "epoch": 2.6834381551362685,
      "grad_norm": 0.27497345209121704,
      "learning_rate": 4.982262899452909e-05,
      "loss": 0.3972,
      "num_input_tokens_seen": 3354192,
      "step": 5120
    },
    {
      "epoch": 2.6860587002096437,
      "grad_norm": 0.5002570748329163,
      "learning_rate": 4.982126675920151e-05,
      "loss": 0.3615,
      "num_input_tokens_seen": 3358960,
      "step": 5125
    },
    {
      "epoch": 2.688679245283019,
      "grad_norm": 0.333095520734787,
      "learning_rate": 4.98198993315481e-05,
      "loss": 0.3675,
      "num_input_tokens_seen": 3361744,
      "step": 5130
    },
    {
      "epoch": 2.691299790356394,
      "grad_norm": 0.47345370054244995,
      "learning_rate": 4.981852671185492e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 3364464,
      "step": 5135
    },
    {
      "epoch": 2.6939203354297696,
      "grad_norm": 0.3117271065711975,
      "learning_rate": 4.9817148900409093e-05,
      "loss": 0.4021,
      "num_input_tokens_seen": 3367728,
      "step": 5140
    },
    {
      "epoch": 2.6965408805031448,
      "grad_norm": 0.5306217670440674,
      "learning_rate": 4.981576589749885e-05,
      "loss": 0.5231,
      "num_input_tokens_seen": 3371056,
      "step": 5145
    },
    {
      "epoch": 2.69916142557652,
      "grad_norm": 1.240605354309082,
      "learning_rate": 4.981437770341348e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 3374256,
      "step": 5150
    },
    {
      "epoch": 2.701781970649895,
      "grad_norm": 0.386744886636734,
      "learning_rate": 4.9812984318443414e-05,
      "loss": 0.5593,
      "num_input_tokens_seen": 3377776,
      "step": 5155
    },
    {
      "epoch": 2.7044025157232703,
      "grad_norm": 0.7372726202011108,
      "learning_rate": 4.98115857428801e-05,
      "loss": 0.7589,
      "num_input_tokens_seen": 3381200,
      "step": 5160
    },
    {
      "epoch": 2.7070230607966455,
      "grad_norm": 0.5692746043205261,
      "learning_rate": 4.9810181977016104e-05,
      "loss": 0.5103,
      "num_input_tokens_seen": 3384048,
      "step": 5165
    },
    {
      "epoch": 2.709643605870021,
      "grad_norm": 0.3897453248500824,
      "learning_rate": 4.9808773021145096e-05,
      "loss": 0.373,
      "num_input_tokens_seen": 3386800,
      "step": 5170
    },
    {
      "epoch": 2.7122641509433962,
      "grad_norm": 0.48945048451423645,
      "learning_rate": 4.98073588755618e-05,
      "loss": 0.5282,
      "num_input_tokens_seen": 3389840,
      "step": 5175
    },
    {
      "epoch": 2.7148846960167714,
      "grad_norm": 0.6534160375595093,
      "learning_rate": 4.980593954056205e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 3393264,
      "step": 5180
    },
    {
      "epoch": 2.717505241090147,
      "grad_norm": 0.4928184151649475,
      "learning_rate": 4.9804515016442754e-05,
      "loss": 0.4958,
      "num_input_tokens_seen": 3396720,
      "step": 5185
    },
    {
      "epoch": 2.720125786163522,
      "grad_norm": 0.3863520324230194,
      "learning_rate": 4.9803085303501896e-05,
      "loss": 0.6211,
      "num_input_tokens_seen": 3399792,
      "step": 5190
    },
    {
      "epoch": 2.7227463312368974,
      "grad_norm": 0.6342717409133911,
      "learning_rate": 4.9801650402038555e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 3402576,
      "step": 5195
    },
    {
      "epoch": 2.7253668763102725,
      "grad_norm": 0.2807915508747101,
      "learning_rate": 4.980021031235291e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 3405648,
      "step": 5200
    },
    {
      "epoch": 2.7279874213836477,
      "grad_norm": 0.421073317527771,
      "learning_rate": 4.9798765034746206e-05,
      "loss": 0.4647,
      "num_input_tokens_seen": 3408816,
      "step": 5205
    },
    {
      "epoch": 2.730607966457023,
      "grad_norm": 0.3467962145805359,
      "learning_rate": 4.979731456952077e-05,
      "loss": 0.4405,
      "num_input_tokens_seen": 3412400,
      "step": 5210
    },
    {
      "epoch": 2.7332285115303985,
      "grad_norm": 0.38976529240608215,
      "learning_rate": 4.9795858916980034e-05,
      "loss": 0.408,
      "num_input_tokens_seen": 3415440,
      "step": 5215
    },
    {
      "epoch": 2.7358490566037736,
      "grad_norm": 0.45022621750831604,
      "learning_rate": 4.9794398077428495e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 3418736,
      "step": 5220
    },
    {
      "epoch": 2.738469601677149,
      "grad_norm": 0.6268121004104614,
      "learning_rate": 4.979293205117175e-05,
      "loss": 0.5019,
      "num_input_tokens_seen": 3421456,
      "step": 5225
    },
    {
      "epoch": 2.741090146750524,
      "grad_norm": 0.6105245351791382,
      "learning_rate": 4.979146083851647e-05,
      "loss": 0.5907,
      "num_input_tokens_seen": 3424432,
      "step": 5230
    },
    {
      "epoch": 2.7437106918238996,
      "grad_norm": 0.3682062029838562,
      "learning_rate": 4.978998443977042e-05,
      "loss": 0.4181,
      "num_input_tokens_seen": 3427504,
      "step": 5235
    },
    {
      "epoch": 2.7463312368972748,
      "grad_norm": 0.5814736485481262,
      "learning_rate": 4.9788502855242436e-05,
      "loss": 0.3625,
      "num_input_tokens_seen": 3431568,
      "step": 5240
    },
    {
      "epoch": 2.74895178197065,
      "grad_norm": 0.4117201864719391,
      "learning_rate": 4.978701608524247e-05,
      "loss": 0.6512,
      "num_input_tokens_seen": 3434736,
      "step": 5245
    },
    {
      "epoch": 2.751572327044025,
      "grad_norm": 0.6644492745399475,
      "learning_rate": 4.978552413008152e-05,
      "loss": 0.4611,
      "num_input_tokens_seen": 3437360,
      "step": 5250
    },
    {
      "epoch": 2.7541928721174003,
      "grad_norm": 0.4836416244506836,
      "learning_rate": 4.9784026990071687e-05,
      "loss": 0.4341,
      "num_input_tokens_seen": 3441008,
      "step": 5255
    },
    {
      "epoch": 2.7568134171907754,
      "grad_norm": 0.24522343277931213,
      "learning_rate": 4.978252466552615e-05,
      "loss": 0.4145,
      "num_input_tokens_seen": 3445072,
      "step": 5260
    },
    {
      "epoch": 2.759433962264151,
      "grad_norm": 0.3489711284637451,
      "learning_rate": 4.9781017156759205e-05,
      "loss": 0.5035,
      "num_input_tokens_seen": 3447952,
      "step": 5265
    },
    {
      "epoch": 2.762054507337526,
      "grad_norm": 0.4423973262310028,
      "learning_rate": 4.977950446408618e-05,
      "loss": 0.4829,
      "num_input_tokens_seen": 3450640,
      "step": 5270
    },
    {
      "epoch": 2.7646750524109014,
      "grad_norm": 0.7296715378761292,
      "learning_rate": 4.9777986587823523e-05,
      "loss": 0.6579,
      "num_input_tokens_seen": 3454256,
      "step": 5275
    },
    {
      "epoch": 2.767295597484277,
      "grad_norm": 1.144197940826416,
      "learning_rate": 4.977646352828875e-05,
      "loss": 0.5688,
      "num_input_tokens_seen": 3457104,
      "step": 5280
    },
    {
      "epoch": 2.769916142557652,
      "grad_norm": 0.54212886095047,
      "learning_rate": 4.9774935285800476e-05,
      "loss": 0.3705,
      "num_input_tokens_seen": 3460016,
      "step": 5285
    },
    {
      "epoch": 2.7725366876310273,
      "grad_norm": 0.3107507526874542,
      "learning_rate": 4.977340186067838e-05,
      "loss": 0.5938,
      "num_input_tokens_seen": 3464464,
      "step": 5290
    },
    {
      "epoch": 2.7751572327044025,
      "grad_norm": 0.4051345884799957,
      "learning_rate": 4.977186325324325e-05,
      "loss": 0.4152,
      "num_input_tokens_seen": 3468048,
      "step": 5295
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 0.4746112525463104,
      "learning_rate": 4.977031946381694e-05,
      "loss": 0.4604,
      "num_input_tokens_seen": 3470736,
      "step": 5300
    },
    {
      "epoch": 2.780398322851153,
      "grad_norm": 0.3573901057243347,
      "learning_rate": 4.9768770492722394e-05,
      "loss": 0.6819,
      "num_input_tokens_seen": 3473424,
      "step": 5305
    },
    {
      "epoch": 2.7830188679245285,
      "grad_norm": 0.7069255709648132,
      "learning_rate": 4.976721634028363e-05,
      "loss": 0.4475,
      "num_input_tokens_seen": 3476816,
      "step": 5310
    },
    {
      "epoch": 2.7856394129979036,
      "grad_norm": 0.4593561887741089,
      "learning_rate": 4.976565700682578e-05,
      "loss": 0.6204,
      "num_input_tokens_seen": 3480016,
      "step": 5315
    },
    {
      "epoch": 2.788259958071279,
      "grad_norm": 0.5655591487884521,
      "learning_rate": 4.976409249267502e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 3483824,
      "step": 5320
    },
    {
      "epoch": 2.790880503144654,
      "grad_norm": 0.5075322389602661,
      "learning_rate": 4.976252279815863e-05,
      "loss": 0.4364,
      "num_input_tokens_seen": 3486416,
      "step": 5325
    },
    {
      "epoch": 2.7935010482180296,
      "grad_norm": 0.4033028483390808,
      "learning_rate": 4.976094792360498e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 3490384,
      "step": 5330
    },
    {
      "epoch": 2.7961215932914047,
      "grad_norm": 0.3166969418525696,
      "learning_rate": 4.975936786934351e-05,
      "loss": 0.4145,
      "num_input_tokens_seen": 3493136,
      "step": 5335
    },
    {
      "epoch": 2.79874213836478,
      "grad_norm": 0.28374502062797546,
      "learning_rate": 4.975778263570475e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 3496112,
      "step": 5340
    },
    {
      "epoch": 2.801362683438155,
      "grad_norm": 1.293016791343689,
      "learning_rate": 4.975619222302032e-05,
      "loss": 0.7568,
      "num_input_tokens_seen": 3498864,
      "step": 5345
    },
    {
      "epoch": 2.8039832285115303,
      "grad_norm": 0.6935149431228638,
      "learning_rate": 4.975459663162291e-05,
      "loss": 0.4241,
      "num_input_tokens_seen": 3502224,
      "step": 5350
    },
    {
      "epoch": 2.8066037735849054,
      "grad_norm": 0.5603355169296265,
      "learning_rate": 4.9752995861846296e-05,
      "loss": 0.4635,
      "num_input_tokens_seen": 3505488,
      "step": 5355
    },
    {
      "epoch": 2.809224318658281,
      "grad_norm": 0.45663753151893616,
      "learning_rate": 4.975138991402535e-05,
      "loss": 0.4488,
      "num_input_tokens_seen": 3508560,
      "step": 5360
    },
    {
      "epoch": 2.811844863731656,
      "grad_norm": 0.2648579776287079,
      "learning_rate": 4.974977878849601e-05,
      "loss": 0.385,
      "num_input_tokens_seen": 3512944,
      "step": 5365
    },
    {
      "epoch": 2.8144654088050314,
      "grad_norm": 0.3401930332183838,
      "learning_rate": 4.9748162485595303e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 3517232,
      "step": 5370
    },
    {
      "epoch": 2.8170859538784065,
      "grad_norm": 0.5147987008094788,
      "learning_rate": 4.974654100566135e-05,
      "loss": 0.5403,
      "num_input_tokens_seen": 3520304,
      "step": 5375
    },
    {
      "epoch": 2.819706498951782,
      "grad_norm": 0.9584813117980957,
      "learning_rate": 4.974491434903335e-05,
      "loss": 0.7124,
      "num_input_tokens_seen": 3523472,
      "step": 5380
    },
    {
      "epoch": 2.8223270440251573,
      "grad_norm": 0.6584447622299194,
      "learning_rate": 4.9743282516051556e-05,
      "loss": 0.53,
      "num_input_tokens_seen": 3528336,
      "step": 5385
    },
    {
      "epoch": 2.8249475890985325,
      "grad_norm": 0.5147920250892639,
      "learning_rate": 4.974164550705736e-05,
      "loss": 0.4305,
      "num_input_tokens_seen": 3533680,
      "step": 5390
    },
    {
      "epoch": 2.8275681341719077,
      "grad_norm": 0.49531981348991394,
      "learning_rate": 4.974000332239319e-05,
      "loss": 0.5109,
      "num_input_tokens_seen": 3536656,
      "step": 5395
    },
    {
      "epoch": 2.830188679245283,
      "grad_norm": 0.5558215379714966,
      "learning_rate": 4.973835596240257e-05,
      "loss": 0.6278,
      "num_input_tokens_seen": 3539280,
      "step": 5400
    },
    {
      "epoch": 2.832809224318658,
      "grad_norm": 0.5002131462097168,
      "learning_rate": 4.9736703427430114e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 3542000,
      "step": 5405
    },
    {
      "epoch": 2.8354297693920336,
      "grad_norm": 0.4754027724266052,
      "learning_rate": 4.9735045717821514e-05,
      "loss": 0.4702,
      "num_input_tokens_seen": 3544624,
      "step": 5410
    },
    {
      "epoch": 2.838050314465409,
      "grad_norm": 0.2588275969028473,
      "learning_rate": 4.973338283392354e-05,
      "loss": 0.3867,
      "num_input_tokens_seen": 3547568,
      "step": 5415
    },
    {
      "epoch": 2.840670859538784,
      "grad_norm": 0.3981804847717285,
      "learning_rate": 4.9731714776084046e-05,
      "loss": 0.5885,
      "num_input_tokens_seen": 3550992,
      "step": 5420
    },
    {
      "epoch": 2.8432914046121596,
      "grad_norm": 0.33010929822921753,
      "learning_rate": 4.973004154465198e-05,
      "loss": 0.5036,
      "num_input_tokens_seen": 3554896,
      "step": 5425
    },
    {
      "epoch": 2.8459119496855347,
      "grad_norm": 0.5656936168670654,
      "learning_rate": 4.972836313997736e-05,
      "loss": 0.5268,
      "num_input_tokens_seen": 3557744,
      "step": 5430
    },
    {
      "epoch": 2.84853249475891,
      "grad_norm": 0.35625457763671875,
      "learning_rate": 4.9726679562411285e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 3561136,
      "step": 5435
    },
    {
      "epoch": 2.851153039832285,
      "grad_norm": 0.43030405044555664,
      "learning_rate": 4.972499081230595e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 3563984,
      "step": 5440
    },
    {
      "epoch": 2.8537735849056602,
      "grad_norm": 0.7148581147193909,
      "learning_rate": 4.9723296890014605e-05,
      "loss": 0.4175,
      "num_input_tokens_seen": 3567184,
      "step": 5445
    },
    {
      "epoch": 2.8563941299790354,
      "grad_norm": 0.4049590528011322,
      "learning_rate": 4.9721597795891605e-05,
      "loss": 0.4432,
      "num_input_tokens_seen": 3571120,
      "step": 5450
    },
    {
      "epoch": 2.859014675052411,
      "grad_norm": 0.5626150369644165,
      "learning_rate": 4.9719893530292396e-05,
      "loss": 0.4334,
      "num_input_tokens_seen": 3574000,
      "step": 5455
    },
    {
      "epoch": 2.861635220125786,
      "grad_norm": 0.6836811900138855,
      "learning_rate": 4.9718184093573475e-05,
      "loss": 0.641,
      "num_input_tokens_seen": 3577296,
      "step": 5460
    },
    {
      "epoch": 2.8642557651991614,
      "grad_norm": 0.5622202754020691,
      "learning_rate": 4.9716469486092435e-05,
      "loss": 0.59,
      "num_input_tokens_seen": 3580368,
      "step": 5465
    },
    {
      "epoch": 2.8668763102725365,
      "grad_norm": 0.8169687986373901,
      "learning_rate": 4.971474970820797e-05,
      "loss": 0.3653,
      "num_input_tokens_seen": 3583984,
      "step": 5470
    },
    {
      "epoch": 2.869496855345912,
      "grad_norm": 0.27051714062690735,
      "learning_rate": 4.971302476027982e-05,
      "loss": 0.4794,
      "num_input_tokens_seen": 3588048,
      "step": 5475
    },
    {
      "epoch": 2.8721174004192873,
      "grad_norm": 0.7724807858467102,
      "learning_rate": 4.971129464266884e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 3590576,
      "step": 5480
    },
    {
      "epoch": 2.8747379454926625,
      "grad_norm": 0.543565571308136,
      "learning_rate": 4.970955935573693e-05,
      "loss": 0.5996,
      "num_input_tokens_seen": 3593072,
      "step": 5485
    },
    {
      "epoch": 2.8773584905660377,
      "grad_norm": 1.0215004682540894,
      "learning_rate": 4.9707818899847106e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 3596144,
      "step": 5490
    },
    {
      "epoch": 2.879979035639413,
      "grad_norm": 0.3643256425857544,
      "learning_rate": 4.970607327536346e-05,
      "loss": 0.5812,
      "num_input_tokens_seen": 3599952,
      "step": 5495
    },
    {
      "epoch": 2.882599580712788,
      "grad_norm": 0.5330453515052795,
      "learning_rate": 4.970432248265114e-05,
      "loss": 0.5552,
      "num_input_tokens_seen": 3603152,
      "step": 5500
    },
    {
      "epoch": 2.8852201257861636,
      "grad_norm": 0.31486183404922485,
      "learning_rate": 4.97025665220764e-05,
      "loss": 0.5177,
      "num_input_tokens_seen": 3607504,
      "step": 5505
    },
    {
      "epoch": 2.8878406708595388,
      "grad_norm": 0.4546512961387634,
      "learning_rate": 4.970080539400656e-05,
      "loss": 0.412,
      "num_input_tokens_seen": 3611440,
      "step": 5510
    },
    {
      "epoch": 2.890461215932914,
      "grad_norm": 0.8672499060630798,
      "learning_rate": 4.969903909881004e-05,
      "loss": 0.507,
      "num_input_tokens_seen": 3614128,
      "step": 5515
    },
    {
      "epoch": 2.8930817610062896,
      "grad_norm": 0.4169726073741913,
      "learning_rate": 4.9697267636856317e-05,
      "loss": 0.4056,
      "num_input_tokens_seen": 3618256,
      "step": 5520
    },
    {
      "epoch": 2.8957023060796647,
      "grad_norm": 0.5929568409919739,
      "learning_rate": 4.969549100851597e-05,
      "loss": 0.5166,
      "num_input_tokens_seen": 3621360,
      "step": 5525
    },
    {
      "epoch": 2.89832285115304,
      "grad_norm": 0.34469839930534363,
      "learning_rate": 4.9693709214160645e-05,
      "loss": 0.3811,
      "num_input_tokens_seen": 3624432,
      "step": 5530
    },
    {
      "epoch": 2.900943396226415,
      "grad_norm": 0.41709500551223755,
      "learning_rate": 4.969192225416307e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 3626800,
      "step": 5535
    },
    {
      "epoch": 2.9035639412997902,
      "grad_norm": 0.6022326350212097,
      "learning_rate": 4.969013012889706e-05,
      "loss": 0.5862,
      "num_input_tokens_seen": 3629424,
      "step": 5540
    },
    {
      "epoch": 2.9061844863731654,
      "grad_norm": 0.30819904804229736,
      "learning_rate": 4.9688332838737504e-05,
      "loss": 0.6043,
      "num_input_tokens_seen": 3632496,
      "step": 5545
    },
    {
      "epoch": 2.908805031446541,
      "grad_norm": 0.25410744547843933,
      "learning_rate": 4.9686530384060386e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 3635568,
      "step": 5550
    },
    {
      "epoch": 2.911425576519916,
      "grad_norm": 0.530881404876709,
      "learning_rate": 4.968472276524275e-05,
      "loss": 0.5744,
      "num_input_tokens_seen": 3638704,
      "step": 5555
    },
    {
      "epoch": 2.9140461215932913,
      "grad_norm": 0.5055753588676453,
      "learning_rate": 4.9682909982662725e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 3641904,
      "step": 5560
    },
    {
      "epoch": 2.9166666666666665,
      "grad_norm": 0.6861050724983215,
      "learning_rate": 4.968109203669953e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 3645296,
      "step": 5565
    },
    {
      "epoch": 2.919287211740042,
      "grad_norm": 0.3533729016780853,
      "learning_rate": 4.967926892773346e-05,
      "loss": 0.6347,
      "num_input_tokens_seen": 3648400,
      "step": 5570
    },
    {
      "epoch": 2.9219077568134173,
      "grad_norm": 0.6892625689506531,
      "learning_rate": 4.967744065614588e-05,
      "loss": 0.4417,
      "num_input_tokens_seen": 3651600,
      "step": 5575
    },
    {
      "epoch": 2.9245283018867925,
      "grad_norm": 0.3260384798049927,
      "learning_rate": 4.9675607222319255e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 3655280,
      "step": 5580
    },
    {
      "epoch": 2.9271488469601676,
      "grad_norm": 0.3726431429386139,
      "learning_rate": 4.967376862663711e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 3657584,
      "step": 5585
    },
    {
      "epoch": 2.929769392033543,
      "grad_norm": 0.3583603501319885,
      "learning_rate": 4.9671924869484064e-05,
      "loss": 0.4008,
      "num_input_tokens_seen": 3660592,
      "step": 5590
    },
    {
      "epoch": 2.932389937106918,
      "grad_norm": 0.7880954146385193,
      "learning_rate": 4.967007595124581e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 3663120,
      "step": 5595
    },
    {
      "epoch": 2.9350104821802936,
      "grad_norm": 0.7594504356384277,
      "learning_rate": 4.966822187230911e-05,
      "loss": 0.4074,
      "num_input_tokens_seen": 3665456,
      "step": 5600
    },
    {
      "epoch": 2.9376310272536688,
      "grad_norm": 0.943690836429596,
      "learning_rate": 4.966636263306183e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 3668080,
      "step": 5605
    },
    {
      "epoch": 2.940251572327044,
      "grad_norm": 0.7130374312400818,
      "learning_rate": 4.9664498233892895e-05,
      "loss": 0.4737,
      "num_input_tokens_seen": 3670672,
      "step": 5610
    },
    {
      "epoch": 2.9428721174004195,
      "grad_norm": 0.4646592438220978,
      "learning_rate": 4.966262867519232e-05,
      "loss": 0.5136,
      "num_input_tokens_seen": 3675312,
      "step": 5615
    },
    {
      "epoch": 2.9454926624737947,
      "grad_norm": 0.3857830762863159,
      "learning_rate": 4.966075395735119e-05,
      "loss": 0.3583,
      "num_input_tokens_seen": 3678864,
      "step": 5620
    },
    {
      "epoch": 2.94811320754717,
      "grad_norm": 0.778363049030304,
      "learning_rate": 4.965887408076167e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 3681808,
      "step": 5625
    },
    {
      "epoch": 2.950733752620545,
      "grad_norm": 0.5290806293487549,
      "learning_rate": 4.965698904581703e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 3684720,
      "step": 5630
    },
    {
      "epoch": 2.95335429769392,
      "grad_norm": 0.37082338333129883,
      "learning_rate": 4.965509885291157e-05,
      "loss": 0.5456,
      "num_input_tokens_seen": 3687920,
      "step": 5635
    },
    {
      "epoch": 2.9559748427672954,
      "grad_norm": 0.31015723943710327,
      "learning_rate": 4.9653203502440716e-05,
      "loss": 0.5191,
      "num_input_tokens_seen": 3691152,
      "step": 5640
    },
    {
      "epoch": 2.958595387840671,
      "grad_norm": 0.583585262298584,
      "learning_rate": 4.965130299480095e-05,
      "loss": 0.5591,
      "num_input_tokens_seen": 3694000,
      "step": 5645
    },
    {
      "epoch": 2.961215932914046,
      "grad_norm": 0.5950997471809387,
      "learning_rate": 4.9649397330389834e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 3696816,
      "step": 5650
    },
    {
      "epoch": 2.9638364779874213,
      "grad_norm": 0.5448876619338989,
      "learning_rate": 4.964748650960601e-05,
      "loss": 0.4978,
      "num_input_tokens_seen": 3699184,
      "step": 5655
    },
    {
      "epoch": 2.9664570230607965,
      "grad_norm": 0.5217930674552917,
      "learning_rate": 4.96455705328492e-05,
      "loss": 0.5827,
      "num_input_tokens_seen": 3701552,
      "step": 5660
    },
    {
      "epoch": 2.969077568134172,
      "grad_norm": 0.3835771381855011,
      "learning_rate": 4.964364940052021e-05,
      "loss": 0.5043,
      "num_input_tokens_seen": 3704784,
      "step": 5665
    },
    {
      "epoch": 2.9716981132075473,
      "grad_norm": 0.4788651466369629,
      "learning_rate": 4.964172311302091e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 3707760,
      "step": 5670
    },
    {
      "epoch": 2.9743186582809225,
      "grad_norm": 0.38735780119895935,
      "learning_rate": 4.9639791670754265e-05,
      "loss": 0.5465,
      "num_input_tokens_seen": 3711536,
      "step": 5675
    },
    {
      "epoch": 2.9769392033542976,
      "grad_norm": 0.344823956489563,
      "learning_rate": 4.9637855074124305e-05,
      "loss": 0.3912,
      "num_input_tokens_seen": 3714224,
      "step": 5680
    },
    {
      "epoch": 2.979559748427673,
      "grad_norm": 0.47083890438079834,
      "learning_rate": 4.963591332353615e-05,
      "loss": 0.4547,
      "num_input_tokens_seen": 3717264,
      "step": 5685
    },
    {
      "epoch": 2.982180293501048,
      "grad_norm": 0.3319435715675354,
      "learning_rate": 4.9633966419396e-05,
      "loss": 0.5576,
      "num_input_tokens_seen": 3720496,
      "step": 5690
    },
    {
      "epoch": 2.9848008385744236,
      "grad_norm": 0.5389329791069031,
      "learning_rate": 4.96320143621111e-05,
      "loss": 0.5045,
      "num_input_tokens_seen": 3723280,
      "step": 5695
    },
    {
      "epoch": 2.9874213836477987,
      "grad_norm": 0.29538625478744507,
      "learning_rate": 4.963005715208981e-05,
      "loss": 0.6917,
      "num_input_tokens_seen": 3727888,
      "step": 5700
    },
    {
      "epoch": 2.990041928721174,
      "grad_norm": 0.4038738012313843,
      "learning_rate": 4.962809478974157e-05,
      "loss": 0.4317,
      "num_input_tokens_seen": 3731248,
      "step": 5705
    },
    {
      "epoch": 2.9926624737945495,
      "grad_norm": 0.7060444951057434,
      "learning_rate": 4.962612727547686e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 3734032,
      "step": 5710
    },
    {
      "epoch": 2.9952830188679247,
      "grad_norm": 0.9082356095314026,
      "learning_rate": 4.962415460970728e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 3737872,
      "step": 5715
    },
    {
      "epoch": 2.9979035639413,
      "grad_norm": 0.3306988775730133,
      "learning_rate": 4.962217679284548e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 3741360,
      "step": 5720
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.4709044396877289,
      "eval_runtime": 9.7147,
      "eval_samples_per_second": 87.29,
      "eval_steps_per_second": 21.823,
      "num_input_tokens_seen": 3743760,
      "step": 5724
    },
    {
      "epoch": 3.000524109014675,
      "grad_norm": 0.43108952045440674,
      "learning_rate": 4.962019382530521e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 3744240,
      "step": 5725
    },
    {
      "epoch": 3.00314465408805,
      "grad_norm": 0.6587929129600525,
      "learning_rate": 4.9618205707501256e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 3746832,
      "step": 5730
    },
    {
      "epoch": 3.0057651991614254,
      "grad_norm": 0.4679100811481476,
      "learning_rate": 4.961621243984954e-05,
      "loss": 0.4397,
      "num_input_tokens_seen": 3750608,
      "step": 5735
    },
    {
      "epoch": 3.008385744234801,
      "grad_norm": 1.4770194292068481,
      "learning_rate": 4.9614214022767014e-05,
      "loss": 0.477,
      "num_input_tokens_seen": 3753360,
      "step": 5740
    },
    {
      "epoch": 3.011006289308176,
      "grad_norm": 0.5402474999427795,
      "learning_rate": 4.9612210456671735e-05,
      "loss": 0.5138,
      "num_input_tokens_seen": 3755824,
      "step": 5745
    },
    {
      "epoch": 3.0136268343815513,
      "grad_norm": 0.44007086753845215,
      "learning_rate": 4.961020174198281e-05,
      "loss": 0.387,
      "num_input_tokens_seen": 3758640,
      "step": 5750
    },
    {
      "epoch": 3.0162473794549265,
      "grad_norm": 0.5771164298057556,
      "learning_rate": 4.9608187879120454e-05,
      "loss": 0.4876,
      "num_input_tokens_seen": 3761712,
      "step": 5755
    },
    {
      "epoch": 3.018867924528302,
      "grad_norm": 0.8602034449577332,
      "learning_rate": 4.9606168868505936e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 3764368,
      "step": 5760
    },
    {
      "epoch": 3.0214884696016773,
      "grad_norm": 0.276877760887146,
      "learning_rate": 4.9604144710561616e-05,
      "loss": 0.5899,
      "num_input_tokens_seen": 3768112,
      "step": 5765
    },
    {
      "epoch": 3.0241090146750524,
      "grad_norm": 0.283273309469223,
      "learning_rate": 4.960211540571092e-05,
      "loss": 0.4974,
      "num_input_tokens_seen": 3773200,
      "step": 5770
    },
    {
      "epoch": 3.0267295597484276,
      "grad_norm": 0.45980069041252136,
      "learning_rate": 4.960008095437836e-05,
      "loss": 0.6096,
      "num_input_tokens_seen": 3777360,
      "step": 5775
    },
    {
      "epoch": 3.029350104821803,
      "grad_norm": 0.7385683059692383,
      "learning_rate": 4.959804135698951e-05,
      "loss": 0.6053,
      "num_input_tokens_seen": 3780240,
      "step": 5780
    },
    {
      "epoch": 3.0319706498951784,
      "grad_norm": 0.5284780859947205,
      "learning_rate": 4.959599661397104e-05,
      "loss": 0.4404,
      "num_input_tokens_seen": 3784528,
      "step": 5785
    },
    {
      "epoch": 3.0345911949685536,
      "grad_norm": 0.32872289419174194,
      "learning_rate": 4.959394672575068e-05,
      "loss": 0.538,
      "num_input_tokens_seen": 3787376,
      "step": 5790
    },
    {
      "epoch": 3.0372117400419287,
      "grad_norm": 0.27115193009376526,
      "learning_rate": 4.959189169275725e-05,
      "loss": 0.4677,
      "num_input_tokens_seen": 3790512,
      "step": 5795
    },
    {
      "epoch": 3.039832285115304,
      "grad_norm": 0.5262227058410645,
      "learning_rate": 4.958983151542065e-05,
      "loss": 0.49,
      "num_input_tokens_seen": 3793744,
      "step": 5800
    },
    {
      "epoch": 3.042452830188679,
      "grad_norm": 0.41249018907546997,
      "learning_rate": 4.958776619417182e-05,
      "loss": 0.5455,
      "num_input_tokens_seen": 3796560,
      "step": 5805
    },
    {
      "epoch": 3.0450733752620547,
      "grad_norm": 0.39022499322891235,
      "learning_rate": 4.9585695729442814e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 3799440,
      "step": 5810
    },
    {
      "epoch": 3.04769392033543,
      "grad_norm": 0.31195321679115295,
      "learning_rate": 4.9583620121666754e-05,
      "loss": 0.5462,
      "num_input_tokens_seen": 3802320,
      "step": 5815
    },
    {
      "epoch": 3.050314465408805,
      "grad_norm": 0.4019836485385895,
      "learning_rate": 4.958153937127783e-05,
      "loss": 0.4703,
      "num_input_tokens_seen": 3805552,
      "step": 5820
    },
    {
      "epoch": 3.05293501048218,
      "grad_norm": 0.614571750164032,
      "learning_rate": 4.95794534787113e-05,
      "loss": 0.5062,
      "num_input_tokens_seen": 3808688,
      "step": 5825
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 0.3387293815612793,
      "learning_rate": 4.957736244440353e-05,
      "loss": 0.4303,
      "num_input_tokens_seen": 3811952,
      "step": 5830
    },
    {
      "epoch": 3.058176100628931,
      "grad_norm": 0.474707692861557,
      "learning_rate": 4.957526626879193e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 3818960,
      "step": 5835
    },
    {
      "epoch": 3.060796645702306,
      "grad_norm": 0.37554284930229187,
      "learning_rate": 4.957316495231499e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 3822384,
      "step": 5840
    },
    {
      "epoch": 3.0634171907756813,
      "grad_norm": 0.3591671288013458,
      "learning_rate": 4.9571058495412294e-05,
      "loss": 0.5618,
      "num_input_tokens_seen": 3825136,
      "step": 5845
    },
    {
      "epoch": 3.0660377358490565,
      "grad_norm": 0.3522745370864868,
      "learning_rate": 4.9568946898524485e-05,
      "loss": 0.4671,
      "num_input_tokens_seen": 3829680,
      "step": 5850
    },
    {
      "epoch": 3.068658280922432,
      "grad_norm": 0.6631748080253601,
      "learning_rate": 4.956683016209327e-05,
      "loss": 0.4526,
      "num_input_tokens_seen": 3832688,
      "step": 5855
    },
    {
      "epoch": 3.0712788259958073,
      "grad_norm": 0.47947850823402405,
      "learning_rate": 4.956470828656147e-05,
      "loss": 0.467,
      "num_input_tokens_seen": 3835472,
      "step": 5860
    },
    {
      "epoch": 3.0738993710691824,
      "grad_norm": 0.4139750599861145,
      "learning_rate": 4.956258127237294e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 3838192,
      "step": 5865
    },
    {
      "epoch": 3.0765199161425576,
      "grad_norm": 0.7554381489753723,
      "learning_rate": 4.956044911997263e-05,
      "loss": 0.5496,
      "num_input_tokens_seen": 3841584,
      "step": 5870
    },
    {
      "epoch": 3.0791404612159328,
      "grad_norm": 0.3339274227619171,
      "learning_rate": 4.955831182980657e-05,
      "loss": 0.4787,
      "num_input_tokens_seen": 3844400,
      "step": 5875
    },
    {
      "epoch": 3.0817610062893084,
      "grad_norm": 0.07432769984006882,
      "learning_rate": 4.955616940232185e-05,
      "loss": 0.3784,
      "num_input_tokens_seen": 3851632,
      "step": 5880
    },
    {
      "epoch": 3.0843815513626835,
      "grad_norm": 0.5795523524284363,
      "learning_rate": 4.9554021837966636e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 3854448,
      "step": 5885
    },
    {
      "epoch": 3.0870020964360587,
      "grad_norm": 0.29636767506599426,
      "learning_rate": 4.955186913719019e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 3858128,
      "step": 5890
    },
    {
      "epoch": 3.089622641509434,
      "grad_norm": 0.45257845520973206,
      "learning_rate": 4.954971130044281e-05,
      "loss": 0.4552,
      "num_input_tokens_seen": 3861104,
      "step": 5895
    },
    {
      "epoch": 3.092243186582809,
      "grad_norm": 0.3638854920864105,
      "learning_rate": 4.954754832817592e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 3864368,
      "step": 5900
    },
    {
      "epoch": 3.0948637316561847,
      "grad_norm": 0.34449639916419983,
      "learning_rate": 4.954538022084196e-05,
      "loss": 0.5087,
      "num_input_tokens_seen": 3867184,
      "step": 5905
    },
    {
      "epoch": 3.09748427672956,
      "grad_norm": 0.3772208094596863,
      "learning_rate": 4.9543206978894496e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 3869776,
      "step": 5910
    },
    {
      "epoch": 3.100104821802935,
      "grad_norm": 1.637475609779358,
      "learning_rate": 4.954102860278813e-05,
      "loss": 0.6497,
      "num_input_tokens_seen": 3871952,
      "step": 5915
    },
    {
      "epoch": 3.10272536687631,
      "grad_norm": 0.28185924887657166,
      "learning_rate": 4.9538845092978567e-05,
      "loss": 0.4536,
      "num_input_tokens_seen": 3875408,
      "step": 5920
    },
    {
      "epoch": 3.1053459119496853,
      "grad_norm": 0.7172935605049133,
      "learning_rate": 4.953665644992256e-05,
      "loss": 0.661,
      "num_input_tokens_seen": 3879664,
      "step": 5925
    },
    {
      "epoch": 3.107966457023061,
      "grad_norm": 0.495191752910614,
      "learning_rate": 4.953446267407795e-05,
      "loss": 0.6323,
      "num_input_tokens_seen": 3882928,
      "step": 5930
    },
    {
      "epoch": 3.110587002096436,
      "grad_norm": 0.3870283365249634,
      "learning_rate": 4.953226376590365e-05,
      "loss": 0.5696,
      "num_input_tokens_seen": 3886288,
      "step": 5935
    },
    {
      "epoch": 3.1132075471698113,
      "grad_norm": 0.4639912545681,
      "learning_rate": 4.953005972585965e-05,
      "loss": 0.6402,
      "num_input_tokens_seen": 3889232,
      "step": 5940
    },
    {
      "epoch": 3.1158280922431865,
      "grad_norm": 0.6032627820968628,
      "learning_rate": 4.952785055440701e-05,
      "loss": 0.5228,
      "num_input_tokens_seen": 3892528,
      "step": 5945
    },
    {
      "epoch": 3.1184486373165616,
      "grad_norm": 0.3669591546058655,
      "learning_rate": 4.9525636252007864e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 3894992,
      "step": 5950
    },
    {
      "epoch": 3.1210691823899372,
      "grad_norm": 0.8554599285125732,
      "learning_rate": 4.9523416819125425e-05,
      "loss": 0.4566,
      "num_input_tokens_seen": 3897744,
      "step": 5955
    },
    {
      "epoch": 3.1236897274633124,
      "grad_norm": 0.5545940399169922,
      "learning_rate": 4.952119225622396e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 3899888,
      "step": 5960
    },
    {
      "epoch": 3.1263102725366876,
      "grad_norm": 0.441685289144516,
      "learning_rate": 4.951896256376882e-05,
      "loss": 0.5792,
      "num_input_tokens_seen": 3905904,
      "step": 5965
    },
    {
      "epoch": 3.1289308176100628,
      "grad_norm": 0.5194721817970276,
      "learning_rate": 4.9516727742226444e-05,
      "loss": 0.6101,
      "num_input_tokens_seen": 3908688,
      "step": 5970
    },
    {
      "epoch": 3.131551362683438,
      "grad_norm": 0.5214908123016357,
      "learning_rate": 4.9514487792064323e-05,
      "loss": 0.6059,
      "num_input_tokens_seen": 3912048,
      "step": 5975
    },
    {
      "epoch": 3.1341719077568135,
      "grad_norm": 0.5783547163009644,
      "learning_rate": 4.951224271375104e-05,
      "loss": 0.5443,
      "num_input_tokens_seen": 3915344,
      "step": 5980
    },
    {
      "epoch": 3.1367924528301887,
      "grad_norm": 0.33914414048194885,
      "learning_rate": 4.950999250775623e-05,
      "loss": 0.5705,
      "num_input_tokens_seen": 3918448,
      "step": 5985
    },
    {
      "epoch": 3.139412997903564,
      "grad_norm": 0.8860571980476379,
      "learning_rate": 4.950773717455061e-05,
      "loss": 0.5509,
      "num_input_tokens_seen": 3921040,
      "step": 5990
    },
    {
      "epoch": 3.142033542976939,
      "grad_norm": 0.5168073773384094,
      "learning_rate": 4.950547671460596e-05,
      "loss": 0.5578,
      "num_input_tokens_seen": 3924560,
      "step": 5995
    },
    {
      "epoch": 3.1446540880503147,
      "grad_norm": 0.28080129623413086,
      "learning_rate": 4.950321112839517e-05,
      "loss": 0.4692,
      "num_input_tokens_seen": 3927056,
      "step": 6000
    },
    {
      "epoch": 3.14727463312369,
      "grad_norm": 0.33383798599243164,
      "learning_rate": 4.9500940416392156e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 3930032,
      "step": 6005
    },
    {
      "epoch": 3.149895178197065,
      "grad_norm": 0.44961613416671753,
      "learning_rate": 4.949866457907193e-05,
      "loss": 0.4154,
      "num_input_tokens_seen": 3933488,
      "step": 6010
    },
    {
      "epoch": 3.15251572327044,
      "grad_norm": 0.4563346207141876,
      "learning_rate": 4.9496383616910565e-05,
      "loss": 0.4849,
      "num_input_tokens_seen": 3937488,
      "step": 6015
    },
    {
      "epoch": 3.1551362683438153,
      "grad_norm": 0.33539822697639465,
      "learning_rate": 4.949409753038522e-05,
      "loss": 0.5538,
      "num_input_tokens_seen": 3940528,
      "step": 6020
    },
    {
      "epoch": 3.157756813417191,
      "grad_norm": 0.5164479613304138,
      "learning_rate": 4.949180631997411e-05,
      "loss": 0.4275,
      "num_input_tokens_seen": 3944112,
      "step": 6025
    },
    {
      "epoch": 3.160377358490566,
      "grad_norm": 0.48679351806640625,
      "learning_rate": 4.948950998615654e-05,
      "loss": 0.4884,
      "num_input_tokens_seen": 3947568,
      "step": 6030
    },
    {
      "epoch": 3.1629979035639413,
      "grad_norm": 0.5193427205085754,
      "learning_rate": 4.9487208529412876e-05,
      "loss": 0.4331,
      "num_input_tokens_seen": 3950896,
      "step": 6035
    },
    {
      "epoch": 3.1656184486373165,
      "grad_norm": 0.44517824053764343,
      "learning_rate": 4.9484901950224546e-05,
      "loss": 0.4921,
      "num_input_tokens_seen": 3954256,
      "step": 6040
    },
    {
      "epoch": 3.1682389937106916,
      "grad_norm": 0.5127482414245605,
      "learning_rate": 4.948259024907407e-05,
      "loss": 0.4985,
      "num_input_tokens_seen": 3958640,
      "step": 6045
    },
    {
      "epoch": 3.1708595387840672,
      "grad_norm": 1.0775656700134277,
      "learning_rate": 4.948027342644502e-05,
      "loss": 0.6151,
      "num_input_tokens_seen": 3961072,
      "step": 6050
    },
    {
      "epoch": 3.1734800838574424,
      "grad_norm": 0.7056604027748108,
      "learning_rate": 4.947795148282206e-05,
      "loss": 0.4118,
      "num_input_tokens_seen": 3964816,
      "step": 6055
    },
    {
      "epoch": 3.1761006289308176,
      "grad_norm": 0.3729551434516907,
      "learning_rate": 4.947562441869091e-05,
      "loss": 0.4977,
      "num_input_tokens_seen": 3967824,
      "step": 6060
    },
    {
      "epoch": 3.1787211740041927,
      "grad_norm": 0.6344398856163025,
      "learning_rate": 4.947329223453836e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 3970960,
      "step": 6065
    },
    {
      "epoch": 3.181341719077568,
      "grad_norm": 0.4964798390865326,
      "learning_rate": 4.947095493085229e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 3973872,
      "step": 6070
    },
    {
      "epoch": 3.1839622641509435,
      "grad_norm": 1.6729307174682617,
      "learning_rate": 4.9468612508121616e-05,
      "loss": 0.6022,
      "num_input_tokens_seen": 3976496,
      "step": 6075
    },
    {
      "epoch": 3.1865828092243187,
      "grad_norm": 0.6343591809272766,
      "learning_rate": 4.9466264966836365e-05,
      "loss": 0.4838,
      "num_input_tokens_seen": 3979248,
      "step": 6080
    },
    {
      "epoch": 3.189203354297694,
      "grad_norm": 0.38800325989723206,
      "learning_rate": 4.94639123074876e-05,
      "loss": 0.4955,
      "num_input_tokens_seen": 3986352,
      "step": 6085
    },
    {
      "epoch": 3.191823899371069,
      "grad_norm": 0.4884684681892395,
      "learning_rate": 4.946155453056749e-05,
      "loss": 0.6213,
      "num_input_tokens_seen": 3989808,
      "step": 6090
    },
    {
      "epoch": 3.1944444444444446,
      "grad_norm": 0.4666501581668854,
      "learning_rate": 4.945919163656923e-05,
      "loss": 0.6523,
      "num_input_tokens_seen": 3993104,
      "step": 6095
    },
    {
      "epoch": 3.19706498951782,
      "grad_norm": 0.3167598843574524,
      "learning_rate": 4.945682362598715e-05,
      "loss": 0.4304,
      "num_input_tokens_seen": 3996240,
      "step": 6100
    },
    {
      "epoch": 3.199685534591195,
      "grad_norm": 0.3001266121864319,
      "learning_rate": 4.945445049931656e-05,
      "loss": 0.3457,
      "num_input_tokens_seen": 4000336,
      "step": 6105
    },
    {
      "epoch": 3.20230607966457,
      "grad_norm": 0.382230669260025,
      "learning_rate": 4.945207225705393e-05,
      "loss": 0.4879,
      "num_input_tokens_seen": 4004400,
      "step": 6110
    },
    {
      "epoch": 3.2049266247379453,
      "grad_norm": 0.5815504789352417,
      "learning_rate": 4.944968889969674e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 4007920,
      "step": 6115
    },
    {
      "epoch": 3.207547169811321,
      "grad_norm": 0.6585891246795654,
      "learning_rate": 4.944730042774358e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 4011152,
      "step": 6120
    },
    {
      "epoch": 3.210167714884696,
      "grad_norm": 0.36638930439949036,
      "learning_rate": 4.944490684169408e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 4013968,
      "step": 6125
    },
    {
      "epoch": 3.2127882599580713,
      "grad_norm": 0.5797913074493408,
      "learning_rate": 4.9442508142048945e-05,
      "loss": 0.5478,
      "num_input_tokens_seen": 4017072,
      "step": 6130
    },
    {
      "epoch": 3.2154088050314464,
      "grad_norm": 0.39761075377464294,
      "learning_rate": 4.9440104329309956e-05,
      "loss": 0.5622,
      "num_input_tokens_seen": 4020976,
      "step": 6135
    },
    {
      "epoch": 3.2180293501048216,
      "grad_norm": 0.3144296109676361,
      "learning_rate": 4.943769540397998e-05,
      "loss": 0.5965,
      "num_input_tokens_seen": 4025008,
      "step": 6140
    },
    {
      "epoch": 3.220649895178197,
      "grad_norm": 0.28093042969703674,
      "learning_rate": 4.943528136656292e-05,
      "loss": 0.4623,
      "num_input_tokens_seen": 4029744,
      "step": 6145
    },
    {
      "epoch": 3.2232704402515724,
      "grad_norm": 0.5630958080291748,
      "learning_rate": 4.943286221756377e-05,
      "loss": 0.3961,
      "num_input_tokens_seen": 4033904,
      "step": 6150
    },
    {
      "epoch": 3.2258909853249476,
      "grad_norm": 0.44024795293807983,
      "learning_rate": 4.9430437957488586e-05,
      "loss": 0.4993,
      "num_input_tokens_seen": 4037392,
      "step": 6155
    },
    {
      "epoch": 3.2285115303983227,
      "grad_norm": 1.6257507801055908,
      "learning_rate": 4.94280085868445e-05,
      "loss": 0.6336,
      "num_input_tokens_seen": 4040336,
      "step": 6160
    },
    {
      "epoch": 3.231132075471698,
      "grad_norm": 0.4415695071220398,
      "learning_rate": 4.942557410613972e-05,
      "loss": 0.5545,
      "num_input_tokens_seen": 4044016,
      "step": 6165
    },
    {
      "epoch": 3.2337526205450735,
      "grad_norm": 0.5560727715492249,
      "learning_rate": 4.942313451588348e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 4046832,
      "step": 6170
    },
    {
      "epoch": 3.2363731656184487,
      "grad_norm": 1.0868229866027832,
      "learning_rate": 4.942068981658614e-05,
      "loss": 0.5067,
      "num_input_tokens_seen": 4048816,
      "step": 6175
    },
    {
      "epoch": 3.238993710691824,
      "grad_norm": 0.5430749654769897,
      "learning_rate": 4.9418240008759097e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 4051728,
      "step": 6180
    },
    {
      "epoch": 3.241614255765199,
      "grad_norm": 0.3921974301338196,
      "learning_rate": 4.9415785092914826e-05,
      "loss": 0.3891,
      "num_input_tokens_seen": 4055472,
      "step": 6185
    },
    {
      "epoch": 3.2442348008385746,
      "grad_norm": 0.27931442856788635,
      "learning_rate": 4.941332506956686e-05,
      "loss": 0.5966,
      "num_input_tokens_seen": 4059056,
      "step": 6190
    },
    {
      "epoch": 3.24685534591195,
      "grad_norm": 0.5748060941696167,
      "learning_rate": 4.941085993922981e-05,
      "loss": 0.3888,
      "num_input_tokens_seen": 4061840,
      "step": 6195
    },
    {
      "epoch": 3.249475890985325,
      "grad_norm": 0.49998509883880615,
      "learning_rate": 4.9408389702419354e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 4064656,
      "step": 6200
    },
    {
      "epoch": 3.2520964360587,
      "grad_norm": 0.6555277705192566,
      "learning_rate": 4.940591435965224e-05,
      "loss": 0.648,
      "num_input_tokens_seen": 4068368,
      "step": 6205
    },
    {
      "epoch": 3.2547169811320753,
      "grad_norm": 0.7678993344306946,
      "learning_rate": 4.9403433911446285e-05,
      "loss": 0.557,
      "num_input_tokens_seen": 4071312,
      "step": 6210
    },
    {
      "epoch": 3.257337526205451,
      "grad_norm": 0.5225664973258972,
      "learning_rate": 4.9400948358320356e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 4074736,
      "step": 6215
    },
    {
      "epoch": 3.259958071278826,
      "grad_norm": 0.24036137759685516,
      "learning_rate": 4.939845770079442e-05,
      "loss": 0.4678,
      "num_input_tokens_seen": 4078960,
      "step": 6220
    },
    {
      "epoch": 3.2625786163522013,
      "grad_norm": 0.4253312349319458,
      "learning_rate": 4.9395961939389475e-05,
      "loss": 0.5002,
      "num_input_tokens_seen": 4082512,
      "step": 6225
    },
    {
      "epoch": 3.2651991614255764,
      "grad_norm": 0.45474937558174133,
      "learning_rate": 4.939346107462762e-05,
      "loss": 0.4138,
      "num_input_tokens_seen": 4086352,
      "step": 6230
    },
    {
      "epoch": 3.2678197064989516,
      "grad_norm": 1.392463207244873,
      "learning_rate": 4.9390955107032e-05,
      "loss": 0.6481,
      "num_input_tokens_seen": 4090832,
      "step": 6235
    },
    {
      "epoch": 3.270440251572327,
      "grad_norm": 0.3935515880584717,
      "learning_rate": 4.938844403712685e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 4093936,
      "step": 6240
    },
    {
      "epoch": 3.2730607966457024,
      "grad_norm": 0.5398693084716797,
      "learning_rate": 4.938592786543743e-05,
      "loss": 0.5799,
      "num_input_tokens_seen": 4097136,
      "step": 6245
    },
    {
      "epoch": 3.2756813417190775,
      "grad_norm": 0.6335437893867493,
      "learning_rate": 4.938340659249012e-05,
      "loss": 0.7258,
      "num_input_tokens_seen": 4099632,
      "step": 6250
    },
    {
      "epoch": 3.2783018867924527,
      "grad_norm": 0.5445836782455444,
      "learning_rate": 4.938088021881233e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 4102576,
      "step": 6255
    },
    {
      "epoch": 3.280922431865828,
      "grad_norm": 0.4070099890232086,
      "learning_rate": 4.937834874493255e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 4106928,
      "step": 6260
    },
    {
      "epoch": 3.2835429769392035,
      "grad_norm": 0.6193175911903381,
      "learning_rate": 4.9375812171380334e-05,
      "loss": 0.6016,
      "num_input_tokens_seen": 4111152,
      "step": 6265
    },
    {
      "epoch": 3.2861635220125787,
      "grad_norm": 0.717820942401886,
      "learning_rate": 4.9373270498686306e-05,
      "loss": 0.4723,
      "num_input_tokens_seen": 4113712,
      "step": 6270
    },
    {
      "epoch": 3.288784067085954,
      "grad_norm": 0.5794321298599243,
      "learning_rate": 4.937072372738216e-05,
      "loss": 0.6302,
      "num_input_tokens_seen": 4116944,
      "step": 6275
    },
    {
      "epoch": 3.291404612159329,
      "grad_norm": 0.7136775255203247,
      "learning_rate": 4.9368171858000635e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 4122128,
      "step": 6280
    },
    {
      "epoch": 3.2940251572327046,
      "grad_norm": 0.4348977208137512,
      "learning_rate": 4.9365614891075576e-05,
      "loss": 0.5454,
      "num_input_tokens_seen": 4125584,
      "step": 6285
    },
    {
      "epoch": 3.29664570230608,
      "grad_norm": 0.777754008769989,
      "learning_rate": 4.936305282714185e-05,
      "loss": 0.3696,
      "num_input_tokens_seen": 4128784,
      "step": 6290
    },
    {
      "epoch": 3.299266247379455,
      "grad_norm": 0.6797719597816467,
      "learning_rate": 4.936048566673542e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 4131120,
      "step": 6295
    },
    {
      "epoch": 3.30188679245283,
      "grad_norm": 0.41842347383499146,
      "learning_rate": 4.935791341039332e-05,
      "loss": 0.4508,
      "num_input_tokens_seen": 4133904,
      "step": 6300
    },
    {
      "epoch": 3.3045073375262053,
      "grad_norm": 0.20485709607601166,
      "learning_rate": 4.935533605865361e-05,
      "loss": 0.3903,
      "num_input_tokens_seen": 4137456,
      "step": 6305
    },
    {
      "epoch": 3.307127882599581,
      "grad_norm": 0.6469826698303223,
      "learning_rate": 4.935275361205547e-05,
      "loss": 0.5665,
      "num_input_tokens_seen": 4140272,
      "step": 6310
    },
    {
      "epoch": 3.309748427672956,
      "grad_norm": 0.312051922082901,
      "learning_rate": 4.93501660711391e-05,
      "loss": 0.5033,
      "num_input_tokens_seen": 4143600,
      "step": 6315
    },
    {
      "epoch": 3.3123689727463312,
      "grad_norm": 1.0144439935684204,
      "learning_rate": 4.934757343644578e-05,
      "loss": 0.4179,
      "num_input_tokens_seen": 4146896,
      "step": 6320
    },
    {
      "epoch": 3.3149895178197064,
      "grad_norm": 0.38186120986938477,
      "learning_rate": 4.934497570851788e-05,
      "loss": 0.4877,
      "num_input_tokens_seen": 4150448,
      "step": 6325
    },
    {
      "epoch": 3.3176100628930816,
      "grad_norm": 0.2870626747608185,
      "learning_rate": 4.9342372887898804e-05,
      "loss": 0.3898,
      "num_input_tokens_seen": 4155504,
      "step": 6330
    },
    {
      "epoch": 3.320230607966457,
      "grad_norm": 0.2654648423194885,
      "learning_rate": 4.9339764975133026e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 4158800,
      "step": 6335
    },
    {
      "epoch": 3.3228511530398324,
      "grad_norm": 0.3006167411804199,
      "learning_rate": 4.93371519707661e-05,
      "loss": 0.4067,
      "num_input_tokens_seen": 4162832,
      "step": 6340
    },
    {
      "epoch": 3.3254716981132075,
      "grad_norm": 0.5301170349121094,
      "learning_rate": 4.933453387534464e-05,
      "loss": 0.4693,
      "num_input_tokens_seen": 4166544,
      "step": 6345
    },
    {
      "epoch": 3.3280922431865827,
      "grad_norm": 0.31952911615371704,
      "learning_rate": 4.933191068941631e-05,
      "loss": 0.6627,
      "num_input_tokens_seen": 4170448,
      "step": 6350
    },
    {
      "epoch": 3.330712788259958,
      "grad_norm": 0.49301984906196594,
      "learning_rate": 4.932928241352987e-05,
      "loss": 0.5267,
      "num_input_tokens_seen": 4174160,
      "step": 6355
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 0.5952290892601013,
      "learning_rate": 4.9326649048235095e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 4177584,
      "step": 6360
    },
    {
      "epoch": 3.3359538784067087,
      "grad_norm": 0.5869919061660767,
      "learning_rate": 4.932401059408287e-05,
      "loss": 0.3654,
      "num_input_tokens_seen": 4179472,
      "step": 6365
    },
    {
      "epoch": 3.338574423480084,
      "grad_norm": 0.7980066537857056,
      "learning_rate": 4.932136705162514e-05,
      "loss": 0.5614,
      "num_input_tokens_seen": 4182192,
      "step": 6370
    },
    {
      "epoch": 3.341194968553459,
      "grad_norm": 0.6278151273727417,
      "learning_rate": 4.93187184214149e-05,
      "loss": 0.347,
      "num_input_tokens_seen": 4185040,
      "step": 6375
    },
    {
      "epoch": 3.3438155136268346,
      "grad_norm": 0.5003727674484253,
      "learning_rate": 4.9316064704006205e-05,
      "loss": 0.5159,
      "num_input_tokens_seen": 4187952,
      "step": 6380
    },
    {
      "epoch": 3.3464360587002098,
      "grad_norm": 0.48922908306121826,
      "learning_rate": 4.931340589995419e-05,
      "loss": 0.4982,
      "num_input_tokens_seen": 4191184,
      "step": 6385
    },
    {
      "epoch": 3.349056603773585,
      "grad_norm": 0.6386975049972534,
      "learning_rate": 4.9310742009815035e-05,
      "loss": 0.528,
      "num_input_tokens_seen": 4194064,
      "step": 6390
    },
    {
      "epoch": 3.35167714884696,
      "grad_norm": 0.2856541872024536,
      "learning_rate": 4.930807303414601e-05,
      "loss": 0.3969,
      "num_input_tokens_seen": 4196688,
      "step": 6395
    },
    {
      "epoch": 3.3542976939203353,
      "grad_norm": 0.4211120009422302,
      "learning_rate": 4.930539897350542e-05,
      "loss": 0.3928,
      "num_input_tokens_seen": 4199952,
      "step": 6400
    },
    {
      "epoch": 3.3569182389937104,
      "grad_norm": 0.5172712802886963,
      "learning_rate": 4.930271982845266e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 4203792,
      "step": 6405
    },
    {
      "epoch": 3.359538784067086,
      "grad_norm": 0.49927324056625366,
      "learning_rate": 4.930003559954817e-05,
      "loss": 0.3692,
      "num_input_tokens_seen": 4206736,
      "step": 6410
    },
    {
      "epoch": 3.3621593291404612,
      "grad_norm": 0.41900384426116943,
      "learning_rate": 4.929734628735346e-05,
      "loss": 0.5456,
      "num_input_tokens_seen": 4210032,
      "step": 6415
    },
    {
      "epoch": 3.3647798742138364,
      "grad_norm": 0.5001506209373474,
      "learning_rate": 4.9294651892431115e-05,
      "loss": 0.4769,
      "num_input_tokens_seen": 4212848,
      "step": 6420
    },
    {
      "epoch": 3.3674004192872116,
      "grad_norm": 0.7517035007476807,
      "learning_rate": 4.929195241534476e-05,
      "loss": 0.6098,
      "num_input_tokens_seen": 4215376,
      "step": 6425
    },
    {
      "epoch": 3.370020964360587,
      "grad_norm": 0.5506497025489807,
      "learning_rate": 4.9289247856659095e-05,
      "loss": 0.5388,
      "num_input_tokens_seen": 4218288,
      "step": 6430
    },
    {
      "epoch": 3.3726415094339623,
      "grad_norm": 0.3402002453804016,
      "learning_rate": 4.9286538216939883e-05,
      "loss": 0.4126,
      "num_input_tokens_seen": 4221488,
      "step": 6435
    },
    {
      "epoch": 3.3752620545073375,
      "grad_norm": 0.34918567538261414,
      "learning_rate": 4.9283823496753944e-05,
      "loss": 0.8219,
      "num_input_tokens_seen": 4224432,
      "step": 6440
    },
    {
      "epoch": 3.3778825995807127,
      "grad_norm": 0.41505149006843567,
      "learning_rate": 4.928110369666918e-05,
      "loss": 0.3295,
      "num_input_tokens_seen": 4229296,
      "step": 6445
    },
    {
      "epoch": 3.380503144654088,
      "grad_norm": 0.4988510310649872,
      "learning_rate": 4.927837881725455e-05,
      "loss": 0.5375,
      "num_input_tokens_seen": 4232016,
      "step": 6450
    },
    {
      "epoch": 3.3831236897274635,
      "grad_norm": 0.5229523777961731,
      "learning_rate": 4.927564885908003e-05,
      "loss": 0.3844,
      "num_input_tokens_seen": 4235248,
      "step": 6455
    },
    {
      "epoch": 3.3857442348008386,
      "grad_norm": 0.4721682667732239,
      "learning_rate": 4.9272913822716735e-05,
      "loss": 0.49,
      "num_input_tokens_seen": 4238544,
      "step": 6460
    },
    {
      "epoch": 3.388364779874214,
      "grad_norm": 0.5029228329658508,
      "learning_rate": 4.927017370873678e-05,
      "loss": 0.5362,
      "num_input_tokens_seen": 4241648,
      "step": 6465
    },
    {
      "epoch": 3.390985324947589,
      "grad_norm": 0.24045243859291077,
      "learning_rate": 4.926742851771338e-05,
      "loss": 0.5191,
      "num_input_tokens_seen": 4245520,
      "step": 6470
    },
    {
      "epoch": 3.3936058700209646,
      "grad_norm": 0.6428686380386353,
      "learning_rate": 4.926467825022079e-05,
      "loss": 0.5103,
      "num_input_tokens_seen": 4248400,
      "step": 6475
    },
    {
      "epoch": 3.3962264150943398,
      "grad_norm": 0.39397966861724854,
      "learning_rate": 4.926192290683434e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 4251568,
      "step": 6480
    },
    {
      "epoch": 3.398846960167715,
      "grad_norm": 0.2951471209526062,
      "learning_rate": 4.9259162488130404e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 4255888,
      "step": 6485
    },
    {
      "epoch": 3.40146750524109,
      "grad_norm": 0.5009044408798218,
      "learning_rate": 4.925639699468645e-05,
      "loss": 0.578,
      "num_input_tokens_seen": 4258960,
      "step": 6490
    },
    {
      "epoch": 3.4040880503144653,
      "grad_norm": 0.7604691386222839,
      "learning_rate": 4.925362642708097e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 4262256,
      "step": 6495
    },
    {
      "epoch": 3.4067085953878404,
      "grad_norm": 0.2977413535118103,
      "learning_rate": 4.9250850785893536e-05,
      "loss": 0.4387,
      "num_input_tokens_seen": 4266352,
      "step": 6500
    },
    {
      "epoch": 3.409329140461216,
      "grad_norm": 0.29754966497421265,
      "learning_rate": 4.924807007170479e-05,
      "loss": 0.434,
      "num_input_tokens_seen": 4269136,
      "step": 6505
    },
    {
      "epoch": 3.411949685534591,
      "grad_norm": 0.5129263997077942,
      "learning_rate": 4.924528428509643e-05,
      "loss": 0.4016,
      "num_input_tokens_seen": 4271536,
      "step": 6510
    },
    {
      "epoch": 3.4145702306079664,
      "grad_norm": 0.529309093952179,
      "learning_rate": 4.9242493426651196e-05,
      "loss": 0.4257,
      "num_input_tokens_seen": 4275248,
      "step": 6515
    },
    {
      "epoch": 3.4171907756813416,
      "grad_norm": 0.5157094597816467,
      "learning_rate": 4.92396974969529e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 4278608,
      "step": 6520
    },
    {
      "epoch": 3.419811320754717,
      "grad_norm": 0.4509941637516022,
      "learning_rate": 4.923689649658644e-05,
      "loss": 0.6065,
      "num_input_tokens_seen": 4282064,
      "step": 6525
    },
    {
      "epoch": 3.4224318658280923,
      "grad_norm": 0.5503265261650085,
      "learning_rate": 4.923409042613774e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 4285872,
      "step": 6530
    },
    {
      "epoch": 3.4250524109014675,
      "grad_norm": 0.24954207241535187,
      "learning_rate": 4.92312792861938e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 4288816,
      "step": 6535
    },
    {
      "epoch": 3.4276729559748427,
      "grad_norm": 0.4236598610877991,
      "learning_rate": 4.922846307734268e-05,
      "loss": 0.5125,
      "num_input_tokens_seen": 4292304,
      "step": 6540
    },
    {
      "epoch": 3.430293501048218,
      "grad_norm": 0.3564295172691345,
      "learning_rate": 4.9225641800173496e-05,
      "loss": 0.5027,
      "num_input_tokens_seen": 4295504,
      "step": 6545
    },
    {
      "epoch": 3.4329140461215935,
      "grad_norm": 0.4175662696361542,
      "learning_rate": 4.922281545527643e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 4298160,
      "step": 6550
    },
    {
      "epoch": 3.4355345911949686,
      "grad_norm": 0.6320603489875793,
      "learning_rate": 4.9219984043242715e-05,
      "loss": 0.4667,
      "num_input_tokens_seen": 4300944,
      "step": 6555
    },
    {
      "epoch": 3.438155136268344,
      "grad_norm": 0.8882912993431091,
      "learning_rate": 4.921714756466466e-05,
      "loss": 0.5037,
      "num_input_tokens_seen": 4303600,
      "step": 6560
    },
    {
      "epoch": 3.440775681341719,
      "grad_norm": 0.8784312009811401,
      "learning_rate": 4.921430602013561e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 4306256,
      "step": 6565
    },
    {
      "epoch": 3.4433962264150946,
      "grad_norm": 0.5194002389907837,
      "learning_rate": 4.9211459410249995e-05,
      "loss": 0.4579,
      "num_input_tokens_seen": 4309264,
      "step": 6570
    },
    {
      "epoch": 3.4460167714884697,
      "grad_norm": 0.5544384121894836,
      "learning_rate": 4.92086077356033e-05,
      "loss": 0.5242,
      "num_input_tokens_seen": 4312176,
      "step": 6575
    },
    {
      "epoch": 3.448637316561845,
      "grad_norm": 0.26559990644454956,
      "learning_rate": 4.920575099679204e-05,
      "loss": 0.5683,
      "num_input_tokens_seen": 4317488,
      "step": 6580
    },
    {
      "epoch": 3.45125786163522,
      "grad_norm": 0.47572842240333557,
      "learning_rate": 4.920288919441383e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 4320592,
      "step": 6585
    },
    {
      "epoch": 3.4538784067085953,
      "grad_norm": 0.3974612057209015,
      "learning_rate": 4.9200022329067316e-05,
      "loss": 0.5006,
      "num_input_tokens_seen": 4323632,
      "step": 6590
    },
    {
      "epoch": 3.4564989517819704,
      "grad_norm": 0.3561129868030548,
      "learning_rate": 4.9197150401352224e-05,
      "loss": 0.3428,
      "num_input_tokens_seen": 4326576,
      "step": 6595
    },
    {
      "epoch": 3.459119496855346,
      "grad_norm": 1.0493541955947876,
      "learning_rate": 4.919427341186933e-05,
      "loss": 0.52,
      "num_input_tokens_seen": 4329424,
      "step": 6600
    },
    {
      "epoch": 3.461740041928721,
      "grad_norm": 0.44021275639533997,
      "learning_rate": 4.919139136122045e-05,
      "loss": 0.6727,
      "num_input_tokens_seen": 4332656,
      "step": 6605
    },
    {
      "epoch": 3.4643605870020964,
      "grad_norm": 0.7058243751525879,
      "learning_rate": 4.918850425000849e-05,
      "loss": 0.5368,
      "num_input_tokens_seen": 4335536,
      "step": 6610
    },
    {
      "epoch": 3.4669811320754715,
      "grad_norm": 0.2470504641532898,
      "learning_rate": 4.918561207883741e-05,
      "loss": 0.5319,
      "num_input_tokens_seen": 4339248,
      "step": 6615
    },
    {
      "epoch": 3.469601677148847,
      "grad_norm": 0.3826963007450104,
      "learning_rate": 4.9182714848312195e-05,
      "loss": 0.5069,
      "num_input_tokens_seen": 4342320,
      "step": 6620
    },
    {
      "epoch": 3.4722222222222223,
      "grad_norm": 0.36929649114608765,
      "learning_rate": 4.917981255903893e-05,
      "loss": 0.4784,
      "num_input_tokens_seen": 4346256,
      "step": 6625
    },
    {
      "epoch": 3.4748427672955975,
      "grad_norm": 0.48455795645713806,
      "learning_rate": 4.917690521162474e-05,
      "loss": 0.6004,
      "num_input_tokens_seen": 4349872,
      "step": 6630
    },
    {
      "epoch": 3.4774633123689727,
      "grad_norm": 0.4727185070514679,
      "learning_rate": 4.9173992806677796e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 4352368,
      "step": 6635
    },
    {
      "epoch": 3.480083857442348,
      "grad_norm": 0.35793471336364746,
      "learning_rate": 4.917107534480735e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 4356880,
      "step": 6640
    },
    {
      "epoch": 3.4827044025157234,
      "grad_norm": 0.3068297803401947,
      "learning_rate": 4.916815282662369e-05,
      "loss": 0.3156,
      "num_input_tokens_seen": 4360688,
      "step": 6645
    },
    {
      "epoch": 3.4853249475890986,
      "grad_norm": 0.4395028054714203,
      "learning_rate": 4.91652252527382e-05,
      "loss": 0.4176,
      "num_input_tokens_seen": 4363856,
      "step": 6650
    },
    {
      "epoch": 3.487945492662474,
      "grad_norm": 0.26618823409080505,
      "learning_rate": 4.9162292623763275e-05,
      "loss": 0.4571,
      "num_input_tokens_seen": 4366640,
      "step": 6655
    },
    {
      "epoch": 3.490566037735849,
      "grad_norm": 0.49743029475212097,
      "learning_rate": 4.9159354940312386e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 4369584,
      "step": 6660
    },
    {
      "epoch": 3.4931865828092246,
      "grad_norm": 0.5526586174964905,
      "learning_rate": 4.915641220300007e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 4372752,
      "step": 6665
    },
    {
      "epoch": 3.4958071278825997,
      "grad_norm": 0.40052592754364014,
      "learning_rate": 4.9153464412441915e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 4375760,
      "step": 6670
    },
    {
      "epoch": 3.498427672955975,
      "grad_norm": 0.3742617666721344,
      "learning_rate": 4.915051156925456e-05,
      "loss": 0.4557,
      "num_input_tokens_seen": 4379248,
      "step": 6675
    },
    {
      "epoch": 3.50104821802935,
      "grad_norm": 0.46348631381988525,
      "learning_rate": 4.91475536740557e-05,
      "loss": 0.6743,
      "num_input_tokens_seen": 4382352,
      "step": 6680
    },
    {
      "epoch": 3.5036687631027252,
      "grad_norm": 0.712692379951477,
      "learning_rate": 4.914459072746411e-05,
      "loss": 0.6169,
      "num_input_tokens_seen": 4385968,
      "step": 6685
    },
    {
      "epoch": 3.5062893081761004,
      "grad_norm": 0.5556931495666504,
      "learning_rate": 4.91416227300996e-05,
      "loss": 0.5094,
      "num_input_tokens_seen": 4388720,
      "step": 6690
    },
    {
      "epoch": 3.508909853249476,
      "grad_norm": 0.61551433801651,
      "learning_rate": 4.913864968258303e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 4391568,
      "step": 6695
    },
    {
      "epoch": 3.511530398322851,
      "grad_norm": 0.4296224117279053,
      "learning_rate": 4.913567158553634e-05,
      "loss": 0.5307,
      "num_input_tokens_seen": 4394256,
      "step": 6700
    },
    {
      "epoch": 3.5141509433962264,
      "grad_norm": 0.9371670484542847,
      "learning_rate": 4.9132688439582505e-05,
      "loss": 0.4942,
      "num_input_tokens_seen": 4396816,
      "step": 6705
    },
    {
      "epoch": 3.5167714884696015,
      "grad_norm": 0.8277596235275269,
      "learning_rate": 4.9129700245345565e-05,
      "loss": 0.551,
      "num_input_tokens_seen": 4399952,
      "step": 6710
    },
    {
      "epoch": 3.519392033542977,
      "grad_norm": 0.33561328053474426,
      "learning_rate": 4.912670700345063e-05,
      "loss": 0.4695,
      "num_input_tokens_seen": 4403440,
      "step": 6715
    },
    {
      "epoch": 3.5220125786163523,
      "grad_norm": 0.3037186563014984,
      "learning_rate": 4.912370871452383e-05,
      "loss": 0.5479,
      "num_input_tokens_seen": 4406736,
      "step": 6720
    },
    {
      "epoch": 3.5246331236897275,
      "grad_norm": 0.2943130433559418,
      "learning_rate": 4.91207053791924e-05,
      "loss": 0.4415,
      "num_input_tokens_seen": 4410640,
      "step": 6725
    },
    {
      "epoch": 3.5272536687631026,
      "grad_norm": 0.8162481188774109,
      "learning_rate": 4.911769699808457e-05,
      "loss": 0.4037,
      "num_input_tokens_seen": 4414096,
      "step": 6730
    },
    {
      "epoch": 3.529874213836478,
      "grad_norm": 0.38686513900756836,
      "learning_rate": 4.911468357182969e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 4416976,
      "step": 6735
    },
    {
      "epoch": 3.532494758909853,
      "grad_norm": 0.3028999865055084,
      "learning_rate": 4.9111665101058116e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 4419792,
      "step": 6740
    },
    {
      "epoch": 3.5351153039832286,
      "grad_norm": 0.337155282497406,
      "learning_rate": 4.910864158640128e-05,
      "loss": 0.4795,
      "num_input_tokens_seen": 4422928,
      "step": 6745
    },
    {
      "epoch": 3.5377358490566038,
      "grad_norm": 0.38384896516799927,
      "learning_rate": 4.910561302849168e-05,
      "loss": 0.6309,
      "num_input_tokens_seen": 4426576,
      "step": 6750
    },
    {
      "epoch": 3.540356394129979,
      "grad_norm": 0.36410653591156006,
      "learning_rate": 4.910257942796284e-05,
      "loss": 0.5259,
      "num_input_tokens_seen": 4431408,
      "step": 6755
    },
    {
      "epoch": 3.5429769392033545,
      "grad_norm": 0.5231142640113831,
      "learning_rate": 4.909954078544936e-05,
      "loss": 0.5774,
      "num_input_tokens_seen": 4434352,
      "step": 6760
    },
    {
      "epoch": 3.5455974842767297,
      "grad_norm": 0.41044795513153076,
      "learning_rate": 4.909649710158689e-05,
      "loss": 0.469,
      "num_input_tokens_seen": 4437488,
      "step": 6765
    },
    {
      "epoch": 3.548218029350105,
      "grad_norm": 0.8422642350196838,
      "learning_rate": 4.909344837701214e-05,
      "loss": 0.4076,
      "num_input_tokens_seen": 4440208,
      "step": 6770
    },
    {
      "epoch": 3.55083857442348,
      "grad_norm": 0.5338645577430725,
      "learning_rate": 4.909039461236285e-05,
      "loss": 0.5204,
      "num_input_tokens_seen": 4443024,
      "step": 6775
    },
    {
      "epoch": 3.5534591194968552,
      "grad_norm": 0.5226659774780273,
      "learning_rate": 4.908733580827786e-05,
      "loss": 0.6119,
      "num_input_tokens_seen": 4446512,
      "step": 6780
    },
    {
      "epoch": 3.5560796645702304,
      "grad_norm": 0.5604325532913208,
      "learning_rate": 4.9084271965397014e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 4449360,
      "step": 6785
    },
    {
      "epoch": 3.558700209643606,
      "grad_norm": 0.40304338932037354,
      "learning_rate": 4.908120308436124e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 4451728,
      "step": 6790
    },
    {
      "epoch": 3.561320754716981,
      "grad_norm": 0.3388858437538147,
      "learning_rate": 4.907812916581251e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 4455568,
      "step": 6795
    },
    {
      "epoch": 3.5639412997903563,
      "grad_norm": 0.6296138763427734,
      "learning_rate": 4.907505021039386e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 4458768,
      "step": 6800
    },
    {
      "epoch": 3.5665618448637315,
      "grad_norm": 0.3662056624889374,
      "learning_rate": 4.907196621874937e-05,
      "loss": 0.5451,
      "num_input_tokens_seen": 4461648,
      "step": 6805
    },
    {
      "epoch": 3.569182389937107,
      "grad_norm": 0.46683821082115173,
      "learning_rate": 4.9068877191524176e-05,
      "loss": 0.5256,
      "num_input_tokens_seen": 4465232,
      "step": 6810
    },
    {
      "epoch": 3.5718029350104823,
      "grad_norm": 0.5211235880851746,
      "learning_rate": 4.906578312936446e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 4467984,
      "step": 6815
    },
    {
      "epoch": 3.5744234800838575,
      "grad_norm": 0.43919670581817627,
      "learning_rate": 4.906268403291747e-05,
      "loss": 0.3825,
      "num_input_tokens_seen": 4471088,
      "step": 6820
    },
    {
      "epoch": 3.5770440251572326,
      "grad_norm": 0.539177417755127,
      "learning_rate": 4.905957990283151e-05,
      "loss": 0.582,
      "num_input_tokens_seen": 4474128,
      "step": 6825
    },
    {
      "epoch": 3.579664570230608,
      "grad_norm": 0.4557861387729645,
      "learning_rate": 4.905647073975592e-05,
      "loss": 0.6781,
      "num_input_tokens_seen": 4478160,
      "step": 6830
    },
    {
      "epoch": 3.582285115303983,
      "grad_norm": 0.35791999101638794,
      "learning_rate": 4.9053356544341095e-05,
      "loss": 0.5261,
      "num_input_tokens_seen": 4480784,
      "step": 6835
    },
    {
      "epoch": 3.5849056603773586,
      "grad_norm": 0.5312778353691101,
      "learning_rate": 4.90502373172385e-05,
      "loss": 0.5919,
      "num_input_tokens_seen": 4483632,
      "step": 6840
    },
    {
      "epoch": 3.5875262054507338,
      "grad_norm": 0.3420924246311188,
      "learning_rate": 4.9047113059100635e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 4486448,
      "step": 6845
    },
    {
      "epoch": 3.590146750524109,
      "grad_norm": 0.4663442373275757,
      "learning_rate": 4.904398377058106e-05,
      "loss": 0.3672,
      "num_input_tokens_seen": 4490192,
      "step": 6850
    },
    {
      "epoch": 3.5927672955974845,
      "grad_norm": 0.8085440397262573,
      "learning_rate": 4.90408494523344e-05,
      "loss": 0.5537,
      "num_input_tokens_seen": 4492880,
      "step": 6855
    },
    {
      "epoch": 3.5953878406708597,
      "grad_norm": 0.48681479692459106,
      "learning_rate": 4.90377101050163e-05,
      "loss": 0.463,
      "num_input_tokens_seen": 4495952,
      "step": 6860
    },
    {
      "epoch": 3.598008385744235,
      "grad_norm": 0.3782612085342407,
      "learning_rate": 4.903456572928348e-05,
      "loss": 0.5175,
      "num_input_tokens_seen": 4498512,
      "step": 6865
    },
    {
      "epoch": 3.60062893081761,
      "grad_norm": 0.6349164843559265,
      "learning_rate": 4.903141632579371e-05,
      "loss": 0.4584,
      "num_input_tokens_seen": 4503152,
      "step": 6870
    },
    {
      "epoch": 3.603249475890985,
      "grad_norm": 0.9052577614784241,
      "learning_rate": 4.902826189520581e-05,
      "loss": 0.5205,
      "num_input_tokens_seen": 4506128,
      "step": 6875
    },
    {
      "epoch": 3.6058700209643604,
      "grad_norm": 0.827372133731842,
      "learning_rate": 4.9025102438179656e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 4509776,
      "step": 6880
    },
    {
      "epoch": 3.608490566037736,
      "grad_norm": 0.5885990262031555,
      "learning_rate": 4.902193795537616e-05,
      "loss": 0.5914,
      "num_input_tokens_seen": 4513136,
      "step": 6885
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 0.37415939569473267,
      "learning_rate": 4.90187684474573e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 4516176,
      "step": 6890
    },
    {
      "epoch": 3.6137316561844863,
      "grad_norm": 0.6135234236717224,
      "learning_rate": 4.901559391508611e-05,
      "loss": 0.5432,
      "num_input_tokens_seen": 4519120,
      "step": 6895
    },
    {
      "epoch": 3.6163522012578615,
      "grad_norm": 0.4659941792488098,
      "learning_rate": 4.9012414358926644e-05,
      "loss": 0.444,
      "num_input_tokens_seen": 4522352,
      "step": 6900
    },
    {
      "epoch": 3.618972746331237,
      "grad_norm": 0.34520235657691956,
      "learning_rate": 4.900922977964405e-05,
      "loss": 0.5225,
      "num_input_tokens_seen": 4525104,
      "step": 6905
    },
    {
      "epoch": 3.6215932914046123,
      "grad_norm": 0.37141895294189453,
      "learning_rate": 4.90060401779045e-05,
      "loss": 0.5236,
      "num_input_tokens_seen": 4528528,
      "step": 6910
    },
    {
      "epoch": 3.6242138364779874,
      "grad_norm": 0.4617548882961273,
      "learning_rate": 4.900284555437521e-05,
      "loss": 0.3668,
      "num_input_tokens_seen": 4531312,
      "step": 6915
    },
    {
      "epoch": 3.6268343815513626,
      "grad_norm": 1.1207144260406494,
      "learning_rate": 4.899964590972448e-05,
      "loss": 0.632,
      "num_input_tokens_seen": 4534128,
      "step": 6920
    },
    {
      "epoch": 3.629454926624738,
      "grad_norm": 0.2941242456436157,
      "learning_rate": 4.899644124462163e-05,
      "loss": 0.5523,
      "num_input_tokens_seen": 4537168,
      "step": 6925
    },
    {
      "epoch": 3.632075471698113,
      "grad_norm": 0.543639063835144,
      "learning_rate": 4.8993231559737026e-05,
      "loss": 0.7267,
      "num_input_tokens_seen": 4540464,
      "step": 6930
    },
    {
      "epoch": 3.6346960167714886,
      "grad_norm": 0.3841390907764435,
      "learning_rate": 4.8990016855742114e-05,
      "loss": 0.4357,
      "num_input_tokens_seen": 4544208,
      "step": 6935
    },
    {
      "epoch": 3.6373165618448637,
      "grad_norm": 1.5127538442611694,
      "learning_rate": 4.898679713330937e-05,
      "loss": 0.4159,
      "num_input_tokens_seen": 4547440,
      "step": 6940
    },
    {
      "epoch": 3.639937106918239,
      "grad_norm": 1.3928166627883911,
      "learning_rate": 4.898357239311232e-05,
      "loss": 0.4871,
      "num_input_tokens_seen": 4550512,
      "step": 6945
    },
    {
      "epoch": 3.6425576519916145,
      "grad_norm": 0.5549226999282837,
      "learning_rate": 4.898034263582554e-05,
      "loss": 0.6912,
      "num_input_tokens_seen": 4553168,
      "step": 6950
    },
    {
      "epoch": 3.6451781970649897,
      "grad_norm": 0.3705584704875946,
      "learning_rate": 4.897710786212467e-05,
      "loss": 0.4937,
      "num_input_tokens_seen": 4556720,
      "step": 6955
    },
    {
      "epoch": 3.647798742138365,
      "grad_norm": 0.40265318751335144,
      "learning_rate": 4.897386807268638e-05,
      "loss": 0.6621,
      "num_input_tokens_seen": 4560432,
      "step": 6960
    },
    {
      "epoch": 3.65041928721174,
      "grad_norm": 0.3254236876964569,
      "learning_rate": 4.897062326818839e-05,
      "loss": 0.5487,
      "num_input_tokens_seen": 4563792,
      "step": 6965
    },
    {
      "epoch": 3.653039832285115,
      "grad_norm": 0.44671645760536194,
      "learning_rate": 4.8967373449309495e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 4568144,
      "step": 6970
    },
    {
      "epoch": 3.6556603773584904,
      "grad_norm": 0.5043753981590271,
      "learning_rate": 4.89641186167295e-05,
      "loss": 0.4546,
      "num_input_tokens_seen": 4570448,
      "step": 6975
    },
    {
      "epoch": 3.658280922431866,
      "grad_norm": 0.8225173354148865,
      "learning_rate": 4.8960858771129294e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 4573232,
      "step": 6980
    },
    {
      "epoch": 3.660901467505241,
      "grad_norm": 1.2601350545883179,
      "learning_rate": 4.895759391319079e-05,
      "loss": 0.5608,
      "num_input_tokens_seen": 4576784,
      "step": 6985
    },
    {
      "epoch": 3.6635220125786163,
      "grad_norm": 0.371011346578598,
      "learning_rate": 4.8954324043596954e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 4580592,
      "step": 6990
    },
    {
      "epoch": 3.6661425576519915,
      "grad_norm": 0.5074666142463684,
      "learning_rate": 4.8951049163031826e-05,
      "loss": 0.4902,
      "num_input_tokens_seen": 4583792,
      "step": 6995
    },
    {
      "epoch": 3.668763102725367,
      "grad_norm": 0.9278416633605957,
      "learning_rate": 4.894776927218045e-05,
      "loss": 0.6353,
      "num_input_tokens_seen": 4586896,
      "step": 7000
    },
    {
      "epoch": 3.6713836477987423,
      "grad_norm": 0.28567618131637573,
      "learning_rate": 4.894448437172896e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 4590096,
      "step": 7005
    },
    {
      "epoch": 3.6740041928721174,
      "grad_norm": 0.3589094877243042,
      "learning_rate": 4.894119446236452e-05,
      "loss": 0.3805,
      "num_input_tokens_seen": 4592976,
      "step": 7010
    },
    {
      "epoch": 3.6766247379454926,
      "grad_norm": 0.3501511216163635,
      "learning_rate": 4.893789954477532e-05,
      "loss": 0.5553,
      "num_input_tokens_seen": 4595664,
      "step": 7015
    },
    {
      "epoch": 3.6792452830188678,
      "grad_norm": 0.49313825368881226,
      "learning_rate": 4.893459961965063e-05,
      "loss": 0.5915,
      "num_input_tokens_seen": 4598320,
      "step": 7020
    },
    {
      "epoch": 3.681865828092243,
      "grad_norm": 0.3856448829174042,
      "learning_rate": 4.8931294687680775e-05,
      "loss": 0.34,
      "num_input_tokens_seen": 4604976,
      "step": 7025
    },
    {
      "epoch": 3.6844863731656186,
      "grad_norm": 0.3408699929714203,
      "learning_rate": 4.8927984749557086e-05,
      "loss": 0.4563,
      "num_input_tokens_seen": 4608528,
      "step": 7030
    },
    {
      "epoch": 3.6871069182389937,
      "grad_norm": 0.43491122126579285,
      "learning_rate": 4.892466980597198e-05,
      "loss": 0.4371,
      "num_input_tokens_seen": 4611536,
      "step": 7035
    },
    {
      "epoch": 3.689727463312369,
      "grad_norm": 0.445855051279068,
      "learning_rate": 4.892134985761889e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 4614544,
      "step": 7040
    },
    {
      "epoch": 3.6923480083857445,
      "grad_norm": 0.41325443983078003,
      "learning_rate": 4.891802490519232e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 4617424,
      "step": 7045
    },
    {
      "epoch": 3.6949685534591197,
      "grad_norm": 0.5375360250473022,
      "learning_rate": 4.891469494938781e-05,
      "loss": 0.3682,
      "num_input_tokens_seen": 4620112,
      "step": 7050
    },
    {
      "epoch": 3.697589098532495,
      "grad_norm": 0.27221304178237915,
      "learning_rate": 4.891135999090195e-05,
      "loss": 0.5536,
      "num_input_tokens_seen": 4623888,
      "step": 7055
    },
    {
      "epoch": 3.70020964360587,
      "grad_norm": 0.36172252893447876,
      "learning_rate": 4.890802003043238e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 4627696,
      "step": 7060
    },
    {
      "epoch": 3.702830188679245,
      "grad_norm": 0.24998418986797333,
      "learning_rate": 4.890467506867777e-05,
      "loss": 0.4359,
      "num_input_tokens_seen": 4630288,
      "step": 7065
    },
    {
      "epoch": 3.7054507337526204,
      "grad_norm": 0.5584595799446106,
      "learning_rate": 4.890132510633786e-05,
      "loss": 0.5275,
      "num_input_tokens_seen": 4633168,
      "step": 7070
    },
    {
      "epoch": 3.708071278825996,
      "grad_norm": 0.7911328077316284,
      "learning_rate": 4.8897970144113416e-05,
      "loss": 0.549,
      "num_input_tokens_seen": 4635984,
      "step": 7075
    },
    {
      "epoch": 3.710691823899371,
      "grad_norm": 0.8071057796478271,
      "learning_rate": 4.889461018270626e-05,
      "loss": 0.5645,
      "num_input_tokens_seen": 4639312,
      "step": 7080
    },
    {
      "epoch": 3.7133123689727463,
      "grad_norm": 0.3280823230743408,
      "learning_rate": 4.889124522281926e-05,
      "loss": 0.5143,
      "num_input_tokens_seen": 4642320,
      "step": 7085
    },
    {
      "epoch": 3.7159329140461215,
      "grad_norm": 0.2969493567943573,
      "learning_rate": 4.888787526515632e-05,
      "loss": 0.5179,
      "num_input_tokens_seen": 4645040,
      "step": 7090
    },
    {
      "epoch": 3.718553459119497,
      "grad_norm": 0.6269901990890503,
      "learning_rate": 4.8884500310422406e-05,
      "loss": 0.4984,
      "num_input_tokens_seen": 4647856,
      "step": 7095
    },
    {
      "epoch": 3.7211740041928723,
      "grad_norm": 0.8981401920318604,
      "learning_rate": 4.8881120359323516e-05,
      "loss": 0.4631,
      "num_input_tokens_seen": 4651280,
      "step": 7100
    },
    {
      "epoch": 3.7237945492662474,
      "grad_norm": 0.5004167556762695,
      "learning_rate": 4.887773541256669e-05,
      "loss": 0.4081,
      "num_input_tokens_seen": 4654512,
      "step": 7105
    },
    {
      "epoch": 3.7264150943396226,
      "grad_norm": 0.8022763133049011,
      "learning_rate": 4.887434547086004e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 4657456,
      "step": 7110
    },
    {
      "epoch": 3.7290356394129978,
      "grad_norm": 0.7642826437950134,
      "learning_rate": 4.887095053491269e-05,
      "loss": 0.3449,
      "num_input_tokens_seen": 4660560,
      "step": 7115
    },
    {
      "epoch": 3.731656184486373,
      "grad_norm": 0.5450635552406311,
      "learning_rate": 4.8867550605434806e-05,
      "loss": 0.6064,
      "num_input_tokens_seen": 4663344,
      "step": 7120
    },
    {
      "epoch": 3.7342767295597485,
      "grad_norm": 0.7212480306625366,
      "learning_rate": 4.8864145683137644e-05,
      "loss": 0.5308,
      "num_input_tokens_seen": 4666384,
      "step": 7125
    },
    {
      "epoch": 3.7368972746331237,
      "grad_norm": 0.5170443654060364,
      "learning_rate": 4.886073576873346e-05,
      "loss": 0.4274,
      "num_input_tokens_seen": 4668880,
      "step": 7130
    },
    {
      "epoch": 3.739517819706499,
      "grad_norm": 0.37144166231155396,
      "learning_rate": 4.885732086293556e-05,
      "loss": 0.5542,
      "num_input_tokens_seen": 4672368,
      "step": 7135
    },
    {
      "epoch": 3.742138364779874,
      "grad_norm": 0.3407185673713684,
      "learning_rate": 4.8853900966458335e-05,
      "loss": 0.4905,
      "num_input_tokens_seen": 4676368,
      "step": 7140
    },
    {
      "epoch": 3.7447589098532497,
      "grad_norm": 0.8024254441261292,
      "learning_rate": 4.8850476080017146e-05,
      "loss": 0.4353,
      "num_input_tokens_seen": 4679632,
      "step": 7145
    },
    {
      "epoch": 3.747379454926625,
      "grad_norm": 0.47895658016204834,
      "learning_rate": 4.884704620432848e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 4683120,
      "step": 7150
    },
    {
      "epoch": 3.75,
      "grad_norm": 0.2931441068649292,
      "learning_rate": 4.8843611340109794e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 4687504,
      "step": 7155
    },
    {
      "epoch": 3.752620545073375,
      "grad_norm": 0.2496490180492401,
      "learning_rate": 4.884017148807965e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 4690448,
      "step": 7160
    },
    {
      "epoch": 3.7552410901467503,
      "grad_norm": 0.44449323415756226,
      "learning_rate": 4.883672664895761e-05,
      "loss": 0.4997,
      "num_input_tokens_seen": 4693104,
      "step": 7165
    },
    {
      "epoch": 3.757861635220126,
      "grad_norm": 0.38556256890296936,
      "learning_rate": 4.8833276823464296e-05,
      "loss": 0.3382,
      "num_input_tokens_seen": 4696080,
      "step": 7170
    },
    {
      "epoch": 3.760482180293501,
      "grad_norm": 0.3738654851913452,
      "learning_rate": 4.882982201232138e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 4698416,
      "step": 7175
    },
    {
      "epoch": 3.7631027253668763,
      "grad_norm": 0.39361631870269775,
      "learning_rate": 4.882636221625156e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 4701072,
      "step": 7180
    },
    {
      "epoch": 3.7657232704402515,
      "grad_norm": 0.5317738056182861,
      "learning_rate": 4.8822897435978586e-05,
      "loss": 0.5552,
      "num_input_tokens_seen": 4703696,
      "step": 7185
    },
    {
      "epoch": 3.768343815513627,
      "grad_norm": 0.5536747574806213,
      "learning_rate": 4.8819427672227264e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 4706288,
      "step": 7190
    },
    {
      "epoch": 3.7709643605870022,
      "grad_norm": 0.40226244926452637,
      "learning_rate": 4.881595292572341e-05,
      "loss": 0.5275,
      "num_input_tokens_seen": 4709040,
      "step": 7195
    },
    {
      "epoch": 3.7735849056603774,
      "grad_norm": 0.4528597891330719,
      "learning_rate": 4.881247319719392e-05,
      "loss": 0.4607,
      "num_input_tokens_seen": 4712944,
      "step": 7200
    },
    {
      "epoch": 3.7762054507337526,
      "grad_norm": 0.41922473907470703,
      "learning_rate": 4.88089884873667e-05,
      "loss": 0.428,
      "num_input_tokens_seen": 4715664,
      "step": 7205
    },
    {
      "epoch": 3.7788259958071277,
      "grad_norm": 0.8451485633850098,
      "learning_rate": 4.880549879697071e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 4718544,
      "step": 7210
    },
    {
      "epoch": 3.781446540880503,
      "grad_norm": 0.5481235980987549,
      "learning_rate": 4.8802004126735965e-05,
      "loss": 0.4778,
      "num_input_tokens_seen": 4721584,
      "step": 7215
    },
    {
      "epoch": 3.7840670859538785,
      "grad_norm": 0.49241966009140015,
      "learning_rate": 4.87985044773935e-05,
      "loss": 0.5958,
      "num_input_tokens_seen": 4724720,
      "step": 7220
    },
    {
      "epoch": 3.7866876310272537,
      "grad_norm": 0.8231102824211121,
      "learning_rate": 4.879499984967541e-05,
      "loss": 0.4845,
      "num_input_tokens_seen": 4727696,
      "step": 7225
    },
    {
      "epoch": 3.789308176100629,
      "grad_norm": 0.5029483437538147,
      "learning_rate": 4.8791490244314824e-05,
      "loss": 0.5068,
      "num_input_tokens_seen": 4730832,
      "step": 7230
    },
    {
      "epoch": 3.791928721174004,
      "grad_norm": 0.40084758400917053,
      "learning_rate": 4.8787975662045904e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 4733456,
      "step": 7235
    },
    {
      "epoch": 3.7945492662473796,
      "grad_norm": 0.3878041207790375,
      "learning_rate": 4.878445610360386e-05,
      "loss": 0.525,
      "num_input_tokens_seen": 4737008,
      "step": 7240
    },
    {
      "epoch": 3.797169811320755,
      "grad_norm": 0.40110310912132263,
      "learning_rate": 4.878093156972494e-05,
      "loss": 0.7625,
      "num_input_tokens_seen": 4740560,
      "step": 7245
    },
    {
      "epoch": 3.79979035639413,
      "grad_norm": 0.2508857250213623,
      "learning_rate": 4.877740206114645e-05,
      "loss": 0.4497,
      "num_input_tokens_seen": 4743536,
      "step": 7250
    },
    {
      "epoch": 3.802410901467505,
      "grad_norm": 0.20483607053756714,
      "learning_rate": 4.8773867578606716e-05,
      "loss": 0.3927,
      "num_input_tokens_seen": 4748144,
      "step": 7255
    },
    {
      "epoch": 3.8050314465408803,
      "grad_norm": 0.38376620411872864,
      "learning_rate": 4.877032812284511e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 4751248,
      "step": 7260
    },
    {
      "epoch": 3.8076519916142555,
      "grad_norm": 0.317624568939209,
      "learning_rate": 4.876678369460205e-05,
      "loss": 0.5224,
      "num_input_tokens_seen": 4754832,
      "step": 7265
    },
    {
      "epoch": 3.810272536687631,
      "grad_norm": 0.28078436851501465,
      "learning_rate": 4.8763234294618985e-05,
      "loss": 0.6439,
      "num_input_tokens_seen": 4758224,
      "step": 7270
    },
    {
      "epoch": 3.8128930817610063,
      "grad_norm": 0.2732778489589691,
      "learning_rate": 4.875967992363841e-05,
      "loss": 0.5069,
      "num_input_tokens_seen": 4761040,
      "step": 7275
    },
    {
      "epoch": 3.8155136268343814,
      "grad_norm": 0.7119266986846924,
      "learning_rate": 4.875612058240386e-05,
      "loss": 0.5034,
      "num_input_tokens_seen": 4763728,
      "step": 7280
    },
    {
      "epoch": 3.818134171907757,
      "grad_norm": 0.5597074627876282,
      "learning_rate": 4.875255627165991e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 4766864,
      "step": 7285
    },
    {
      "epoch": 3.8207547169811322,
      "grad_norm": 1.3794105052947998,
      "learning_rate": 4.874898699215217e-05,
      "loss": 0.6005,
      "num_input_tokens_seen": 4769872,
      "step": 7290
    },
    {
      "epoch": 3.8233752620545074,
      "grad_norm": 0.38024717569351196,
      "learning_rate": 4.8745412744627296e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 4772816,
      "step": 7295
    },
    {
      "epoch": 3.8259958071278826,
      "grad_norm": 0.34289100766181946,
      "learning_rate": 4.874183352983297e-05,
      "loss": 0.5465,
      "num_input_tokens_seen": 4776336,
      "step": 7300
    },
    {
      "epoch": 3.8286163522012577,
      "grad_norm": 0.3185069262981415,
      "learning_rate": 4.873824934851794e-05,
      "loss": 0.6208,
      "num_input_tokens_seen": 4780240,
      "step": 7305
    },
    {
      "epoch": 3.831236897274633,
      "grad_norm": 0.4937554895877838,
      "learning_rate": 4.873466020143196e-05,
      "loss": 0.6864,
      "num_input_tokens_seen": 4782544,
      "step": 7310
    },
    {
      "epoch": 3.8338574423480085,
      "grad_norm": 0.5746698975563049,
      "learning_rate": 4.873106608932585e-05,
      "loss": 0.478,
      "num_input_tokens_seen": 4785104,
      "step": 7315
    },
    {
      "epoch": 3.8364779874213837,
      "grad_norm": 0.2786664366722107,
      "learning_rate": 4.872746701295145e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 4788016,
      "step": 7320
    },
    {
      "epoch": 3.839098532494759,
      "grad_norm": 0.4647200107574463,
      "learning_rate": 4.872386297306165e-05,
      "loss": 0.491,
      "num_input_tokens_seen": 4790640,
      "step": 7325
    },
    {
      "epoch": 3.841719077568134,
      "grad_norm": 0.5991224646568298,
      "learning_rate": 4.872025397041037e-05,
      "loss": 0.6248,
      "num_input_tokens_seen": 4793232,
      "step": 7330
    },
    {
      "epoch": 3.8443396226415096,
      "grad_norm": 0.7208259105682373,
      "learning_rate": 4.871664000575258e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 4797136,
      "step": 7335
    },
    {
      "epoch": 3.846960167714885,
      "grad_norm": 0.5603818297386169,
      "learning_rate": 4.871302107984427e-05,
      "loss": 0.7628,
      "num_input_tokens_seen": 4800784,
      "step": 7340
    },
    {
      "epoch": 3.84958071278826,
      "grad_norm": 0.27272507548332214,
      "learning_rate": 4.8709397193442474e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 4804432,
      "step": 7345
    },
    {
      "epoch": 3.852201257861635,
      "grad_norm": 0.30088770389556885,
      "learning_rate": 4.870576834730529e-05,
      "loss": 0.5471,
      "num_input_tokens_seen": 4808208,
      "step": 7350
    },
    {
      "epoch": 3.8548218029350103,
      "grad_norm": 0.5184746980667114,
      "learning_rate": 4.8702134542191816e-05,
      "loss": 0.573,
      "num_input_tokens_seen": 4811248,
      "step": 7355
    },
    {
      "epoch": 3.8574423480083855,
      "grad_norm": 0.5844084024429321,
      "learning_rate": 4.86984957788622e-05,
      "loss": 0.5266,
      "num_input_tokens_seen": 4814352,
      "step": 7360
    },
    {
      "epoch": 3.860062893081761,
      "grad_norm": 0.4091830849647522,
      "learning_rate": 4.869485205807763e-05,
      "loss": 0.5079,
      "num_input_tokens_seen": 4819152,
      "step": 7365
    },
    {
      "epoch": 3.8626834381551363,
      "grad_norm": 0.4451589286327362,
      "learning_rate": 4.869120338060034e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 4822224,
      "step": 7370
    },
    {
      "epoch": 3.8653039832285114,
      "grad_norm": 0.4955470860004425,
      "learning_rate": 4.8687549747193585e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 4825648,
      "step": 7375
    },
    {
      "epoch": 3.867924528301887,
      "grad_norm": 0.23475369811058044,
      "learning_rate": 4.868389115862166e-05,
      "loss": 0.3666,
      "num_input_tokens_seen": 4828816,
      "step": 7380
    },
    {
      "epoch": 3.870545073375262,
      "grad_norm": 0.6199975609779358,
      "learning_rate": 4.868022761564991e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 4831408,
      "step": 7385
    },
    {
      "epoch": 3.8731656184486374,
      "grad_norm": 0.2596244215965271,
      "learning_rate": 4.8676559119044696e-05,
      "loss": 0.4405,
      "num_input_tokens_seen": 4834864,
      "step": 7390
    },
    {
      "epoch": 3.8757861635220126,
      "grad_norm": 0.4456653296947479,
      "learning_rate": 4.867288566957344e-05,
      "loss": 0.5044,
      "num_input_tokens_seen": 4837552,
      "step": 7395
    },
    {
      "epoch": 3.8784067085953877,
      "grad_norm": 0.6299834847450256,
      "learning_rate": 4.866920726800457e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 4840336,
      "step": 7400
    },
    {
      "epoch": 3.881027253668763,
      "grad_norm": 0.3656795918941498,
      "learning_rate": 4.8665523915107576e-05,
      "loss": 0.5374,
      "num_input_tokens_seen": 4843184,
      "step": 7405
    },
    {
      "epoch": 3.8836477987421385,
      "grad_norm": 0.35575130581855774,
      "learning_rate": 4.866183561165296e-05,
      "loss": 0.5685,
      "num_input_tokens_seen": 4846576,
      "step": 7410
    },
    {
      "epoch": 3.8862683438155137,
      "grad_norm": 0.5259768962860107,
      "learning_rate": 4.86581423584123e-05,
      "loss": 0.4657,
      "num_input_tokens_seen": 4849392,
      "step": 7415
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 0.45726269483566284,
      "learning_rate": 4.8654444156158154e-05,
      "loss": 0.3541,
      "num_input_tokens_seen": 4852752,
      "step": 7420
    },
    {
      "epoch": 3.891509433962264,
      "grad_norm": 0.32892706990242004,
      "learning_rate": 4.865074100566415e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 4856880,
      "step": 7425
    },
    {
      "epoch": 3.8941299790356396,
      "grad_norm": 0.5452679395675659,
      "learning_rate": 4.864703290770495e-05,
      "loss": 0.5472,
      "num_input_tokens_seen": 4859984,
      "step": 7430
    },
    {
      "epoch": 3.896750524109015,
      "grad_norm": 0.48041877150535583,
      "learning_rate": 4.864331986305626e-05,
      "loss": 0.4239,
      "num_input_tokens_seen": 4862640,
      "step": 7435
    },
    {
      "epoch": 3.89937106918239,
      "grad_norm": 1.188927173614502,
      "learning_rate": 4.8639601872494775e-05,
      "loss": 0.5539,
      "num_input_tokens_seen": 4865168,
      "step": 7440
    },
    {
      "epoch": 3.901991614255765,
      "grad_norm": 0.384124219417572,
      "learning_rate": 4.8635878936798287e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 4869520,
      "step": 7445
    },
    {
      "epoch": 3.9046121593291403,
      "grad_norm": 0.33435508608818054,
      "learning_rate": 4.863215105674557e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 4872944,
      "step": 7450
    },
    {
      "epoch": 3.9072327044025155,
      "grad_norm": 0.5585997700691223,
      "learning_rate": 4.8628418233116457e-05,
      "loss": 0.4444,
      "num_input_tokens_seen": 4876624,
      "step": 7455
    },
    {
      "epoch": 3.909853249475891,
      "grad_norm": 0.46226587891578674,
      "learning_rate": 4.8624680466691824e-05,
      "loss": 0.5065,
      "num_input_tokens_seen": 4879152,
      "step": 7460
    },
    {
      "epoch": 3.9124737945492662,
      "grad_norm": 0.3799794912338257,
      "learning_rate": 4.862093775825356e-05,
      "loss": 0.4086,
      "num_input_tokens_seen": 4883760,
      "step": 7465
    },
    {
      "epoch": 3.9150943396226414,
      "grad_norm": 0.3970862329006195,
      "learning_rate": 4.86171901085846e-05,
      "loss": 0.4719,
      "num_input_tokens_seen": 4886640,
      "step": 7470
    },
    {
      "epoch": 3.917714884696017,
      "grad_norm": 0.4858344495296478,
      "learning_rate": 4.861343751846891e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 4890896,
      "step": 7475
    },
    {
      "epoch": 3.920335429769392,
      "grad_norm": 0.5809312462806702,
      "learning_rate": 4.860967998869148e-05,
      "loss": 0.5437,
      "num_input_tokens_seen": 4893744,
      "step": 7480
    },
    {
      "epoch": 3.9229559748427674,
      "grad_norm": 0.4346202313899994,
      "learning_rate": 4.860591752003836e-05,
      "loss": 0.6735,
      "num_input_tokens_seen": 4896240,
      "step": 7485
    },
    {
      "epoch": 3.9255765199161425,
      "grad_norm": 0.5646572709083557,
      "learning_rate": 4.8602150113296596e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 4899856,
      "step": 7490
    },
    {
      "epoch": 3.9281970649895177,
      "grad_norm": 0.3330306112766266,
      "learning_rate": 4.8598377769254305e-05,
      "loss": 0.4388,
      "num_input_tokens_seen": 4903312,
      "step": 7495
    },
    {
      "epoch": 3.930817610062893,
      "grad_norm": 0.31318050622940063,
      "learning_rate": 4.85946004887006e-05,
      "loss": 0.4705,
      "num_input_tokens_seen": 4905744,
      "step": 7500
    },
    {
      "epoch": 3.9334381551362685,
      "grad_norm": 0.6413263082504272,
      "learning_rate": 4.859081827242566e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 4908368,
      "step": 7505
    },
    {
      "epoch": 3.9360587002096437,
      "grad_norm": 0.4184214472770691,
      "learning_rate": 4.858703112122068e-05,
      "loss": 0.401,
      "num_input_tokens_seen": 4912112,
      "step": 7510
    },
    {
      "epoch": 3.938679245283019,
      "grad_norm": 0.3468869626522064,
      "learning_rate": 4.8583239035877876e-05,
      "loss": 0.4839,
      "num_input_tokens_seen": 4915216,
      "step": 7515
    },
    {
      "epoch": 3.941299790356394,
      "grad_norm": 0.367424875497818,
      "learning_rate": 4.8579442017190526e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 4918320,
      "step": 7520
    },
    {
      "epoch": 3.9439203354297696,
      "grad_norm": 0.5549091696739197,
      "learning_rate": 4.8575640065952905e-05,
      "loss": 0.4118,
      "num_input_tokens_seen": 4922384,
      "step": 7525
    },
    {
      "epoch": 3.9465408805031448,
      "grad_norm": 0.39500218629837036,
      "learning_rate": 4.857183318296035e-05,
      "loss": 0.4837,
      "num_input_tokens_seen": 4925904,
      "step": 7530
    },
    {
      "epoch": 3.94916142557652,
      "grad_norm": 0.9955672025680542,
      "learning_rate": 4.856802136900922e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 4928112,
      "step": 7535
    },
    {
      "epoch": 3.951781970649895,
      "grad_norm": 0.5526247024536133,
      "learning_rate": 4.856420462489689e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 4931536,
      "step": 7540
    },
    {
      "epoch": 3.9544025157232703,
      "grad_norm": 0.32646432518959045,
      "learning_rate": 4.856038295142179e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 4935184,
      "step": 7545
    },
    {
      "epoch": 3.9570230607966455,
      "grad_norm": 0.48020225763320923,
      "learning_rate": 4.855655634938337e-05,
      "loss": 0.3744,
      "num_input_tokens_seen": 4938704,
      "step": 7550
    },
    {
      "epoch": 3.959643605870021,
      "grad_norm": 0.45446786284446716,
      "learning_rate": 4.8552724819582106e-05,
      "loss": 0.49,
      "num_input_tokens_seen": 4941360,
      "step": 7555
    },
    {
      "epoch": 3.9622641509433962,
      "grad_norm": 0.4850861430168152,
      "learning_rate": 4.8548888362819515e-05,
      "loss": 0.5718,
      "num_input_tokens_seen": 4944656,
      "step": 7560
    },
    {
      "epoch": 3.9648846960167714,
      "grad_norm": 0.8152313828468323,
      "learning_rate": 4.854504697989814e-05,
      "loss": 0.3973,
      "num_input_tokens_seen": 4947248,
      "step": 7565
    },
    {
      "epoch": 3.967505241090147,
      "grad_norm": 0.4323869049549103,
      "learning_rate": 4.8541200671621554e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 4950160,
      "step": 7570
    },
    {
      "epoch": 3.970125786163522,
      "grad_norm": 0.8035544157028198,
      "learning_rate": 4.8537349438794354e-05,
      "loss": 0.5549,
      "num_input_tokens_seen": 4953264,
      "step": 7575
    },
    {
      "epoch": 3.9727463312368974,
      "grad_norm": 0.5596507787704468,
      "learning_rate": 4.853349328222219e-05,
      "loss": 0.5128,
      "num_input_tokens_seen": 4955760,
      "step": 7580
    },
    {
      "epoch": 3.9753668763102725,
      "grad_norm": 0.4686371386051178,
      "learning_rate": 4.8529632202711706e-05,
      "loss": 0.4929,
      "num_input_tokens_seen": 4961616,
      "step": 7585
    },
    {
      "epoch": 3.9779874213836477,
      "grad_norm": 1.0545135736465454,
      "learning_rate": 4.852576620107061e-05,
      "loss": 0.3409,
      "num_input_tokens_seen": 4964240,
      "step": 7590
    },
    {
      "epoch": 3.980607966457023,
      "grad_norm": 0.8650411367416382,
      "learning_rate": 4.8521895278107615e-05,
      "loss": 0.4254,
      "num_input_tokens_seen": 4968016,
      "step": 7595
    },
    {
      "epoch": 3.9832285115303985,
      "grad_norm": 0.23860488831996918,
      "learning_rate": 4.8518019434632484e-05,
      "loss": 0.4877,
      "num_input_tokens_seen": 4971952,
      "step": 7600
    },
    {
      "epoch": 3.9858490566037736,
      "grad_norm": 0.4314138889312744,
      "learning_rate": 4.8514138671455995e-05,
      "loss": 0.5027,
      "num_input_tokens_seen": 4975984,
      "step": 7605
    },
    {
      "epoch": 3.988469601677149,
      "grad_norm": 0.5006566047668457,
      "learning_rate": 4.851025298938996e-05,
      "loss": 0.5367,
      "num_input_tokens_seen": 4978832,
      "step": 7610
    },
    {
      "epoch": 3.991090146750524,
      "grad_norm": 0.4341861307621002,
      "learning_rate": 4.850636238924722e-05,
      "loss": 0.423,
      "num_input_tokens_seen": 4981808,
      "step": 7615
    },
    {
      "epoch": 3.9937106918238996,
      "grad_norm": 0.3789421617984772,
      "learning_rate": 4.850246687184163e-05,
      "loss": 0.6379,
      "num_input_tokens_seen": 4985072,
      "step": 7620
    },
    {
      "epoch": 3.9963312368972748,
      "grad_norm": 0.4476439952850342,
      "learning_rate": 4.849856643798812e-05,
      "loss": 0.4152,
      "num_input_tokens_seen": 4987920,
      "step": 7625
    },
    {
      "epoch": 3.99895178197065,
      "grad_norm": 0.2983989715576172,
      "learning_rate": 4.849466108850259e-05,
      "loss": 0.4606,
      "num_input_tokens_seen": 4990768,
      "step": 7630
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.46252498030662537,
      "eval_runtime": 9.6443,
      "eval_samples_per_second": 87.927,
      "eval_steps_per_second": 21.982,
      "num_input_tokens_seen": 4991472,
      "step": 7632
    },
    {
      "epoch": 4.001572327044025,
      "grad_norm": 0.3628941476345062,
      "learning_rate": 4.8490750824202e-05,
      "loss": 0.58,
      "num_input_tokens_seen": 4993136,
      "step": 7635
    },
    {
      "epoch": 4.0041928721174,
      "grad_norm": 0.5987470149993896,
      "learning_rate": 4.8486835645904324e-05,
      "loss": 0.4366,
      "num_input_tokens_seen": 4997680,
      "step": 7640
    },
    {
      "epoch": 4.006813417190775,
      "grad_norm": 0.3909796476364136,
      "learning_rate": 4.84829155544286e-05,
      "loss": 0.4717,
      "num_input_tokens_seen": 5000976,
      "step": 7645
    },
    {
      "epoch": 4.009433962264151,
      "grad_norm": 0.5959615707397461,
      "learning_rate": 4.847899055059484e-05,
      "loss": 0.4303,
      "num_input_tokens_seen": 5003664,
      "step": 7650
    },
    {
      "epoch": 4.012054507337526,
      "grad_norm": 0.2506575584411621,
      "learning_rate": 4.847506063522412e-05,
      "loss": 0.4915,
      "num_input_tokens_seen": 5006832,
      "step": 7655
    },
    {
      "epoch": 4.014675052410902,
      "grad_norm": 0.5814336538314819,
      "learning_rate": 4.847112580913854e-05,
      "loss": 0.4387,
      "num_input_tokens_seen": 5010512,
      "step": 7660
    },
    {
      "epoch": 4.017295597484277,
      "grad_norm": 0.4425448775291443,
      "learning_rate": 4.846718607316121e-05,
      "loss": 0.5616,
      "num_input_tokens_seen": 5014352,
      "step": 7665
    },
    {
      "epoch": 4.019916142557652,
      "grad_norm": 0.41732484102249146,
      "learning_rate": 4.846324142811628e-05,
      "loss": 0.4519,
      "num_input_tokens_seen": 5017072,
      "step": 7670
    },
    {
      "epoch": 4.022536687631027,
      "grad_norm": 0.3216921389102936,
      "learning_rate": 4.8459291874828924e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 5019824,
      "step": 7675
    },
    {
      "epoch": 4.0251572327044025,
      "grad_norm": 0.44719964265823364,
      "learning_rate": 4.845533741412534e-05,
      "loss": 0.5426,
      "num_input_tokens_seen": 5022800,
      "step": 7680
    },
    {
      "epoch": 4.027777777777778,
      "grad_norm": 0.3681773841381073,
      "learning_rate": 4.845137804683277e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 5025424,
      "step": 7685
    },
    {
      "epoch": 4.030398322851153,
      "grad_norm": 0.3308062255382538,
      "learning_rate": 4.844741377377946e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 5028144,
      "step": 7690
    },
    {
      "epoch": 4.033018867924528,
      "grad_norm": 0.746774435043335,
      "learning_rate": 4.844344459579469e-05,
      "loss": 0.4866,
      "num_input_tokens_seen": 5031568,
      "step": 7695
    },
    {
      "epoch": 4.035639412997903,
      "grad_norm": 0.43867620825767517,
      "learning_rate": 4.8439470513708766e-05,
      "loss": 0.5209,
      "num_input_tokens_seen": 5034960,
      "step": 7700
    },
    {
      "epoch": 4.038259958071279,
      "grad_norm": 0.34572556614875793,
      "learning_rate": 4.8435491528353026e-05,
      "loss": 0.3664,
      "num_input_tokens_seen": 5037552,
      "step": 7705
    },
    {
      "epoch": 4.040880503144654,
      "grad_norm": 0.41245853900909424,
      "learning_rate": 4.843150764055982e-05,
      "loss": 0.4704,
      "num_input_tokens_seen": 5041136,
      "step": 7710
    },
    {
      "epoch": 4.04350104821803,
      "grad_norm": 0.3895730972290039,
      "learning_rate": 4.842751885116254e-05,
      "loss": 0.5178,
      "num_input_tokens_seen": 5044240,
      "step": 7715
    },
    {
      "epoch": 4.046121593291405,
      "grad_norm": 0.3385066092014313,
      "learning_rate": 4.842352516099559e-05,
      "loss": 0.5779,
      "num_input_tokens_seen": 5047248,
      "step": 7720
    },
    {
      "epoch": 4.04874213836478,
      "grad_norm": 0.41425809264183044,
      "learning_rate": 4.84195265708944e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 5050128,
      "step": 7725
    },
    {
      "epoch": 4.051362683438155,
      "grad_norm": 0.3308907449245453,
      "learning_rate": 4.841552308169545e-05,
      "loss": 0.5052,
      "num_input_tokens_seen": 5053744,
      "step": 7730
    },
    {
      "epoch": 4.05398322851153,
      "grad_norm": 0.3523765504360199,
      "learning_rate": 4.841151469423619e-05,
      "loss": 0.5662,
      "num_input_tokens_seen": 5056912,
      "step": 7735
    },
    {
      "epoch": 4.056603773584905,
      "grad_norm": 0.478884220123291,
      "learning_rate": 4.840750140935516e-05,
      "loss": 0.4696,
      "num_input_tokens_seen": 5060880,
      "step": 7740
    },
    {
      "epoch": 4.059224318658281,
      "grad_norm": 0.4720120429992676,
      "learning_rate": 4.840348322789189e-05,
      "loss": 0.6218,
      "num_input_tokens_seen": 5063888,
      "step": 7745
    },
    {
      "epoch": 4.061844863731656,
      "grad_norm": 0.3160932660102844,
      "learning_rate": 4.839946015068692e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 5066960,
      "step": 7750
    },
    {
      "epoch": 4.064465408805032,
      "grad_norm": 1.2741137742996216,
      "learning_rate": 4.839543217858184e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 5070352,
      "step": 7755
    },
    {
      "epoch": 4.067085953878407,
      "grad_norm": 0.43520045280456543,
      "learning_rate": 4.839139931241926e-05,
      "loss": 0.6216,
      "num_input_tokens_seen": 5073488,
      "step": 7760
    },
    {
      "epoch": 4.069706498951782,
      "grad_norm": 0.4893611669540405,
      "learning_rate": 4.838736155304281e-05,
      "loss": 0.4063,
      "num_input_tokens_seen": 5076720,
      "step": 7765
    },
    {
      "epoch": 4.072327044025157,
      "grad_norm": 0.5871918201446533,
      "learning_rate": 4.8383318901297135e-05,
      "loss": 0.5566,
      "num_input_tokens_seen": 5080080,
      "step": 7770
    },
    {
      "epoch": 4.0749475890985325,
      "grad_norm": 0.472781777381897,
      "learning_rate": 4.837927135802792e-05,
      "loss": 0.4917,
      "num_input_tokens_seen": 5083056,
      "step": 7775
    },
    {
      "epoch": 4.077568134171908,
      "grad_norm": 0.37866541743278503,
      "learning_rate": 4.8375218924081856e-05,
      "loss": 0.5486,
      "num_input_tokens_seen": 5085904,
      "step": 7780
    },
    {
      "epoch": 4.080188679245283,
      "grad_norm": 0.40734535455703735,
      "learning_rate": 4.837116160030668e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 5088816,
      "step": 7785
    },
    {
      "epoch": 4.082809224318658,
      "grad_norm": 0.3112063407897949,
      "learning_rate": 4.8367099387551126e-05,
      "loss": 0.5312,
      "num_input_tokens_seen": 5092112,
      "step": 7790
    },
    {
      "epoch": 4.085429769392033,
      "grad_norm": 0.8378486037254333,
      "learning_rate": 4.8363032286664964e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 5095536,
      "step": 7795
    },
    {
      "epoch": 4.088050314465409,
      "grad_norm": 0.37174543738365173,
      "learning_rate": 4.835896029849899e-05,
      "loss": 0.5722,
      "num_input_tokens_seen": 5099216,
      "step": 7800
    },
    {
      "epoch": 4.090670859538784,
      "grad_norm": 0.2920002043247223,
      "learning_rate": 4.835488342390502e-05,
      "loss": 0.4034,
      "num_input_tokens_seen": 5102288,
      "step": 7805
    },
    {
      "epoch": 4.09329140461216,
      "grad_norm": 0.5126864910125732,
      "learning_rate": 4.835080166373587e-05,
      "loss": 0.5241,
      "num_input_tokens_seen": 5105232,
      "step": 7810
    },
    {
      "epoch": 4.095911949685535,
      "grad_norm": 0.36013638973236084,
      "learning_rate": 4.834671501884542e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 5109840,
      "step": 7815
    },
    {
      "epoch": 4.09853249475891,
      "grad_norm": 0.7054063677787781,
      "learning_rate": 4.8342623490088543e-05,
      "loss": 0.4884,
      "num_input_tokens_seen": 5112240,
      "step": 7820
    },
    {
      "epoch": 4.101153039832285,
      "grad_norm": 0.20897535979747772,
      "learning_rate": 4.833852707832114e-05,
      "loss": 0.4671,
      "num_input_tokens_seen": 5115536,
      "step": 7825
    },
    {
      "epoch": 4.10377358490566,
      "grad_norm": 0.5967658162117004,
      "learning_rate": 4.833442578440013e-05,
      "loss": 0.4838,
      "num_input_tokens_seen": 5118160,
      "step": 7830
    },
    {
      "epoch": 4.106394129979035,
      "grad_norm": 0.6640020608901978,
      "learning_rate": 4.833031960918346e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 5120720,
      "step": 7835
    },
    {
      "epoch": 4.109014675052411,
      "grad_norm": 0.2908872663974762,
      "learning_rate": 4.8326208553530095e-05,
      "loss": 0.5281,
      "num_input_tokens_seen": 5123632,
      "step": 7840
    },
    {
      "epoch": 4.111635220125786,
      "grad_norm": 0.6143691539764404,
      "learning_rate": 4.832209261830002e-05,
      "loss": 0.5068,
      "num_input_tokens_seen": 5126288,
      "step": 7845
    },
    {
      "epoch": 4.114255765199162,
      "grad_norm": 1.423507809638977,
      "learning_rate": 4.831797180435423e-05,
      "loss": 0.741,
      "num_input_tokens_seen": 5128752,
      "step": 7850
    },
    {
      "epoch": 4.116876310272537,
      "grad_norm": 0.31298136711120605,
      "learning_rate": 4.831384611255478e-05,
      "loss": 0.5156,
      "num_input_tokens_seen": 5132048,
      "step": 7855
    },
    {
      "epoch": 4.119496855345912,
      "grad_norm": 0.45017796754837036,
      "learning_rate": 4.8309715543764697e-05,
      "loss": 0.4739,
      "num_input_tokens_seen": 5135024,
      "step": 7860
    },
    {
      "epoch": 4.122117400419287,
      "grad_norm": 0.35425952076911926,
      "learning_rate": 4.8305580098848046e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 5138512,
      "step": 7865
    },
    {
      "epoch": 4.1247379454926625,
      "grad_norm": 0.5566848516464233,
      "learning_rate": 4.830143977866993e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 5141872,
      "step": 7870
    },
    {
      "epoch": 4.127358490566038,
      "grad_norm": 0.526434600353241,
      "learning_rate": 4.829729458409645e-05,
      "loss": 0.5017,
      "num_input_tokens_seen": 5147664,
      "step": 7875
    },
    {
      "epoch": 4.129979035639413,
      "grad_norm": 0.8127632737159729,
      "learning_rate": 4.8293144515994725e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 5151696,
      "step": 7880
    },
    {
      "epoch": 4.132599580712788,
      "grad_norm": 0.4213475286960602,
      "learning_rate": 4.828898957523292e-05,
      "loss": 0.4679,
      "num_input_tokens_seen": 5154800,
      "step": 7885
    },
    {
      "epoch": 4.135220125786163,
      "grad_norm": 0.3787809908390045,
      "learning_rate": 4.8284829762680183e-05,
      "loss": 0.5224,
      "num_input_tokens_seen": 5157744,
      "step": 7890
    },
    {
      "epoch": 4.137840670859539,
      "grad_norm": 0.4883621037006378,
      "learning_rate": 4.828066507920671e-05,
      "loss": 0.3836,
      "num_input_tokens_seen": 5160944,
      "step": 7895
    },
    {
      "epoch": 4.140461215932914,
      "grad_norm": 0.4848870635032654,
      "learning_rate": 4.82764955256837e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 5163920,
      "step": 7900
    },
    {
      "epoch": 4.1430817610062896,
      "grad_norm": 0.9695469737052917,
      "learning_rate": 4.827232110298338e-05,
      "loss": 0.5928,
      "num_input_tokens_seen": 5167760,
      "step": 7905
    },
    {
      "epoch": 4.145702306079665,
      "grad_norm": 0.6198493242263794,
      "learning_rate": 4.8268141811978995e-05,
      "loss": 0.5104,
      "num_input_tokens_seen": 5170800,
      "step": 7910
    },
    {
      "epoch": 4.14832285115304,
      "grad_norm": 0.46923691034317017,
      "learning_rate": 4.826395765354479e-05,
      "loss": 0.4292,
      "num_input_tokens_seen": 5173488,
      "step": 7915
    },
    {
      "epoch": 4.150943396226415,
      "grad_norm": 0.3830249011516571,
      "learning_rate": 4.825976862855606e-05,
      "loss": 0.5298,
      "num_input_tokens_seen": 5176272,
      "step": 7920
    },
    {
      "epoch": 4.15356394129979,
      "grad_norm": 0.5204660892486572,
      "learning_rate": 4.8255574737889094e-05,
      "loss": 0.4467,
      "num_input_tokens_seen": 5178896,
      "step": 7925
    },
    {
      "epoch": 4.156184486373165,
      "grad_norm": 0.5980522036552429,
      "learning_rate": 4.82513759824212e-05,
      "loss": 0.4623,
      "num_input_tokens_seen": 5182192,
      "step": 7930
    },
    {
      "epoch": 4.158805031446541,
      "grad_norm": 0.44209691882133484,
      "learning_rate": 4.824717236303072e-05,
      "loss": 0.7751,
      "num_input_tokens_seen": 5185232,
      "step": 7935
    },
    {
      "epoch": 4.161425576519916,
      "grad_norm": 0.35121944546699524,
      "learning_rate": 4.8242963880597005e-05,
      "loss": 0.6619,
      "num_input_tokens_seen": 5188176,
      "step": 7940
    },
    {
      "epoch": 4.164046121593292,
      "grad_norm": 0.36018383502960205,
      "learning_rate": 4.823875053600041e-05,
      "loss": 0.6081,
      "num_input_tokens_seen": 5190768,
      "step": 7945
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 0.6391671299934387,
      "learning_rate": 4.823453233012233e-05,
      "loss": 0.4417,
      "num_input_tokens_seen": 5193584,
      "step": 7950
    },
    {
      "epoch": 4.169287211740042,
      "grad_norm": 0.5885236859321594,
      "learning_rate": 4.8230309263845155e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 5196592,
      "step": 7955
    },
    {
      "epoch": 4.171907756813417,
      "grad_norm": 0.26471376419067383,
      "learning_rate": 4.822608133805231e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 5199024,
      "step": 7960
    },
    {
      "epoch": 4.1745283018867925,
      "grad_norm": 0.3856058120727539,
      "learning_rate": 4.8221848553628226e-05,
      "loss": 0.4295,
      "num_input_tokens_seen": 5202288,
      "step": 7965
    },
    {
      "epoch": 4.177148846960168,
      "grad_norm": 0.287984162569046,
      "learning_rate": 4.8217610911458346e-05,
      "loss": 0.5816,
      "num_input_tokens_seen": 5205296,
      "step": 7970
    },
    {
      "epoch": 4.179769392033543,
      "grad_norm": 0.27875787019729614,
      "learning_rate": 4.821336841242915e-05,
      "loss": 0.4371,
      "num_input_tokens_seen": 5208912,
      "step": 7975
    },
    {
      "epoch": 4.182389937106918,
      "grad_norm": 1.0782970190048218,
      "learning_rate": 4.820912105742811e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 5212048,
      "step": 7980
    },
    {
      "epoch": 4.185010482180293,
      "grad_norm": 0.5030689239501953,
      "learning_rate": 4.8204868847343736e-05,
      "loss": 0.4583,
      "num_input_tokens_seen": 5214800,
      "step": 7985
    },
    {
      "epoch": 4.187631027253669,
      "grad_norm": 0.3879414200782776,
      "learning_rate": 4.820061178306553e-05,
      "loss": 0.337,
      "num_input_tokens_seen": 5217744,
      "step": 7990
    },
    {
      "epoch": 4.190251572327044,
      "grad_norm": 0.8272335529327393,
      "learning_rate": 4.819634986548403e-05,
      "loss": 0.6327,
      "num_input_tokens_seen": 5220784,
      "step": 7995
    },
    {
      "epoch": 4.1928721174004195,
      "grad_norm": 0.5089852809906006,
      "learning_rate": 4.819208309549077e-05,
      "loss": 0.4266,
      "num_input_tokens_seen": 5223504,
      "step": 8000
    },
    {
      "epoch": 4.195492662473795,
      "grad_norm": 0.461283415555954,
      "learning_rate": 4.8187811473978316e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 5226672,
      "step": 8005
    },
    {
      "epoch": 4.19811320754717,
      "grad_norm": 0.48428210616111755,
      "learning_rate": 4.818353500184025e-05,
      "loss": 0.358,
      "num_input_tokens_seen": 5229808,
      "step": 8010
    },
    {
      "epoch": 4.200733752620545,
      "grad_norm": 0.6151285767555237,
      "learning_rate": 4.817925367997115e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 5233328,
      "step": 8015
    },
    {
      "epoch": 4.20335429769392,
      "grad_norm": 0.2185632586479187,
      "learning_rate": 4.8174967509266615e-05,
      "loss": 0.4069,
      "num_input_tokens_seen": 5238512,
      "step": 8020
    },
    {
      "epoch": 4.205974842767295,
      "grad_norm": 0.3695061206817627,
      "learning_rate": 4.817067649062328e-05,
      "loss": 0.4709,
      "num_input_tokens_seen": 5242256,
      "step": 8025
    },
    {
      "epoch": 4.2085953878406706,
      "grad_norm": 0.7297657132148743,
      "learning_rate": 4.816638062493876e-05,
      "loss": 0.416,
      "num_input_tokens_seen": 5245040,
      "step": 8030
    },
    {
      "epoch": 4.211215932914046,
      "grad_norm": 0.4539424777030945,
      "learning_rate": 4.8162079913111716e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 5248112,
      "step": 8035
    },
    {
      "epoch": 4.213836477987422,
      "grad_norm": 0.36034032702445984,
      "learning_rate": 4.81577743560418e-05,
      "loss": 0.5427,
      "num_input_tokens_seen": 5251792,
      "step": 8040
    },
    {
      "epoch": 4.216457023060797,
      "grad_norm": 0.530923068523407,
      "learning_rate": 4.8153463954629684e-05,
      "loss": 0.6004,
      "num_input_tokens_seen": 5255120,
      "step": 8045
    },
    {
      "epoch": 4.219077568134172,
      "grad_norm": 0.43474698066711426,
      "learning_rate": 4.814914870977706e-05,
      "loss": 0.6357,
      "num_input_tokens_seen": 5259248,
      "step": 8050
    },
    {
      "epoch": 4.221698113207547,
      "grad_norm": 0.47328540682792664,
      "learning_rate": 4.8144828622386627e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 5262832,
      "step": 8055
    },
    {
      "epoch": 4.2243186582809225,
      "grad_norm": 0.5494180917739868,
      "learning_rate": 4.8140503693362095e-05,
      "loss": 0.568,
      "num_input_tokens_seen": 5266064,
      "step": 8060
    },
    {
      "epoch": 4.226939203354298,
      "grad_norm": 0.4031012952327728,
      "learning_rate": 4.813617392360818e-05,
      "loss": 0.5176,
      "num_input_tokens_seen": 5269008,
      "step": 8065
    },
    {
      "epoch": 4.229559748427673,
      "grad_norm": 0.4628792703151703,
      "learning_rate": 4.813183931403065e-05,
      "loss": 0.5624,
      "num_input_tokens_seen": 5272176,
      "step": 8070
    },
    {
      "epoch": 4.232180293501048,
      "grad_norm": 0.28659844398498535,
      "learning_rate": 4.8127499865536226e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 5276016,
      "step": 8075
    },
    {
      "epoch": 4.234800838574423,
      "grad_norm": 0.3385455012321472,
      "learning_rate": 4.812315557903269e-05,
      "loss": 0.5281,
      "num_input_tokens_seen": 5279024,
      "step": 8080
    },
    {
      "epoch": 4.237421383647799,
      "grad_norm": 0.4221336841583252,
      "learning_rate": 4.8118806455428806e-05,
      "loss": 0.5207,
      "num_input_tokens_seen": 5283088,
      "step": 8085
    },
    {
      "epoch": 4.240041928721174,
      "grad_norm": 0.923004686832428,
      "learning_rate": 4.811445249563437e-05,
      "loss": 0.6743,
      "num_input_tokens_seen": 5285936,
      "step": 8090
    },
    {
      "epoch": 4.2426624737945495,
      "grad_norm": 0.4021261930465698,
      "learning_rate": 4.811009370056017e-05,
      "loss": 0.3941,
      "num_input_tokens_seen": 5288912,
      "step": 8095
    },
    {
      "epoch": 4.245283018867925,
      "grad_norm": 0.7436296343803406,
      "learning_rate": 4.8105730071118016e-05,
      "loss": 0.6284,
      "num_input_tokens_seen": 5292912,
      "step": 8100
    },
    {
      "epoch": 4.2479035639413,
      "grad_norm": 0.35830986499786377,
      "learning_rate": 4.810136160822074e-05,
      "loss": 0.42,
      "num_input_tokens_seen": 5296368,
      "step": 8105
    },
    {
      "epoch": 4.250524109014675,
      "grad_norm": 0.6578747034072876,
      "learning_rate": 4.8096988312782174e-05,
      "loss": 0.5424,
      "num_input_tokens_seen": 5299408,
      "step": 8110
    },
    {
      "epoch": 4.25314465408805,
      "grad_norm": 0.7507877349853516,
      "learning_rate": 4.809261018571715e-05,
      "loss": 0.4347,
      "num_input_tokens_seen": 5302160,
      "step": 8115
    },
    {
      "epoch": 4.255765199161425,
      "grad_norm": 0.6283955574035645,
      "learning_rate": 4.808822722794153e-05,
      "loss": 0.4807,
      "num_input_tokens_seen": 5305296,
      "step": 8120
    },
    {
      "epoch": 4.2583857442348005,
      "grad_norm": 0.6558042764663696,
      "learning_rate": 4.8083839440372194e-05,
      "loss": 0.442,
      "num_input_tokens_seen": 5307888,
      "step": 8125
    },
    {
      "epoch": 4.261006289308176,
      "grad_norm": 0.40109673142433167,
      "learning_rate": 4.8079446823926984e-05,
      "loss": 0.4218,
      "num_input_tokens_seen": 5311024,
      "step": 8130
    },
    {
      "epoch": 4.263626834381552,
      "grad_norm": 0.37032729387283325,
      "learning_rate": 4.8075049379524805e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 5314512,
      "step": 8135
    },
    {
      "epoch": 4.266247379454927,
      "grad_norm": 0.7650788426399231,
      "learning_rate": 4.807064710808555e-05,
      "loss": 0.4855,
      "num_input_tokens_seen": 5317744,
      "step": 8140
    },
    {
      "epoch": 4.268867924528302,
      "grad_norm": 0.6534713506698608,
      "learning_rate": 4.806624001053013e-05,
      "loss": 0.4401,
      "num_input_tokens_seen": 5322640,
      "step": 8145
    },
    {
      "epoch": 4.271488469601677,
      "grad_norm": 1.1024054288864136,
      "learning_rate": 4.8061828087780455e-05,
      "loss": 0.5421,
      "num_input_tokens_seen": 5326192,
      "step": 8150
    },
    {
      "epoch": 4.274109014675052,
      "grad_norm": 0.288076788187027,
      "learning_rate": 4.805741134075944e-05,
      "loss": 0.5094,
      "num_input_tokens_seen": 5329104,
      "step": 8155
    },
    {
      "epoch": 4.276729559748428,
      "grad_norm": 0.35860922932624817,
      "learning_rate": 4.8052989770391026e-05,
      "loss": 0.4521,
      "num_input_tokens_seen": 5332368,
      "step": 8160
    },
    {
      "epoch": 4.279350104821803,
      "grad_norm": 0.9114435315132141,
      "learning_rate": 4.804856337760016e-05,
      "loss": 0.5419,
      "num_input_tokens_seen": 5336240,
      "step": 8165
    },
    {
      "epoch": 4.281970649895178,
      "grad_norm": 0.27088040113449097,
      "learning_rate": 4.804413216331278e-05,
      "loss": 0.4432,
      "num_input_tokens_seen": 5341104,
      "step": 8170
    },
    {
      "epoch": 4.284591194968553,
      "grad_norm": 0.4011000990867615,
      "learning_rate": 4.8039696128455856e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 5343440,
      "step": 8175
    },
    {
      "epoch": 4.287211740041929,
      "grad_norm": 0.3112547993659973,
      "learning_rate": 4.803525527395736e-05,
      "loss": 0.4486,
      "num_input_tokens_seen": 5347280,
      "step": 8180
    },
    {
      "epoch": 4.289832285115304,
      "grad_norm": 0.5548213720321655,
      "learning_rate": 4.803080960074625e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 5354512,
      "step": 8185
    },
    {
      "epoch": 4.2924528301886795,
      "grad_norm": 0.22279977798461914,
      "learning_rate": 4.8026359109752525e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 5357840,
      "step": 8190
    },
    {
      "epoch": 4.295073375262055,
      "grad_norm": 0.4075099527835846,
      "learning_rate": 4.8021903801907175e-05,
      "loss": 0.5718,
      "num_input_tokens_seen": 5362096,
      "step": 8195
    },
    {
      "epoch": 4.29769392033543,
      "grad_norm": 0.36661043763160706,
      "learning_rate": 4.801744367814219e-05,
      "loss": 0.531,
      "num_input_tokens_seen": 5364784,
      "step": 8200
    },
    {
      "epoch": 4.300314465408805,
      "grad_norm": 0.6038001179695129,
      "learning_rate": 4.801297873939059e-05,
      "loss": 0.4077,
      "num_input_tokens_seen": 5367504,
      "step": 8205
    },
    {
      "epoch": 4.30293501048218,
      "grad_norm": 0.602475643157959,
      "learning_rate": 4.8008508986586384e-05,
      "loss": 0.5446,
      "num_input_tokens_seen": 5370384,
      "step": 8210
    },
    {
      "epoch": 4.305555555555555,
      "grad_norm": 0.5583343505859375,
      "learning_rate": 4.800403442066459e-05,
      "loss": 0.4958,
      "num_input_tokens_seen": 5373264,
      "step": 8215
    },
    {
      "epoch": 4.3081761006289305,
      "grad_norm": 0.4542606770992279,
      "learning_rate": 4.799955504256125e-05,
      "loss": 0.4631,
      "num_input_tokens_seen": 5377040,
      "step": 8220
    },
    {
      "epoch": 4.310796645702306,
      "grad_norm": 0.7870923280715942,
      "learning_rate": 4.7995070853213376e-05,
      "loss": 0.5151,
      "num_input_tokens_seen": 5379888,
      "step": 8225
    },
    {
      "epoch": 4.313417190775682,
      "grad_norm": 1.1559805870056152,
      "learning_rate": 4.799058185355903e-05,
      "loss": 0.5882,
      "num_input_tokens_seen": 5382864,
      "step": 8230
    },
    {
      "epoch": 4.316037735849057,
      "grad_norm": 0.44580385088920593,
      "learning_rate": 4.7986088044537235e-05,
      "loss": 0.4502,
      "num_input_tokens_seen": 5386224,
      "step": 8235
    },
    {
      "epoch": 4.318658280922432,
      "grad_norm": 0.42789676785469055,
      "learning_rate": 4.798158942708807e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 5389104,
      "step": 8240
    },
    {
      "epoch": 4.321278825995807,
      "grad_norm": 0.4411965310573578,
      "learning_rate": 4.797708600215258e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 5391376,
      "step": 8245
    },
    {
      "epoch": 4.323899371069182,
      "grad_norm": 0.4188184142112732,
      "learning_rate": 4.7972577770672845e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 5394672,
      "step": 8250
    },
    {
      "epoch": 4.326519916142558,
      "grad_norm": 0.37256547808647156,
      "learning_rate": 4.796806473359192e-05,
      "loss": 0.4869,
      "num_input_tokens_seen": 5398096,
      "step": 8255
    },
    {
      "epoch": 4.329140461215933,
      "grad_norm": 0.6285741925239563,
      "learning_rate": 4.796354689185389e-05,
      "loss": 0.6008,
      "num_input_tokens_seen": 5400880,
      "step": 8260
    },
    {
      "epoch": 4.331761006289308,
      "grad_norm": 0.4555758237838745,
      "learning_rate": 4.795902424640383e-05,
      "loss": 0.5934,
      "num_input_tokens_seen": 5404336,
      "step": 8265
    },
    {
      "epoch": 4.334381551362683,
      "grad_norm": 0.5987867116928101,
      "learning_rate": 4.795449679818782e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 5408272,
      "step": 8270
    },
    {
      "epoch": 4.337002096436059,
      "grad_norm": 0.4903946816921234,
      "learning_rate": 4.794996454815297e-05,
      "loss": 0.5643,
      "num_input_tokens_seen": 5411824,
      "step": 8275
    },
    {
      "epoch": 4.339622641509434,
      "grad_norm": 0.9210709929466248,
      "learning_rate": 4.7945427497247355e-05,
      "loss": 0.4584,
      "num_input_tokens_seen": 5415344,
      "step": 8280
    },
    {
      "epoch": 4.3422431865828095,
      "grad_norm": 0.7456410527229309,
      "learning_rate": 4.7940885646420084e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 5418384,
      "step": 8285
    },
    {
      "epoch": 4.344863731656185,
      "grad_norm": 0.410455584526062,
      "learning_rate": 4.7936338996621264e-05,
      "loss": 0.4613,
      "num_input_tokens_seen": 5422192,
      "step": 8290
    },
    {
      "epoch": 4.34748427672956,
      "grad_norm": 0.2929849624633789,
      "learning_rate": 4.7931787548801996e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 5426224,
      "step": 8295
    },
    {
      "epoch": 4.350104821802935,
      "grad_norm": 0.5133172869682312,
      "learning_rate": 4.7927231303914394e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 5428816,
      "step": 8300
    },
    {
      "epoch": 4.35272536687631,
      "grad_norm": 0.8341405987739563,
      "learning_rate": 4.7922670262911565e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 5433712,
      "step": 8305
    },
    {
      "epoch": 4.355345911949685,
      "grad_norm": 0.42231300473213196,
      "learning_rate": 4.791810442674764e-05,
      "loss": 0.5699,
      "num_input_tokens_seen": 5436080,
      "step": 8310
    },
    {
      "epoch": 4.3579664570230605,
      "grad_norm": 0.5665462017059326,
      "learning_rate": 4.791353379637773e-05,
      "loss": 0.5017,
      "num_input_tokens_seen": 5438864,
      "step": 8315
    },
    {
      "epoch": 4.360587002096436,
      "grad_norm": 0.5270713567733765,
      "learning_rate": 4.790895837275796e-05,
      "loss": 0.3322,
      "num_input_tokens_seen": 5441488,
      "step": 8320
    },
    {
      "epoch": 4.363207547169811,
      "grad_norm": 0.34163305163383484,
      "learning_rate": 4.790437815684545e-05,
      "loss": 0.484,
      "num_input_tokens_seen": 5444752,
      "step": 8325
    },
    {
      "epoch": 4.365828092243187,
      "grad_norm": 0.6861332058906555,
      "learning_rate": 4.789979314959835e-05,
      "loss": 0.5327,
      "num_input_tokens_seen": 5447600,
      "step": 8330
    },
    {
      "epoch": 4.368448637316562,
      "grad_norm": 0.8132908940315247,
      "learning_rate": 4.789520335197577e-05,
      "loss": 0.472,
      "num_input_tokens_seen": 5450928,
      "step": 8335
    },
    {
      "epoch": 4.371069182389937,
      "grad_norm": 0.27777808904647827,
      "learning_rate": 4.789060876493785e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 5454192,
      "step": 8340
    },
    {
      "epoch": 4.373689727463312,
      "grad_norm": 0.29998818039894104,
      "learning_rate": 4.788600938944573e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 5457200,
      "step": 8345
    },
    {
      "epoch": 4.376310272536688,
      "grad_norm": 0.406820148229599,
      "learning_rate": 4.788140522646155e-05,
      "loss": 0.4692,
      "num_input_tokens_seen": 5460816,
      "step": 8350
    },
    {
      "epoch": 4.378930817610063,
      "grad_norm": 0.3106505572795868,
      "learning_rate": 4.787679627694843e-05,
      "loss": 0.5433,
      "num_input_tokens_seen": 5466256,
      "step": 8355
    },
    {
      "epoch": 4.381551362683438,
      "grad_norm": 0.486378014087677,
      "learning_rate": 4.787218254187053e-05,
      "loss": 0.5362,
      "num_input_tokens_seen": 5469520,
      "step": 8360
    },
    {
      "epoch": 4.384171907756813,
      "grad_norm": 0.42936912178993225,
      "learning_rate": 4.7867564022192965e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 5473200,
      "step": 8365
    },
    {
      "epoch": 4.386792452830189,
      "grad_norm": 0.8884363174438477,
      "learning_rate": 4.786294071888191e-05,
      "loss": 0.3946,
      "num_input_tokens_seen": 5475888,
      "step": 8370
    },
    {
      "epoch": 4.389412997903564,
      "grad_norm": 0.4982866942882538,
      "learning_rate": 4.785831263290449e-05,
      "loss": 0.3792,
      "num_input_tokens_seen": 5478224,
      "step": 8375
    },
    {
      "epoch": 4.3920335429769395,
      "grad_norm": 0.32591474056243896,
      "learning_rate": 4.7853679765228845e-05,
      "loss": 0.4789,
      "num_input_tokens_seen": 5481264,
      "step": 8380
    },
    {
      "epoch": 4.394654088050315,
      "grad_norm": 0.40352967381477356,
      "learning_rate": 4.7849042116824116e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 5485040,
      "step": 8385
    },
    {
      "epoch": 4.39727463312369,
      "grad_norm": 0.7648398876190186,
      "learning_rate": 4.7844399688660456e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 5488848,
      "step": 8390
    },
    {
      "epoch": 4.399895178197065,
      "grad_norm": 0.5489470958709717,
      "learning_rate": 4.783975248170901e-05,
      "loss": 0.5022,
      "num_input_tokens_seen": 5491504,
      "step": 8395
    },
    {
      "epoch": 4.40251572327044,
      "grad_norm": 0.3659225106239319,
      "learning_rate": 4.783510049694191e-05,
      "loss": 0.6302,
      "num_input_tokens_seen": 5494384,
      "step": 8400
    },
    {
      "epoch": 4.405136268343815,
      "grad_norm": 0.5347582101821899,
      "learning_rate": 4.78304437353323e-05,
      "loss": 0.6704,
      "num_input_tokens_seen": 5497712,
      "step": 8405
    },
    {
      "epoch": 4.4077568134171905,
      "grad_norm": 0.679892897605896,
      "learning_rate": 4.782578219785433e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 5500464,
      "step": 8410
    },
    {
      "epoch": 4.410377358490566,
      "grad_norm": 0.4353889226913452,
      "learning_rate": 4.782111588548312e-05,
      "loss": 0.3647,
      "num_input_tokens_seen": 5503632,
      "step": 8415
    },
    {
      "epoch": 4.412997903563941,
      "grad_norm": 0.2595607042312622,
      "learning_rate": 4.781644479919484e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 5506512,
      "step": 8420
    },
    {
      "epoch": 4.415618448637317,
      "grad_norm": 0.824734091758728,
      "learning_rate": 4.7811768939966605e-05,
      "loss": 0.3763,
      "num_input_tokens_seen": 5509648,
      "step": 8425
    },
    {
      "epoch": 4.418238993710692,
      "grad_norm": 0.542576789855957,
      "learning_rate": 4.780708830877656e-05,
      "loss": 0.484,
      "num_input_tokens_seen": 5512144,
      "step": 8430
    },
    {
      "epoch": 4.420859538784067,
      "grad_norm": 0.40118134021759033,
      "learning_rate": 4.780240290660384e-05,
      "loss": 0.5256,
      "num_input_tokens_seen": 5515312,
      "step": 8435
    },
    {
      "epoch": 4.423480083857442,
      "grad_norm": 0.37190210819244385,
      "learning_rate": 4.7797712734428575e-05,
      "loss": 0.5146,
      "num_input_tokens_seen": 5518000,
      "step": 8440
    },
    {
      "epoch": 4.426100628930818,
      "grad_norm": 0.29288408160209656,
      "learning_rate": 4.779301779323189e-05,
      "loss": 0.3607,
      "num_input_tokens_seen": 5521200,
      "step": 8445
    },
    {
      "epoch": 4.428721174004193,
      "grad_norm": 0.4331675171852112,
      "learning_rate": 4.778831808399593e-05,
      "loss": 0.6008,
      "num_input_tokens_seen": 5524336,
      "step": 8450
    },
    {
      "epoch": 4.431341719077568,
      "grad_norm": 0.9071370959281921,
      "learning_rate": 4.77836136077038e-05,
      "loss": 0.5043,
      "num_input_tokens_seen": 5527344,
      "step": 8455
    },
    {
      "epoch": 4.433962264150943,
      "grad_norm": 0.4974251985549927,
      "learning_rate": 4.777890436533963e-05,
      "loss": 0.3436,
      "num_input_tokens_seen": 5534256,
      "step": 8460
    },
    {
      "epoch": 4.436582809224318,
      "grad_norm": 0.3501285910606384,
      "learning_rate": 4.777419035788855e-05,
      "loss": 0.3724,
      "num_input_tokens_seen": 5537296,
      "step": 8465
    },
    {
      "epoch": 4.439203354297694,
      "grad_norm": 0.6278685927391052,
      "learning_rate": 4.776947158633667e-05,
      "loss": 0.5772,
      "num_input_tokens_seen": 5540656,
      "step": 8470
    },
    {
      "epoch": 4.4418238993710695,
      "grad_norm": 0.45992311835289,
      "learning_rate": 4.7764748051671095e-05,
      "loss": 0.4303,
      "num_input_tokens_seen": 5543632,
      "step": 8475
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 0.43831363320350647,
      "learning_rate": 4.7760019754879946e-05,
      "loss": 0.5477,
      "num_input_tokens_seen": 5546544,
      "step": 8480
    },
    {
      "epoch": 4.44706498951782,
      "grad_norm": 0.8365979790687561,
      "learning_rate": 4.7755286696952324e-05,
      "loss": 0.4885,
      "num_input_tokens_seen": 5549168,
      "step": 8485
    },
    {
      "epoch": 4.449685534591195,
      "grad_norm": 0.22963914275169373,
      "learning_rate": 4.775054887887833e-05,
      "loss": 0.3674,
      "num_input_tokens_seen": 5553264,
      "step": 8490
    },
    {
      "epoch": 4.45230607966457,
      "grad_norm": 0.3046753704547882,
      "learning_rate": 4.774580630164906e-05,
      "loss": 0.4731,
      "num_input_tokens_seen": 5556144,
      "step": 8495
    },
    {
      "epoch": 4.454926624737945,
      "grad_norm": 0.5294002294540405,
      "learning_rate": 4.774105896625661e-05,
      "loss": 0.3448,
      "num_input_tokens_seen": 5558832,
      "step": 8500
    },
    {
      "epoch": 4.4575471698113205,
      "grad_norm": 0.31583261489868164,
      "learning_rate": 4.773630687369407e-05,
      "loss": 0.491,
      "num_input_tokens_seen": 5561968,
      "step": 8505
    },
    {
      "epoch": 4.460167714884696,
      "grad_norm": 0.3932648301124573,
      "learning_rate": 4.773155002495552e-05,
      "loss": 0.4381,
      "num_input_tokens_seen": 5565456,
      "step": 8510
    },
    {
      "epoch": 4.462788259958071,
      "grad_norm": 0.4235745668411255,
      "learning_rate": 4.772678842103605e-05,
      "loss": 0.5095,
      "num_input_tokens_seen": 5568144,
      "step": 8515
    },
    {
      "epoch": 4.465408805031447,
      "grad_norm": 0.36548495292663574,
      "learning_rate": 4.772202206293172e-05,
      "loss": 0.5407,
      "num_input_tokens_seen": 5572080,
      "step": 8520
    },
    {
      "epoch": 4.468029350104822,
      "grad_norm": 0.5035116076469421,
      "learning_rate": 4.7717250951639586e-05,
      "loss": 0.4591,
      "num_input_tokens_seen": 5574992,
      "step": 8525
    },
    {
      "epoch": 4.470649895178197,
      "grad_norm": 0.46204906702041626,
      "learning_rate": 4.771247508815774e-05,
      "loss": 0.3929,
      "num_input_tokens_seen": 5578000,
      "step": 8530
    },
    {
      "epoch": 4.473270440251572,
      "grad_norm": 0.30242404341697693,
      "learning_rate": 4.770769447348522e-05,
      "loss": 0.4527,
      "num_input_tokens_seen": 5581200,
      "step": 8535
    },
    {
      "epoch": 4.475890985324948,
      "grad_norm": 0.45549827814102173,
      "learning_rate": 4.7702909108622074e-05,
      "loss": 0.4182,
      "num_input_tokens_seen": 5584336,
      "step": 8540
    },
    {
      "epoch": 4.478511530398323,
      "grad_norm": 0.5040175318717957,
      "learning_rate": 4.7698118994569355e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 5587440,
      "step": 8545
    },
    {
      "epoch": 4.481132075471698,
      "grad_norm": 0.7594400644302368,
      "learning_rate": 4.769332413232909e-05,
      "loss": 0.6426,
      "num_input_tokens_seen": 5590192,
      "step": 8550
    },
    {
      "epoch": 4.483752620545073,
      "grad_norm": 0.34875866770744324,
      "learning_rate": 4.768852452290432e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 5593712,
      "step": 8555
    },
    {
      "epoch": 4.486373165618448,
      "grad_norm": 0.4526386857032776,
      "learning_rate": 4.768372016729907e-05,
      "loss": 0.5283,
      "num_input_tokens_seen": 5596624,
      "step": 8560
    },
    {
      "epoch": 4.488993710691824,
      "grad_norm": 0.5685678124427795,
      "learning_rate": 4.767891106651834e-05,
      "loss": 0.4675,
      "num_input_tokens_seen": 5599536,
      "step": 8565
    },
    {
      "epoch": 4.4916142557651995,
      "grad_norm": 0.2812809944152832,
      "learning_rate": 4.767409722156816e-05,
      "loss": 0.462,
      "num_input_tokens_seen": 5603056,
      "step": 8570
    },
    {
      "epoch": 4.494234800838575,
      "grad_norm": 0.5637115836143494,
      "learning_rate": 4.7669278633455515e-05,
      "loss": 0.5074,
      "num_input_tokens_seen": 5605968,
      "step": 8575
    },
    {
      "epoch": 4.49685534591195,
      "grad_norm": 0.6973444819450378,
      "learning_rate": 4.7664455303188404e-05,
      "loss": 0.5688,
      "num_input_tokens_seen": 5608656,
      "step": 8580
    },
    {
      "epoch": 4.499475890985325,
      "grad_norm": 0.5465814471244812,
      "learning_rate": 4.7659627231775815e-05,
      "loss": 0.5217,
      "num_input_tokens_seen": 5612176,
      "step": 8585
    },
    {
      "epoch": 4.5020964360587,
      "grad_norm": 0.4102930426597595,
      "learning_rate": 4.765479442022773e-05,
      "loss": 0.3925,
      "num_input_tokens_seen": 5615728,
      "step": 8590
    },
    {
      "epoch": 4.504716981132075,
      "grad_norm": 0.5738059282302856,
      "learning_rate": 4.76499568695551e-05,
      "loss": 0.54,
      "num_input_tokens_seen": 5618448,
      "step": 8595
    },
    {
      "epoch": 4.5073375262054505,
      "grad_norm": 0.2731757164001465,
      "learning_rate": 4.764511458076991e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 5622256,
      "step": 8600
    },
    {
      "epoch": 4.509958071278826,
      "grad_norm": 0.3578195571899414,
      "learning_rate": 4.7640267554885085e-05,
      "loss": 0.5048,
      "num_input_tokens_seen": 5625264,
      "step": 8605
    },
    {
      "epoch": 4.512578616352201,
      "grad_norm": 0.33405110239982605,
      "learning_rate": 4.763541579291459e-05,
      "loss": 0.5219,
      "num_input_tokens_seen": 5627888,
      "step": 8610
    },
    {
      "epoch": 4.515199161425577,
      "grad_norm": 0.32741138339042664,
      "learning_rate": 4.763055929587336e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 5632208,
      "step": 8615
    },
    {
      "epoch": 4.517819706498952,
      "grad_norm": 0.4513361155986786,
      "learning_rate": 4.7625698064777294e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 5634384,
      "step": 8620
    },
    {
      "epoch": 4.520440251572327,
      "grad_norm": 0.3968472480773926,
      "learning_rate": 4.762083210064332e-05,
      "loss": 0.5046,
      "num_input_tokens_seen": 5636944,
      "step": 8625
    },
    {
      "epoch": 4.523060796645702,
      "grad_norm": 0.2813452184200287,
      "learning_rate": 4.761596140448934e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 5640080,
      "step": 8630
    },
    {
      "epoch": 4.5256813417190775,
      "grad_norm": 0.4489630460739136,
      "learning_rate": 4.7611085977334256e-05,
      "loss": 0.5225,
      "num_input_tokens_seen": 5643408,
      "step": 8635
    },
    {
      "epoch": 4.528301886792453,
      "grad_norm": 0.775202751159668,
      "learning_rate": 4.7606205820197946e-05,
      "loss": 0.5602,
      "num_input_tokens_seen": 5647216,
      "step": 8640
    },
    {
      "epoch": 4.530922431865828,
      "grad_norm": 0.8418669700622559,
      "learning_rate": 4.760132093410129e-05,
      "loss": 0.4879,
      "num_input_tokens_seen": 5651024,
      "step": 8645
    },
    {
      "epoch": 4.533542976939203,
      "grad_norm": 0.564548134803772,
      "learning_rate": 4.759643132006614e-05,
      "loss": 0.438,
      "num_input_tokens_seen": 5654032,
      "step": 8650
    },
    {
      "epoch": 4.536163522012579,
      "grad_norm": 0.4607052803039551,
      "learning_rate": 4.759153697911535e-05,
      "loss": 0.3889,
      "num_input_tokens_seen": 5656720,
      "step": 8655
    },
    {
      "epoch": 4.538784067085954,
      "grad_norm": 0.465580016374588,
      "learning_rate": 4.758663791227276e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 5658992,
      "step": 8660
    },
    {
      "epoch": 4.5414046121593294,
      "grad_norm": 0.3898661732673645,
      "learning_rate": 4.7581734120563206e-05,
      "loss": 0.2969,
      "num_input_tokens_seen": 5662448,
      "step": 8665
    },
    {
      "epoch": 4.544025157232705,
      "grad_norm": 0.4867733418941498,
      "learning_rate": 4.757682560501249e-05,
      "loss": 0.6149,
      "num_input_tokens_seen": 5665648,
      "step": 8670
    },
    {
      "epoch": 4.54664570230608,
      "grad_norm": 0.8633899092674255,
      "learning_rate": 4.757191236664744e-05,
      "loss": 0.5437,
      "num_input_tokens_seen": 5668976,
      "step": 8675
    },
    {
      "epoch": 4.549266247379455,
      "grad_norm": 0.23752596974372864,
      "learning_rate": 4.7566994406495825e-05,
      "loss": 0.3549,
      "num_input_tokens_seen": 5672272,
      "step": 8680
    },
    {
      "epoch": 4.55188679245283,
      "grad_norm": 0.47226595878601074,
      "learning_rate": 4.756207172558645e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 5675056,
      "step": 8685
    },
    {
      "epoch": 4.554507337526205,
      "grad_norm": 0.33837923407554626,
      "learning_rate": 4.7557144324949065e-05,
      "loss": 0.4283,
      "num_input_tokens_seen": 5677808,
      "step": 8690
    },
    {
      "epoch": 4.5571278825995805,
      "grad_norm": 0.3726891577243805,
      "learning_rate": 4.755221220561443e-05,
      "loss": 0.4662,
      "num_input_tokens_seen": 5681808,
      "step": 8695
    },
    {
      "epoch": 4.559748427672956,
      "grad_norm": 0.3443498909473419,
      "learning_rate": 4.7547275368614295e-05,
      "loss": 0.522,
      "num_input_tokens_seen": 5684688,
      "step": 8700
    },
    {
      "epoch": 4.562368972746331,
      "grad_norm": 0.33974531292915344,
      "learning_rate": 4.754233381498139e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 5687792,
      "step": 8705
    },
    {
      "epoch": 4.564989517819707,
      "grad_norm": 0.4543100595474243,
      "learning_rate": 4.753738754574942e-05,
      "loss": 0.5659,
      "num_input_tokens_seen": 5690352,
      "step": 8710
    },
    {
      "epoch": 4.567610062893082,
      "grad_norm": 0.40207192301750183,
      "learning_rate": 4.75324365619531e-05,
      "loss": 0.5522,
      "num_input_tokens_seen": 5693232,
      "step": 8715
    },
    {
      "epoch": 4.570230607966457,
      "grad_norm": 0.5191853642463684,
      "learning_rate": 4.752748086462811e-05,
      "loss": 0.4099,
      "num_input_tokens_seen": 5695792,
      "step": 8720
    },
    {
      "epoch": 4.572851153039832,
      "grad_norm": 0.8307037949562073,
      "learning_rate": 4.752252045481113e-05,
      "loss": 0.3963,
      "num_input_tokens_seen": 5699664,
      "step": 8725
    },
    {
      "epoch": 4.5754716981132075,
      "grad_norm": 0.4233620762825012,
      "learning_rate": 4.7517555333539834e-05,
      "loss": 0.3918,
      "num_input_tokens_seen": 5702832,
      "step": 8730
    },
    {
      "epoch": 4.578092243186583,
      "grad_norm": 0.3391372859477997,
      "learning_rate": 4.751258550185284e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 5705520,
      "step": 8735
    },
    {
      "epoch": 4.580712788259958,
      "grad_norm": 0.37946444749832153,
      "learning_rate": 4.7507610960789804e-05,
      "loss": 0.4757,
      "num_input_tokens_seen": 5709584,
      "step": 8740
    },
    {
      "epoch": 4.583333333333333,
      "grad_norm": 0.5317145586013794,
      "learning_rate": 4.7502631711391334e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 5713840,
      "step": 8745
    },
    {
      "epoch": 4.585953878406709,
      "grad_norm": 0.4404304623603821,
      "learning_rate": 4.749764775469904e-05,
      "loss": 0.5936,
      "num_input_tokens_seen": 5716016,
      "step": 8750
    },
    {
      "epoch": 4.588574423480084,
      "grad_norm": 0.3281893730163574,
      "learning_rate": 4.749265909175549e-05,
      "loss": 0.43,
      "num_input_tokens_seen": 5718640,
      "step": 8755
    },
    {
      "epoch": 4.591194968553459,
      "grad_norm": 0.350597620010376,
      "learning_rate": 4.7487665723604275e-05,
      "loss": 0.374,
      "num_input_tokens_seen": 5722448,
      "step": 8760
    },
    {
      "epoch": 4.593815513626835,
      "grad_norm": 0.47048816084861755,
      "learning_rate": 4.7482667651289944e-05,
      "loss": 0.4286,
      "num_input_tokens_seen": 5725520,
      "step": 8765
    },
    {
      "epoch": 4.59643605870021,
      "grad_norm": 0.538112461566925,
      "learning_rate": 4.747766487585803e-05,
      "loss": 0.4927,
      "num_input_tokens_seen": 5728144,
      "step": 8770
    },
    {
      "epoch": 4.599056603773585,
      "grad_norm": 0.5709526538848877,
      "learning_rate": 4.747265739835507e-05,
      "loss": 0.5659,
      "num_input_tokens_seen": 5731280,
      "step": 8775
    },
    {
      "epoch": 4.60167714884696,
      "grad_norm": 0.5022927522659302,
      "learning_rate": 4.746764521982856e-05,
      "loss": 0.6083,
      "num_input_tokens_seen": 5734096,
      "step": 8780
    },
    {
      "epoch": 4.604297693920335,
      "grad_norm": 0.31947726011276245,
      "learning_rate": 4.7462628341326995e-05,
      "loss": 0.4626,
      "num_input_tokens_seen": 5737360,
      "step": 8785
    },
    {
      "epoch": 4.6069182389937104,
      "grad_norm": 1.0670558214187622,
      "learning_rate": 4.7457606763899856e-05,
      "loss": 0.4767,
      "num_input_tokens_seen": 5740432,
      "step": 8790
    },
    {
      "epoch": 4.609538784067086,
      "grad_norm": 0.3680132329463959,
      "learning_rate": 4.745258048859758e-05,
      "loss": 0.4739,
      "num_input_tokens_seen": 5745552,
      "step": 8795
    },
    {
      "epoch": 4.612159329140461,
      "grad_norm": 0.5160056352615356,
      "learning_rate": 4.744754951647162e-05,
      "loss": 0.462,
      "num_input_tokens_seen": 5749712,
      "step": 8800
    },
    {
      "epoch": 4.614779874213837,
      "grad_norm": 0.3592943251132965,
      "learning_rate": 4.74425138485744e-05,
      "loss": 0.4281,
      "num_input_tokens_seen": 5753648,
      "step": 8805
    },
    {
      "epoch": 4.617400419287212,
      "grad_norm": 0.22571077942848206,
      "learning_rate": 4.743747348595932e-05,
      "loss": 0.5588,
      "num_input_tokens_seen": 5757008,
      "step": 8810
    },
    {
      "epoch": 4.620020964360587,
      "grad_norm": 0.2504875063896179,
      "learning_rate": 4.743242842968077e-05,
      "loss": 0.4093,
      "num_input_tokens_seen": 5760816,
      "step": 8815
    },
    {
      "epoch": 4.622641509433962,
      "grad_norm": 0.2980923354625702,
      "learning_rate": 4.742737868079411e-05,
      "loss": 0.7454,
      "num_input_tokens_seen": 5764080,
      "step": 8820
    },
    {
      "epoch": 4.6252620545073375,
      "grad_norm": 0.5467278957366943,
      "learning_rate": 4.7422324240355696e-05,
      "loss": 0.5813,
      "num_input_tokens_seen": 5767728,
      "step": 8825
    },
    {
      "epoch": 4.627882599580713,
      "grad_norm": 0.32057079672813416,
      "learning_rate": 4.741726510942286e-05,
      "loss": 0.5421,
      "num_input_tokens_seen": 5771568,
      "step": 8830
    },
    {
      "epoch": 4.630503144654088,
      "grad_norm": 0.40132221579551697,
      "learning_rate": 4.741220128905391e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 5774576,
      "step": 8835
    },
    {
      "epoch": 4.633123689727463,
      "grad_norm": 0.5721356868743896,
      "learning_rate": 4.740713278030815e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 5778480,
      "step": 8840
    },
    {
      "epoch": 4.635744234800838,
      "grad_norm": 0.3350412845611572,
      "learning_rate": 4.740205958424584e-05,
      "loss": 0.4176,
      "num_input_tokens_seen": 5782160,
      "step": 8845
    },
    {
      "epoch": 4.638364779874214,
      "grad_norm": 0.9622722268104553,
      "learning_rate": 4.739698170192824e-05,
      "loss": 0.5765,
      "num_input_tokens_seen": 5784560,
      "step": 8850
    },
    {
      "epoch": 4.640985324947589,
      "grad_norm": 0.632438063621521,
      "learning_rate": 4.7391899134417596e-05,
      "loss": 0.5107,
      "num_input_tokens_seen": 5787984,
      "step": 8855
    },
    {
      "epoch": 4.643605870020965,
      "grad_norm": 0.40031033754348755,
      "learning_rate": 4.73868118827771e-05,
      "loss": 0.5945,
      "num_input_tokens_seen": 5791056,
      "step": 8860
    },
    {
      "epoch": 4.64622641509434,
      "grad_norm": 0.44284147024154663,
      "learning_rate": 4.738171994807097e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 5793456,
      "step": 8865
    },
    {
      "epoch": 4.648846960167715,
      "grad_norm": 0.8362407088279724,
      "learning_rate": 4.7376623331364364e-05,
      "loss": 0.4358,
      "num_input_tokens_seen": 5797104,
      "step": 8870
    },
    {
      "epoch": 4.65146750524109,
      "grad_norm": 0.44656994938850403,
      "learning_rate": 4.737152203372345e-05,
      "loss": 0.4629,
      "num_input_tokens_seen": 5800048,
      "step": 8875
    },
    {
      "epoch": 4.654088050314465,
      "grad_norm": 0.7754992842674255,
      "learning_rate": 4.7366416056215345e-05,
      "loss": 0.5243,
      "num_input_tokens_seen": 5803600,
      "step": 8880
    },
    {
      "epoch": 4.65670859538784,
      "grad_norm": 0.5147721171379089,
      "learning_rate": 4.7361305399908185e-05,
      "loss": 0.7288,
      "num_input_tokens_seen": 5805776,
      "step": 8885
    },
    {
      "epoch": 4.659329140461216,
      "grad_norm": 0.6647140979766846,
      "learning_rate": 4.735619006587103e-05,
      "loss": 0.469,
      "num_input_tokens_seen": 5808560,
      "step": 8890
    },
    {
      "epoch": 4.661949685534591,
      "grad_norm": 0.36335301399230957,
      "learning_rate": 4.735107005517398e-05,
      "loss": 0.4571,
      "num_input_tokens_seen": 5812048,
      "step": 8895
    },
    {
      "epoch": 4.664570230607967,
      "grad_norm": 0.3966279625892639,
      "learning_rate": 4.734594536888806e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 5815856,
      "step": 8900
    },
    {
      "epoch": 4.667190775681342,
      "grad_norm": 0.36344826221466064,
      "learning_rate": 4.734081600808531e-05,
      "loss": 0.3688,
      "num_input_tokens_seen": 5820208,
      "step": 8905
    },
    {
      "epoch": 4.669811320754717,
      "grad_norm": 0.273205041885376,
      "learning_rate": 4.733568197383872e-05,
      "loss": 0.5204,
      "num_input_tokens_seen": 5823888,
      "step": 8910
    },
    {
      "epoch": 4.672431865828092,
      "grad_norm": 0.38349711894989014,
      "learning_rate": 4.7330543267222294e-05,
      "loss": 0.3563,
      "num_input_tokens_seen": 5826832,
      "step": 8915
    },
    {
      "epoch": 4.6750524109014675,
      "grad_norm": 0.8041329979896545,
      "learning_rate": 4.732539988931097e-05,
      "loss": 0.4286,
      "num_input_tokens_seen": 5831344,
      "step": 8920
    },
    {
      "epoch": 4.677672955974843,
      "grad_norm": 0.5235361456871033,
      "learning_rate": 4.732025184118068e-05,
      "loss": 0.5139,
      "num_input_tokens_seen": 5835824,
      "step": 8925
    },
    {
      "epoch": 4.680293501048218,
      "grad_norm": 0.7007504105567932,
      "learning_rate": 4.731509912390837e-05,
      "loss": 0.6165,
      "num_input_tokens_seen": 5838192,
      "step": 8930
    },
    {
      "epoch": 4.682914046121593,
      "grad_norm": 0.5168800354003906,
      "learning_rate": 4.730994173857189e-05,
      "loss": 0.5953,
      "num_input_tokens_seen": 5841008,
      "step": 8935
    },
    {
      "epoch": 4.685534591194968,
      "grad_norm": 0.8330833911895752,
      "learning_rate": 4.7304779686250126e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 5844592,
      "step": 8940
    },
    {
      "epoch": 4.688155136268344,
      "grad_norm": 0.3945862948894501,
      "learning_rate": 4.729961296802292e-05,
      "loss": 0.4255,
      "num_input_tokens_seen": 5850256,
      "step": 8945
    },
    {
      "epoch": 4.690775681341719,
      "grad_norm": 0.9112400412559509,
      "learning_rate": 4.7294441584971095e-05,
      "loss": 0.4017,
      "num_input_tokens_seen": 5853232,
      "step": 8950
    },
    {
      "epoch": 4.693396226415095,
      "grad_norm": 0.39039909839630127,
      "learning_rate": 4.728926553817643e-05,
      "loss": 0.6163,
      "num_input_tokens_seen": 5857584,
      "step": 8955
    },
    {
      "epoch": 4.69601677148847,
      "grad_norm": 0.5669586658477783,
      "learning_rate": 4.728408482872171e-05,
      "loss": 0.5074,
      "num_input_tokens_seen": 5860592,
      "step": 8960
    },
    {
      "epoch": 4.698637316561845,
      "grad_norm": 0.33912214636802673,
      "learning_rate": 4.727889945769067e-05,
      "loss": 0.5902,
      "num_input_tokens_seen": 5864208,
      "step": 8965
    },
    {
      "epoch": 4.70125786163522,
      "grad_norm": 0.8612805008888245,
      "learning_rate": 4.727370942616803e-05,
      "loss": 0.4749,
      "num_input_tokens_seen": 5867728,
      "step": 8970
    },
    {
      "epoch": 4.703878406708595,
      "grad_norm": 0.5253639817237854,
      "learning_rate": 4.72685147352395e-05,
      "loss": 0.6139,
      "num_input_tokens_seen": 5871472,
      "step": 8975
    },
    {
      "epoch": 4.70649895178197,
      "grad_norm": 0.310272216796875,
      "learning_rate": 4.7263315385991735e-05,
      "loss": 0.3501,
      "num_input_tokens_seen": 5875152,
      "step": 8980
    },
    {
      "epoch": 4.709119496855346,
      "grad_norm": 0.4030962586402893,
      "learning_rate": 4.7258111379512384e-05,
      "loss": 0.4911,
      "num_input_tokens_seen": 5878192,
      "step": 8985
    },
    {
      "epoch": 4.711740041928721,
      "grad_norm": 0.3215034604072571,
      "learning_rate": 4.725290271689007e-05,
      "loss": 0.5172,
      "num_input_tokens_seen": 5881904,
      "step": 8990
    },
    {
      "epoch": 4.714360587002097,
      "grad_norm": 0.6962401866912842,
      "learning_rate": 4.724768939921438e-05,
      "loss": 0.5131,
      "num_input_tokens_seen": 5884560,
      "step": 8995
    },
    {
      "epoch": 4.716981132075472,
      "grad_norm": 0.6000973582267761,
      "learning_rate": 4.7242471427575884e-05,
      "loss": 0.4891,
      "num_input_tokens_seen": 5887344,
      "step": 9000
    },
    {
      "epoch": 4.719601677148847,
      "grad_norm": 0.36822229623794556,
      "learning_rate": 4.723724880306612e-05,
      "loss": 0.5638,
      "num_input_tokens_seen": 5891088,
      "step": 9005
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 0.29353851079940796,
      "learning_rate": 4.72320215267776e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 5894576,
      "step": 9010
    },
    {
      "epoch": 4.7248427672955975,
      "grad_norm": 0.560762345790863,
      "learning_rate": 4.722678959980381e-05,
      "loss": 0.5058,
      "num_input_tokens_seen": 5898128,
      "step": 9015
    },
    {
      "epoch": 4.727463312368973,
      "grad_norm": 0.5139216780662537,
      "learning_rate": 4.7221553023239216e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 5901328,
      "step": 9020
    },
    {
      "epoch": 4.730083857442348,
      "grad_norm": 0.33801591396331787,
      "learning_rate": 4.721631179817925e-05,
      "loss": 0.3626,
      "num_input_tokens_seen": 5904656,
      "step": 9025
    },
    {
      "epoch": 4.732704402515723,
      "grad_norm": 0.3960614800453186,
      "learning_rate": 4.721106592572031e-05,
      "loss": 0.4575,
      "num_input_tokens_seen": 5907728,
      "step": 9030
    },
    {
      "epoch": 4.735324947589098,
      "grad_norm": 0.5603753328323364,
      "learning_rate": 4.7205815406959774e-05,
      "loss": 0.4218,
      "num_input_tokens_seen": 5910544,
      "step": 9035
    },
    {
      "epoch": 4.737945492662474,
      "grad_norm": 0.3615403473377228,
      "learning_rate": 4.7200560242996e-05,
      "loss": 0.5343,
      "num_input_tokens_seen": 5914096,
      "step": 9040
    },
    {
      "epoch": 4.740566037735849,
      "grad_norm": 0.3768449127674103,
      "learning_rate": 4.719530043492829e-05,
      "loss": 0.378,
      "num_input_tokens_seen": 5916688,
      "step": 9045
    },
    {
      "epoch": 4.743186582809225,
      "grad_norm": 0.4402933120727539,
      "learning_rate": 4.7190035983856954e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 5919632,
      "step": 9050
    },
    {
      "epoch": 4.7458071278826,
      "grad_norm": 0.47334742546081543,
      "learning_rate": 4.718476689088325e-05,
      "loss": 0.5141,
      "num_input_tokens_seen": 5923280,
      "step": 9055
    },
    {
      "epoch": 4.748427672955975,
      "grad_norm": 0.39317792654037476,
      "learning_rate": 4.71794931571094e-05,
      "loss": 0.519,
      "num_input_tokens_seen": 5926384,
      "step": 9060
    },
    {
      "epoch": 4.75104821802935,
      "grad_norm": 0.6114296913146973,
      "learning_rate": 4.717421478363862e-05,
      "loss": 0.4966,
      "num_input_tokens_seen": 5929168,
      "step": 9065
    },
    {
      "epoch": 4.753668763102725,
      "grad_norm": 1.2914155721664429,
      "learning_rate": 4.7168931771575085e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 5931632,
      "step": 9070
    },
    {
      "epoch": 4.7562893081761,
      "grad_norm": 0.4285983741283417,
      "learning_rate": 4.716364412202394e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 5934736,
      "step": 9075
    },
    {
      "epoch": 4.758909853249476,
      "grad_norm": 0.49017518758773804,
      "learning_rate": 4.7158351836091306e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 5937616,
      "step": 9080
    },
    {
      "epoch": 4.761530398322851,
      "grad_norm": 0.44292664527893066,
      "learning_rate": 4.7153054914884266e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 5940592,
      "step": 9085
    },
    {
      "epoch": 4.764150943396227,
      "grad_norm": 0.3760435879230499,
      "learning_rate": 4.714775335951086e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 5943568,
      "step": 9090
    },
    {
      "epoch": 4.766771488469602,
      "grad_norm": 0.40099161863327026,
      "learning_rate": 4.7142447171080137e-05,
      "loss": 0.4204,
      "num_input_tokens_seen": 5946960,
      "step": 9095
    },
    {
      "epoch": 4.769392033542977,
      "grad_norm": 0.4855356812477112,
      "learning_rate": 4.713713635070208e-05,
      "loss": 0.4255,
      "num_input_tokens_seen": 5950512,
      "step": 9100
    },
    {
      "epoch": 4.772012578616352,
      "grad_norm": 0.5066821575164795,
      "learning_rate": 4.713182089948765e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 5954192,
      "step": 9105
    },
    {
      "epoch": 4.7746331236897275,
      "grad_norm": 0.3807922303676605,
      "learning_rate": 4.712650081854878e-05,
      "loss": 0.5432,
      "num_input_tokens_seen": 5957584,
      "step": 9110
    },
    {
      "epoch": 4.777253668763103,
      "grad_norm": 0.6022106409072876,
      "learning_rate": 4.712117610899837e-05,
      "loss": 0.5272,
      "num_input_tokens_seen": 5960432,
      "step": 9115
    },
    {
      "epoch": 4.779874213836478,
      "grad_norm": 0.4437089264392853,
      "learning_rate": 4.7115846771950286e-05,
      "loss": 0.4402,
      "num_input_tokens_seen": 5963984,
      "step": 9120
    },
    {
      "epoch": 4.782494758909853,
      "grad_norm": 0.40307074785232544,
      "learning_rate": 4.711051280851938e-05,
      "loss": 0.3519,
      "num_input_tokens_seen": 5967600,
      "step": 9125
    },
    {
      "epoch": 4.785115303983228,
      "grad_norm": 0.7747418284416199,
      "learning_rate": 4.7105174219821435e-05,
      "loss": 0.5553,
      "num_input_tokens_seen": 5970608,
      "step": 9130
    },
    {
      "epoch": 4.787735849056604,
      "grad_norm": 0.4153198003768921,
      "learning_rate": 4.709983100697323e-05,
      "loss": 0.4152,
      "num_input_tokens_seen": 5973840,
      "step": 9135
    },
    {
      "epoch": 4.790356394129979,
      "grad_norm": 0.5294312834739685,
      "learning_rate": 4.709448317109252e-05,
      "loss": 0.4164,
      "num_input_tokens_seen": 5977168,
      "step": 9140
    },
    {
      "epoch": 4.7929769392033545,
      "grad_norm": 0.7249862551689148,
      "learning_rate": 4.708913071329798e-05,
      "loss": 0.4746,
      "num_input_tokens_seen": 5980528,
      "step": 9145
    },
    {
      "epoch": 4.79559748427673,
      "grad_norm": 0.24611619114875793,
      "learning_rate": 4.708377363470932e-05,
      "loss": 0.6061,
      "num_input_tokens_seen": 5984368,
      "step": 9150
    },
    {
      "epoch": 4.798218029350105,
      "grad_norm": 0.4680712819099426,
      "learning_rate": 4.7078411936447144e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 5987568,
      "step": 9155
    },
    {
      "epoch": 4.80083857442348,
      "grad_norm": 0.6636074185371399,
      "learning_rate": 4.707304561963308e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 5990256,
      "step": 9160
    },
    {
      "epoch": 4.803459119496855,
      "grad_norm": 0.40234294533729553,
      "learning_rate": 4.70676746853897e-05,
      "loss": 0.5449,
      "num_input_tokens_seen": 5993680,
      "step": 9165
    },
    {
      "epoch": 4.80607966457023,
      "grad_norm": 0.3965575397014618,
      "learning_rate": 4.706229913484054e-05,
      "loss": 0.4939,
      "num_input_tokens_seen": 6001392,
      "step": 9170
    },
    {
      "epoch": 4.808700209643606,
      "grad_norm": 0.5227388739585876,
      "learning_rate": 4.705691896911009e-05,
      "loss": 0.4524,
      "num_input_tokens_seen": 6004240,
      "step": 9175
    },
    {
      "epoch": 4.811320754716981,
      "grad_norm": 1.2839981317520142,
      "learning_rate": 4.705153418932383e-05,
      "loss": 0.5256,
      "num_input_tokens_seen": 6008304,
      "step": 9180
    },
    {
      "epoch": 4.813941299790356,
      "grad_norm": 0.5032493472099304,
      "learning_rate": 4.7046144796608206e-05,
      "loss": 0.5211,
      "num_input_tokens_seen": 6011504,
      "step": 9185
    },
    {
      "epoch": 4.816561844863732,
      "grad_norm": 0.420076459646225,
      "learning_rate": 4.70407507920906e-05,
      "loss": 0.4642,
      "num_input_tokens_seen": 6014192,
      "step": 9190
    },
    {
      "epoch": 4.819182389937107,
      "grad_norm": 0.49349313974380493,
      "learning_rate": 4.7035352176899384e-05,
      "loss": 0.5729,
      "num_input_tokens_seen": 6016560,
      "step": 9195
    },
    {
      "epoch": 4.821802935010482,
      "grad_norm": 0.7430074214935303,
      "learning_rate": 4.7029948952163885e-05,
      "loss": 0.5866,
      "num_input_tokens_seen": 6019120,
      "step": 9200
    },
    {
      "epoch": 4.8244234800838575,
      "grad_norm": 0.4579275846481323,
      "learning_rate": 4.7024541119014386e-05,
      "loss": 0.475,
      "num_input_tokens_seen": 6022672,
      "step": 9205
    },
    {
      "epoch": 4.827044025157233,
      "grad_norm": 0.4043867886066437,
      "learning_rate": 4.701912867858217e-05,
      "loss": 0.3438,
      "num_input_tokens_seen": 6025520,
      "step": 9210
    },
    {
      "epoch": 4.829664570230608,
      "grad_norm": 0.45108795166015625,
      "learning_rate": 4.7013711631999434e-05,
      "loss": 0.4633,
      "num_input_tokens_seen": 6028208,
      "step": 9215
    },
    {
      "epoch": 4.832285115303983,
      "grad_norm": 0.50236576795578,
      "learning_rate": 4.700828998039937e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 6031760,
      "step": 9220
    },
    {
      "epoch": 4.834905660377358,
      "grad_norm": 0.28155717253685,
      "learning_rate": 4.700286372491612e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 6034128,
      "step": 9225
    },
    {
      "epoch": 4.837526205450734,
      "grad_norm": 0.4590338468551636,
      "learning_rate": 4.6997432866684816e-05,
      "loss": 0.5312,
      "num_input_tokens_seen": 6036976,
      "step": 9230
    },
    {
      "epoch": 4.840146750524109,
      "grad_norm": 0.5165165066719055,
      "learning_rate": 4.6991997406841504e-05,
      "loss": 0.6072,
      "num_input_tokens_seen": 6040272,
      "step": 9235
    },
    {
      "epoch": 4.8427672955974845,
      "grad_norm": 0.378032922744751,
      "learning_rate": 4.698655734652324e-05,
      "loss": 0.5158,
      "num_input_tokens_seen": 6042768,
      "step": 9240
    },
    {
      "epoch": 4.84538784067086,
      "grad_norm": 0.7842647433280945,
      "learning_rate": 4.698111268686801e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 6046672,
      "step": 9245
    },
    {
      "epoch": 4.848008385744235,
      "grad_norm": 0.3421376645565033,
      "learning_rate": 4.6975663429014785e-05,
      "loss": 0.4,
      "num_input_tokens_seen": 6049328,
      "step": 9250
    },
    {
      "epoch": 4.85062893081761,
      "grad_norm": 0.7278099656105042,
      "learning_rate": 4.697020957410347e-05,
      "loss": 0.4541,
      "num_input_tokens_seen": 6053168,
      "step": 9255
    },
    {
      "epoch": 4.853249475890985,
      "grad_norm": 0.3670884370803833,
      "learning_rate": 4.6964751123274974e-05,
      "loss": 0.523,
      "num_input_tokens_seen": 6056560,
      "step": 9260
    },
    {
      "epoch": 4.85587002096436,
      "grad_norm": 0.5382692813873291,
      "learning_rate": 4.695928807767113e-05,
      "loss": 0.3726,
      "num_input_tokens_seen": 6059856,
      "step": 9265
    },
    {
      "epoch": 4.8584905660377355,
      "grad_norm": 0.5207222104072571,
      "learning_rate": 4.695382043843474e-05,
      "loss": 0.5817,
      "num_input_tokens_seen": 6063280,
      "step": 9270
    },
    {
      "epoch": 4.861111111111111,
      "grad_norm": 0.569308876991272,
      "learning_rate": 4.694834820670958e-05,
      "loss": 0.5376,
      "num_input_tokens_seen": 6066256,
      "step": 9275
    },
    {
      "epoch": 4.863731656184486,
      "grad_norm": 0.7889625430107117,
      "learning_rate": 4.694287138364037e-05,
      "loss": 0.5057,
      "num_input_tokens_seen": 6069008,
      "step": 9280
    },
    {
      "epoch": 4.866352201257862,
      "grad_norm": 0.5172746777534485,
      "learning_rate": 4.6937389970372815e-05,
      "loss": 0.4758,
      "num_input_tokens_seen": 6072848,
      "step": 9285
    },
    {
      "epoch": 4.868972746331237,
      "grad_norm": 1.3106828927993774,
      "learning_rate": 4.693190396805354e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 6075536,
      "step": 9290
    },
    {
      "epoch": 4.871593291404612,
      "grad_norm": 0.3781812787055969,
      "learning_rate": 4.692641337783017e-05,
      "loss": 0.4471,
      "num_input_tokens_seen": 6079376,
      "step": 9295
    },
    {
      "epoch": 4.8742138364779874,
      "grad_norm": 0.2608051300048828,
      "learning_rate": 4.6920918200851274e-05,
      "loss": 0.3782,
      "num_input_tokens_seen": 6083056,
      "step": 9300
    },
    {
      "epoch": 4.876834381551363,
      "grad_norm": 0.5969725847244263,
      "learning_rate": 4.691541843826638e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 6085520,
      "step": 9305
    },
    {
      "epoch": 4.879454926624738,
      "grad_norm": 1.1134341955184937,
      "learning_rate": 4.690991409122597e-05,
      "loss": 0.5469,
      "num_input_tokens_seen": 6089040,
      "step": 9310
    },
    {
      "epoch": 4.882075471698113,
      "grad_norm": 0.4647405743598938,
      "learning_rate": 4.690440516088149e-05,
      "loss": 0.3575,
      "num_input_tokens_seen": 6092016,
      "step": 9315
    },
    {
      "epoch": 4.884696016771488,
      "grad_norm": 0.409601628780365,
      "learning_rate": 4.689889164838536e-05,
      "loss": 0.553,
      "num_input_tokens_seen": 6094608,
      "step": 9320
    },
    {
      "epoch": 4.887316561844864,
      "grad_norm": 0.5287918448448181,
      "learning_rate": 4.6893373554890925e-05,
      "loss": 0.3308,
      "num_input_tokens_seen": 6097520,
      "step": 9325
    },
    {
      "epoch": 4.889937106918239,
      "grad_norm": 0.5291329622268677,
      "learning_rate": 4.688785088155251e-05,
      "loss": 0.3738,
      "num_input_tokens_seen": 6100464,
      "step": 9330
    },
    {
      "epoch": 4.8925576519916145,
      "grad_norm": 0.4464346170425415,
      "learning_rate": 4.688232362952541e-05,
      "loss": 0.5192,
      "num_input_tokens_seen": 6104304,
      "step": 9335
    },
    {
      "epoch": 4.89517819706499,
      "grad_norm": 0.4251300096511841,
      "learning_rate": 4.687679179996585e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 6107856,
      "step": 9340
    },
    {
      "epoch": 4.897798742138365,
      "grad_norm": 0.42594045400619507,
      "learning_rate": 4.687125539403103e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 6110832,
      "step": 9345
    },
    {
      "epoch": 4.90041928721174,
      "grad_norm": 0.36519119143486023,
      "learning_rate": 4.68657144128791e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 6114576,
      "step": 9350
    },
    {
      "epoch": 4.903039832285115,
      "grad_norm": 0.45947179198265076,
      "learning_rate": 4.6860168857669175e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 6118512,
      "step": 9355
    },
    {
      "epoch": 4.90566037735849,
      "grad_norm": 0.386229932308197,
      "learning_rate": 4.6854618729561315e-05,
      "loss": 0.4209,
      "num_input_tokens_seen": 6121872,
      "step": 9360
    },
    {
      "epoch": 4.9082809224318655,
      "grad_norm": 0.34352418780326843,
      "learning_rate": 4.684906402971655e-05,
      "loss": 0.5859,
      "num_input_tokens_seen": 6125520,
      "step": 9365
    },
    {
      "epoch": 4.910901467505241,
      "grad_norm": 0.23365047574043274,
      "learning_rate": 4.684350475929685e-05,
      "loss": 0.4478,
      "num_input_tokens_seen": 6129936,
      "step": 9370
    },
    {
      "epoch": 4.913522012578616,
      "grad_norm": 0.6903394460678101,
      "learning_rate": 4.683794091946517e-05,
      "loss": 0.6964,
      "num_input_tokens_seen": 6132656,
      "step": 9375
    },
    {
      "epoch": 4.916142557651992,
      "grad_norm": 0.5516082644462585,
      "learning_rate": 4.683237251138538e-05,
      "loss": 0.4472,
      "num_input_tokens_seen": 6136528,
      "step": 9380
    },
    {
      "epoch": 4.918763102725367,
      "grad_norm": 0.3615407347679138,
      "learning_rate": 4.6826799536222346e-05,
      "loss": 0.4216,
      "num_input_tokens_seen": 6139248,
      "step": 9385
    },
    {
      "epoch": 4.921383647798742,
      "grad_norm": 0.39389699697494507,
      "learning_rate": 4.6821221995141865e-05,
      "loss": 0.4721,
      "num_input_tokens_seen": 6142864,
      "step": 9390
    },
    {
      "epoch": 4.924004192872117,
      "grad_norm": 0.40360867977142334,
      "learning_rate": 4.681563988931068e-05,
      "loss": 0.6894,
      "num_input_tokens_seen": 6146128,
      "step": 9395
    },
    {
      "epoch": 4.926624737945493,
      "grad_norm": 0.5281766057014465,
      "learning_rate": 4.681005321989652e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 6148976,
      "step": 9400
    },
    {
      "epoch": 4.929245283018868,
      "grad_norm": 0.42039987444877625,
      "learning_rate": 4.6804461988068055e-05,
      "loss": 0.6264,
      "num_input_tokens_seen": 6151952,
      "step": 9405
    },
    {
      "epoch": 4.931865828092243,
      "grad_norm": 0.5617814660072327,
      "learning_rate": 4.6798866194994896e-05,
      "loss": 0.6001,
      "num_input_tokens_seen": 6155376,
      "step": 9410
    },
    {
      "epoch": 4.934486373165618,
      "grad_norm": 1.3674933910369873,
      "learning_rate": 4.6793265841847624e-05,
      "loss": 0.5228,
      "num_input_tokens_seen": 6157936,
      "step": 9415
    },
    {
      "epoch": 4.937106918238994,
      "grad_norm": 0.5144702792167664,
      "learning_rate": 4.6787660929797766e-05,
      "loss": 0.4736,
      "num_input_tokens_seen": 6163632,
      "step": 9420
    },
    {
      "epoch": 4.939727463312369,
      "grad_norm": 0.30545884370803833,
      "learning_rate": 4.678205146001781e-05,
      "loss": 0.4494,
      "num_input_tokens_seen": 6166800,
      "step": 9425
    },
    {
      "epoch": 4.9423480083857445,
      "grad_norm": 0.4872092604637146,
      "learning_rate": 4.67764374336812e-05,
      "loss": 0.4879,
      "num_input_tokens_seen": 6170384,
      "step": 9430
    },
    {
      "epoch": 4.94496855345912,
      "grad_norm": 0.2853154242038727,
      "learning_rate": 4.6770818851962305e-05,
      "loss": 0.414,
      "num_input_tokens_seen": 6173488,
      "step": 9435
    },
    {
      "epoch": 4.947589098532495,
      "grad_norm": 0.4046965539455414,
      "learning_rate": 4.676519571603648e-05,
      "loss": 0.4536,
      "num_input_tokens_seen": 6176560,
      "step": 9440
    },
    {
      "epoch": 4.95020964360587,
      "grad_norm": 0.22473680973052979,
      "learning_rate": 4.6759568027080034e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 6180496,
      "step": 9445
    },
    {
      "epoch": 4.952830188679245,
      "grad_norm": 0.6450782418251038,
      "learning_rate": 4.67539357862702e-05,
      "loss": 0.5125,
      "num_input_tokens_seen": 6183344,
      "step": 9450
    },
    {
      "epoch": 4.95545073375262,
      "grad_norm": 0.3543927073478699,
      "learning_rate": 4.6748298994785175e-05,
      "loss": 0.4237,
      "num_input_tokens_seen": 6186320,
      "step": 9455
    },
    {
      "epoch": 4.9580712788259955,
      "grad_norm": 0.469735324382782,
      "learning_rate": 4.6742657653804116e-05,
      "loss": 0.4362,
      "num_input_tokens_seen": 6189904,
      "step": 9460
    },
    {
      "epoch": 4.960691823899371,
      "grad_norm": 0.36466720700263977,
      "learning_rate": 4.673701176450713e-05,
      "loss": 0.5521,
      "num_input_tokens_seen": 6194000,
      "step": 9465
    },
    {
      "epoch": 4.963312368972746,
      "grad_norm": 0.5218120813369751,
      "learning_rate": 4.6731361328075286e-05,
      "loss": 0.4425,
      "num_input_tokens_seen": 6197040,
      "step": 9470
    },
    {
      "epoch": 4.965932914046122,
      "grad_norm": 0.28805452585220337,
      "learning_rate": 4.672570634569056e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 6200944,
      "step": 9475
    },
    {
      "epoch": 4.968553459119497,
      "grad_norm": 1.0786086320877075,
      "learning_rate": 4.672004681853593e-05,
      "loss": 0.4916,
      "num_input_tokens_seen": 6203216,
      "step": 9480
    },
    {
      "epoch": 4.971174004192872,
      "grad_norm": 0.44788414239883423,
      "learning_rate": 4.6714382747795304e-05,
      "loss": 0.4078,
      "num_input_tokens_seen": 6206384,
      "step": 9485
    },
    {
      "epoch": 4.973794549266247,
      "grad_norm": 1.0676918029785156,
      "learning_rate": 4.670871413465354e-05,
      "loss": 0.5098,
      "num_input_tokens_seen": 6209872,
      "step": 9490
    },
    {
      "epoch": 4.976415094339623,
      "grad_norm": 0.39524009823799133,
      "learning_rate": 4.670304098029645e-05,
      "loss": 0.5249,
      "num_input_tokens_seen": 6213328,
      "step": 9495
    },
    {
      "epoch": 4.979035639412998,
      "grad_norm": 0.5363308787345886,
      "learning_rate": 4.669736328591078e-05,
      "loss": 0.458,
      "num_input_tokens_seen": 6216784,
      "step": 9500
    },
    {
      "epoch": 4.981656184486373,
      "grad_norm": 0.500468909740448,
      "learning_rate": 4.669168105268424e-05,
      "loss": 0.501,
      "num_input_tokens_seen": 6220016,
      "step": 9505
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.3878590166568756,
      "learning_rate": 4.6685994281805514e-05,
      "loss": 0.4075,
      "num_input_tokens_seen": 6222672,
      "step": 9510
    },
    {
      "epoch": 4.986897274633124,
      "grad_norm": 0.3355172872543335,
      "learning_rate": 4.6680302974464175e-05,
      "loss": 0.4674,
      "num_input_tokens_seen": 6226224,
      "step": 9515
    },
    {
      "epoch": 4.989517819706499,
      "grad_norm": 0.5763694643974304,
      "learning_rate": 4.66746071318508e-05,
      "loss": 0.4021,
      "num_input_tokens_seen": 6228688,
      "step": 9520
    },
    {
      "epoch": 4.9921383647798745,
      "grad_norm": 0.42635297775268555,
      "learning_rate": 4.6668906755156896e-05,
      "loss": 0.4417,
      "num_input_tokens_seen": 6231440,
      "step": 9525
    },
    {
      "epoch": 4.99475890985325,
      "grad_norm": 0.615874171257019,
      "learning_rate": 4.6663201845574906e-05,
      "loss": 0.584,
      "num_input_tokens_seen": 6234160,
      "step": 9530
    },
    {
      "epoch": 4.997379454926625,
      "grad_norm": 0.560550332069397,
      "learning_rate": 4.6657492404298234e-05,
      "loss": 0.3585,
      "num_input_tokens_seen": 6237232,
      "step": 9535
    },
    {
      "epoch": 5.0,
      "grad_norm": 22.37757110595703,
      "learning_rate": 4.665177843252124e-05,
      "loss": 1.1977,
      "num_input_tokens_seen": 6239608,
      "step": 9540
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.45809313654899597,
      "eval_runtime": 11.2984,
      "eval_samples_per_second": 75.055,
      "eval_steps_per_second": 18.764,
      "num_input_tokens_seen": 6239608,
      "step": 9540
    },
    {
      "epoch": 5.002620545073375,
      "grad_norm": 0.36859774589538574,
      "learning_rate": 4.664605993143921e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 6242936,
      "step": 9545
    },
    {
      "epoch": 5.00524109014675,
      "grad_norm": 0.5108821988105774,
      "learning_rate": 4.6640336902248396e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 6246392,
      "step": 9550
    },
    {
      "epoch": 5.0078616352201255,
      "grad_norm": 0.4653903841972351,
      "learning_rate": 4.6634609346145984e-05,
      "loss": 0.5594,
      "num_input_tokens_seen": 6250008,
      "step": 9555
    },
    {
      "epoch": 5.010482180293501,
      "grad_norm": 0.5098759531974792,
      "learning_rate": 4.6628877264330116e-05,
      "loss": 0.4406,
      "num_input_tokens_seen": 6253336,
      "step": 9560
    },
    {
      "epoch": 5.013102725366877,
      "grad_norm": 0.2830705940723419,
      "learning_rate": 4.6623140657999886e-05,
      "loss": 0.3742,
      "num_input_tokens_seen": 6256888,
      "step": 9565
    },
    {
      "epoch": 5.015723270440252,
      "grad_norm": 0.6150985956192017,
      "learning_rate": 4.661739952835532e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 6259768,
      "step": 9570
    },
    {
      "epoch": 5.018343815513627,
      "grad_norm": 1.3169240951538086,
      "learning_rate": 4.66116538765974e-05,
      "loss": 0.4993,
      "num_input_tokens_seen": 6262616,
      "step": 9575
    },
    {
      "epoch": 5.020964360587002,
      "grad_norm": 0.3150300681591034,
      "learning_rate": 4.6605903703928034e-05,
      "loss": 0.3483,
      "num_input_tokens_seen": 6265624,
      "step": 9580
    },
    {
      "epoch": 5.023584905660377,
      "grad_norm": 0.32883062958717346,
      "learning_rate": 4.6600149011550116e-05,
      "loss": 0.6221,
      "num_input_tokens_seen": 6269304,
      "step": 9585
    },
    {
      "epoch": 5.026205450733753,
      "grad_norm": 0.3093138039112091,
      "learning_rate": 4.659438980066745e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 6272312,
      "step": 9590
    },
    {
      "epoch": 5.028825995807128,
      "grad_norm": 0.6773216128349304,
      "learning_rate": 4.65886260724848e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 6274904,
      "step": 9595
    },
    {
      "epoch": 5.031446540880503,
      "grad_norm": 0.4958578050136566,
      "learning_rate": 4.658285782820787e-05,
      "loss": 0.4575,
      "num_input_tokens_seen": 6278104,
      "step": 9600
    },
    {
      "epoch": 5.034067085953878,
      "grad_norm": 0.30733683705329895,
      "learning_rate": 4.657708506904331e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 6281656,
      "step": 9605
    },
    {
      "epoch": 5.036687631027253,
      "grad_norm": 0.3287902772426605,
      "learning_rate": 4.6571307796198724e-05,
      "loss": 0.5548,
      "num_input_tokens_seen": 6285240,
      "step": 9610
    },
    {
      "epoch": 5.039308176100629,
      "grad_norm": 0.43997833132743835,
      "learning_rate": 4.656552601088265e-05,
      "loss": 0.5081,
      "num_input_tokens_seen": 6287768,
      "step": 9615
    },
    {
      "epoch": 5.0419287211740045,
      "grad_norm": 0.29283618927001953,
      "learning_rate": 4.655973971430456e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 6291128,
      "step": 9620
    },
    {
      "epoch": 5.04454926624738,
      "grad_norm": 0.4205421209335327,
      "learning_rate": 4.6553948907674895e-05,
      "loss": 0.2686,
      "num_input_tokens_seen": 6295416,
      "step": 9625
    },
    {
      "epoch": 5.047169811320755,
      "grad_norm": 0.40138474106788635,
      "learning_rate": 4.654815359220501e-05,
      "loss": 0.466,
      "num_input_tokens_seen": 6299000,
      "step": 9630
    },
    {
      "epoch": 5.04979035639413,
      "grad_norm": 0.5316153764724731,
      "learning_rate": 4.6542353769107226e-05,
      "loss": 0.488,
      "num_input_tokens_seen": 6302680,
      "step": 9635
    },
    {
      "epoch": 5.052410901467505,
      "grad_norm": 0.517509937286377,
      "learning_rate": 4.653654943959481e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 6306744,
      "step": 9640
    },
    {
      "epoch": 5.05503144654088,
      "grad_norm": 0.5703705549240112,
      "learning_rate": 4.653074060488195e-05,
      "loss": 0.6156,
      "num_input_tokens_seen": 6309400,
      "step": 9645
    },
    {
      "epoch": 5.0576519916142555,
      "grad_norm": 0.5296927094459534,
      "learning_rate": 4.6524927266183796e-05,
      "loss": 0.4137,
      "num_input_tokens_seen": 6312728,
      "step": 9650
    },
    {
      "epoch": 5.060272536687631,
      "grad_norm": 0.41497254371643066,
      "learning_rate": 4.6519109424716426e-05,
      "loss": 0.5674,
      "num_input_tokens_seen": 6315960,
      "step": 9655
    },
    {
      "epoch": 5.062893081761007,
      "grad_norm": 0.2105894684791565,
      "learning_rate": 4.6513287081696855e-05,
      "loss": 0.4016,
      "num_input_tokens_seen": 6319320,
      "step": 9660
    },
    {
      "epoch": 5.065513626834382,
      "grad_norm": 0.6831375360488892,
      "learning_rate": 4.6507460238343076e-05,
      "loss": 0.5716,
      "num_input_tokens_seen": 6321656,
      "step": 9665
    },
    {
      "epoch": 5.068134171907757,
      "grad_norm": 0.5176510214805603,
      "learning_rate": 4.6501628895873984e-05,
      "loss": 0.4094,
      "num_input_tokens_seen": 6324664,
      "step": 9670
    },
    {
      "epoch": 5.070754716981132,
      "grad_norm": 0.39402613043785095,
      "learning_rate": 4.649579305550942e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 6327192,
      "step": 9675
    },
    {
      "epoch": 5.073375262054507,
      "grad_norm": 0.5253567695617676,
      "learning_rate": 4.64899527184702e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 6330232,
      "step": 9680
    },
    {
      "epoch": 5.075995807127883,
      "grad_norm": 0.4313686192035675,
      "learning_rate": 4.6484107885978035e-05,
      "loss": 0.5147,
      "num_input_tokens_seen": 6332600,
      "step": 9685
    },
    {
      "epoch": 5.078616352201258,
      "grad_norm": 0.3572958707809448,
      "learning_rate": 4.64782585592556e-05,
      "loss": 0.5266,
      "num_input_tokens_seen": 6336600,
      "step": 9690
    },
    {
      "epoch": 5.081236897274633,
      "grad_norm": 0.5687592625617981,
      "learning_rate": 4.647240473952652e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 6339640,
      "step": 9695
    },
    {
      "epoch": 5.083857442348008,
      "grad_norm": 0.42406371235847473,
      "learning_rate": 4.6466546428015336e-05,
      "loss": 0.6715,
      "num_input_tokens_seen": 6342840,
      "step": 9700
    },
    {
      "epoch": 5.086477987421383,
      "grad_norm": 0.49759337306022644,
      "learning_rate": 4.646068362594753e-05,
      "loss": 0.5345,
      "num_input_tokens_seen": 6346008,
      "step": 9705
    },
    {
      "epoch": 5.089098532494759,
      "grad_norm": 0.5682648420333862,
      "learning_rate": 4.645481633454956e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 6349016,
      "step": 9710
    },
    {
      "epoch": 5.0917190775681345,
      "grad_norm": 0.5041850805282593,
      "learning_rate": 4.644894455504878e-05,
      "loss": 0.4314,
      "num_input_tokens_seen": 6352952,
      "step": 9715
    },
    {
      "epoch": 5.09433962264151,
      "grad_norm": 0.5312981605529785,
      "learning_rate": 4.64430682886735e-05,
      "loss": 0.4474,
      "num_input_tokens_seen": 6356504,
      "step": 9720
    },
    {
      "epoch": 5.096960167714885,
      "grad_norm": 0.30728742480278015,
      "learning_rate": 4.643718753665297e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 6359256,
      "step": 9725
    },
    {
      "epoch": 5.09958071278826,
      "grad_norm": 0.35681483149528503,
      "learning_rate": 4.643130230021737e-05,
      "loss": 0.5025,
      "num_input_tokens_seen": 6362520,
      "step": 9730
    },
    {
      "epoch": 5.102201257861635,
      "grad_norm": 0.3614998161792755,
      "learning_rate": 4.6425412580597836e-05,
      "loss": 0.5487,
      "num_input_tokens_seen": 6365368,
      "step": 9735
    },
    {
      "epoch": 5.10482180293501,
      "grad_norm": 0.4257192015647888,
      "learning_rate": 4.641951837902643e-05,
      "loss": 0.6978,
      "num_input_tokens_seen": 6367896,
      "step": 9740
    },
    {
      "epoch": 5.1074423480083855,
      "grad_norm": 0.6856440901756287,
      "learning_rate": 4.641361969673614e-05,
      "loss": 0.4764,
      "num_input_tokens_seen": 6371384,
      "step": 9745
    },
    {
      "epoch": 5.110062893081761,
      "grad_norm": 0.41200318932533264,
      "learning_rate": 4.640771653496091e-05,
      "loss": 0.4188,
      "num_input_tokens_seen": 6375128,
      "step": 9750
    },
    {
      "epoch": 5.112683438155136,
      "grad_norm": 0.3681190013885498,
      "learning_rate": 4.64018088949356e-05,
      "loss": 0.3454,
      "num_input_tokens_seen": 6377304,
      "step": 9755
    },
    {
      "epoch": 5.115303983228512,
      "grad_norm": 0.6216462254524231,
      "learning_rate": 4.6395896777896055e-05,
      "loss": 0.475,
      "num_input_tokens_seen": 6379992,
      "step": 9760
    },
    {
      "epoch": 5.117924528301887,
      "grad_norm": 0.6644501686096191,
      "learning_rate": 4.638998018507899e-05,
      "loss": 0.4758,
      "num_input_tokens_seen": 6383000,
      "step": 9765
    },
    {
      "epoch": 5.120545073375262,
      "grad_norm": 0.3631719648838043,
      "learning_rate": 4.63840591177221e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 6386776,
      "step": 9770
    },
    {
      "epoch": 5.123165618448637,
      "grad_norm": 0.5307257771492004,
      "learning_rate": 4.637813357706401e-05,
      "loss": 0.4629,
      "num_input_tokens_seen": 6390424,
      "step": 9775
    },
    {
      "epoch": 5.1257861635220126,
      "grad_norm": 0.9471929669380188,
      "learning_rate": 4.6372203564344264e-05,
      "loss": 0.6459,
      "num_input_tokens_seen": 6392888,
      "step": 9780
    },
    {
      "epoch": 5.128406708595388,
      "grad_norm": 0.3250109851360321,
      "learning_rate": 4.6366269080803364e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 6396728,
      "step": 9785
    },
    {
      "epoch": 5.131027253668763,
      "grad_norm": 0.3708447217941284,
      "learning_rate": 4.636033012768273e-05,
      "loss": 0.4717,
      "num_input_tokens_seen": 6400088,
      "step": 9790
    },
    {
      "epoch": 5.133647798742138,
      "grad_norm": 0.5054700374603271,
      "learning_rate": 4.635438670622473e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 6403032,
      "step": 9795
    },
    {
      "epoch": 5.136268343815513,
      "grad_norm": 0.42813315987586975,
      "learning_rate": 4.6348438817672646e-05,
      "loss": 0.4644,
      "num_input_tokens_seen": 6405944,
      "step": 9800
    },
    {
      "epoch": 5.138888888888889,
      "grad_norm": 0.4869573712348938,
      "learning_rate": 4.634248646327072e-05,
      "loss": 0.4546,
      "num_input_tokens_seen": 6409048,
      "step": 9805
    },
    {
      "epoch": 5.1415094339622645,
      "grad_norm": 0.3998188376426697,
      "learning_rate": 4.6336529644264114e-05,
      "loss": 0.4029,
      "num_input_tokens_seen": 6412920,
      "step": 9810
    },
    {
      "epoch": 5.14412997903564,
      "grad_norm": 0.47993728518486023,
      "learning_rate": 4.633056836189893e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 6416120,
      "step": 9815
    },
    {
      "epoch": 5.146750524109015,
      "grad_norm": 0.4094477891921997,
      "learning_rate": 4.632460261742219e-05,
      "loss": 0.6207,
      "num_input_tokens_seen": 6419896,
      "step": 9820
    },
    {
      "epoch": 5.14937106918239,
      "grad_norm": 0.4498511254787445,
      "learning_rate": 4.631863241208187e-05,
      "loss": 0.43,
      "num_input_tokens_seen": 6423096,
      "step": 9825
    },
    {
      "epoch": 5.151991614255765,
      "grad_norm": 0.5396851897239685,
      "learning_rate": 4.6312657747126865e-05,
      "loss": 0.5048,
      "num_input_tokens_seen": 6425592,
      "step": 9830
    },
    {
      "epoch": 5.15461215932914,
      "grad_norm": 0.7003793716430664,
      "learning_rate": 4.630667862380701e-05,
      "loss": 0.432,
      "num_input_tokens_seen": 6429304,
      "step": 9835
    },
    {
      "epoch": 5.1572327044025155,
      "grad_norm": 0.7127403020858765,
      "learning_rate": 4.630069504337306e-05,
      "loss": 0.4109,
      "num_input_tokens_seen": 6432984,
      "step": 9840
    },
    {
      "epoch": 5.159853249475891,
      "grad_norm": 0.33791422843933105,
      "learning_rate": 4.629470700707672e-05,
      "loss": 0.4171,
      "num_input_tokens_seen": 6436376,
      "step": 9845
    },
    {
      "epoch": 5.162473794549266,
      "grad_norm": 0.7657738924026489,
      "learning_rate": 4.628871451617062e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 6439544,
      "step": 9850
    },
    {
      "epoch": 5.165094339622642,
      "grad_norm": 0.5242050886154175,
      "learning_rate": 4.628271757190832e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 6443128,
      "step": 9855
    },
    {
      "epoch": 5.167714884696017,
      "grad_norm": 0.4166945815086365,
      "learning_rate": 4.62767161755443e-05,
      "loss": 0.528,
      "num_input_tokens_seen": 6446456,
      "step": 9860
    },
    {
      "epoch": 5.170335429769392,
      "grad_norm": 0.6119619011878967,
      "learning_rate": 4.6270710328334004e-05,
      "loss": 0.5159,
      "num_input_tokens_seen": 6449272,
      "step": 9865
    },
    {
      "epoch": 5.172955974842767,
      "grad_norm": 0.33638638257980347,
      "learning_rate": 4.626470003153377e-05,
      "loss": 0.5797,
      "num_input_tokens_seen": 6452632,
      "step": 9870
    },
    {
      "epoch": 5.1755765199161425,
      "grad_norm": 0.37137290835380554,
      "learning_rate": 4.625868528640089e-05,
      "loss": 0.3759,
      "num_input_tokens_seen": 6454936,
      "step": 9875
    },
    {
      "epoch": 5.178197064989518,
      "grad_norm": 0.5628800988197327,
      "learning_rate": 4.6252666094193586e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 6458168,
      "step": 9880
    },
    {
      "epoch": 5.180817610062893,
      "grad_norm": 0.6639958024024963,
      "learning_rate": 4.6246642456171e-05,
      "loss": 0.5497,
      "num_input_tokens_seen": 6462392,
      "step": 9885
    },
    {
      "epoch": 5.183438155136268,
      "grad_norm": 0.33639925718307495,
      "learning_rate": 4.6240614373593196e-05,
      "loss": 0.5386,
      "num_input_tokens_seen": 6470968,
      "step": 9890
    },
    {
      "epoch": 5.186058700209643,
      "grad_norm": 0.4841623306274414,
      "learning_rate": 4.623458184772119e-05,
      "loss": 0.4957,
      "num_input_tokens_seen": 6473880,
      "step": 9895
    },
    {
      "epoch": 5.188679245283019,
      "grad_norm": 0.4067640006542206,
      "learning_rate": 4.622854487981693e-05,
      "loss": 0.487,
      "num_input_tokens_seen": 6476504,
      "step": 9900
    },
    {
      "epoch": 5.191299790356394,
      "grad_norm": 0.9226706027984619,
      "learning_rate": 4.622250347114326e-05,
      "loss": 0.4577,
      "num_input_tokens_seen": 6479032,
      "step": 9905
    },
    {
      "epoch": 5.19392033542977,
      "grad_norm": 0.42313793301582336,
      "learning_rate": 4.621645762296399e-05,
      "loss": 0.4043,
      "num_input_tokens_seen": 6482392,
      "step": 9910
    },
    {
      "epoch": 5.196540880503145,
      "grad_norm": 0.4923549294471741,
      "learning_rate": 4.6210407336543835e-05,
      "loss": 0.5311,
      "num_input_tokens_seen": 6485816,
      "step": 9915
    },
    {
      "epoch": 5.19916142557652,
      "grad_norm": 0.3850288987159729,
      "learning_rate": 4.620435261314844e-05,
      "loss": 0.4163,
      "num_input_tokens_seen": 6489112,
      "step": 9920
    },
    {
      "epoch": 5.201781970649895,
      "grad_norm": 0.6767807602882385,
      "learning_rate": 4.61982934540444e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 6492248,
      "step": 9925
    },
    {
      "epoch": 5.20440251572327,
      "grad_norm": 0.6634882092475891,
      "learning_rate": 4.6192229860499206e-05,
      "loss": 0.5449,
      "num_input_tokens_seen": 6496600,
      "step": 9930
    },
    {
      "epoch": 5.2070230607966455,
      "grad_norm": 0.44370371103286743,
      "learning_rate": 4.6186161833781295e-05,
      "loss": 0.3439,
      "num_input_tokens_seen": 6499256,
      "step": 9935
    },
    {
      "epoch": 5.209643605870021,
      "grad_norm": 0.5649805665016174,
      "learning_rate": 4.618008937516003e-05,
      "loss": 0.5016,
      "num_input_tokens_seen": 6502360,
      "step": 9940
    },
    {
      "epoch": 5.212264150943396,
      "grad_norm": 0.5328425765037537,
      "learning_rate": 4.617401248590572e-05,
      "loss": 0.3355,
      "num_input_tokens_seen": 6504792,
      "step": 9945
    },
    {
      "epoch": 5.214884696016772,
      "grad_norm": 0.4016362130641937,
      "learning_rate": 4.616793116728955e-05,
      "loss": 0.4845,
      "num_input_tokens_seen": 6507992,
      "step": 9950
    },
    {
      "epoch": 5.217505241090147,
      "grad_norm": 0.4230624735355377,
      "learning_rate": 4.6161845420583685e-05,
      "loss": 0.37,
      "num_input_tokens_seen": 6511128,
      "step": 9955
    },
    {
      "epoch": 5.220125786163522,
      "grad_norm": 0.3093361556529999,
      "learning_rate": 4.6155755247061174e-05,
      "loss": 0.5867,
      "num_input_tokens_seen": 6514872,
      "step": 9960
    },
    {
      "epoch": 5.222746331236897,
      "grad_norm": 0.7608223557472229,
      "learning_rate": 4.614966064799603e-05,
      "loss": 0.441,
      "num_input_tokens_seen": 6517464,
      "step": 9965
    },
    {
      "epoch": 5.2253668763102725,
      "grad_norm": 0.36637747287750244,
      "learning_rate": 4.6143561624663166e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 6520056,
      "step": 9970
    },
    {
      "epoch": 5.227987421383648,
      "grad_norm": 0.477742463350296,
      "learning_rate": 4.6137458178338425e-05,
      "loss": 0.435,
      "num_input_tokens_seen": 6522840,
      "step": 9975
    },
    {
      "epoch": 5.230607966457023,
      "grad_norm": 0.31924620270729065,
      "learning_rate": 4.613135031029859e-05,
      "loss": 0.4465,
      "num_input_tokens_seen": 6526424,
      "step": 9980
    },
    {
      "epoch": 5.233228511530398,
      "grad_norm": 0.28968265652656555,
      "learning_rate": 4.612523802182134e-05,
      "loss": 0.4644,
      "num_input_tokens_seen": 6530008,
      "step": 9985
    },
    {
      "epoch": 5.235849056603773,
      "grad_norm": 0.4710724651813507,
      "learning_rate": 4.611912131418531e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 6532760,
      "step": 9990
    },
    {
      "epoch": 5.238469601677149,
      "grad_norm": 0.35329923033714294,
      "learning_rate": 4.611300018867004e-05,
      "loss": 0.5039,
      "num_input_tokens_seen": 6536984,
      "step": 9995
    },
    {
      "epoch": 5.241090146750524,
      "grad_norm": 0.4449976682662964,
      "learning_rate": 4.6106874646556e-05,
      "loss": 0.4885,
      "num_input_tokens_seen": 6540696,
      "step": 10000
    },
    {
      "epoch": 5.2437106918239,
      "grad_norm": 0.43159472942352295,
      "learning_rate": 4.6100744689124584e-05,
      "loss": 0.4008,
      "num_input_tokens_seen": 6543960,
      "step": 10005
    },
    {
      "epoch": 5.246331236897275,
      "grad_norm": 0.9445719718933105,
      "learning_rate": 4.6094610317658116e-05,
      "loss": 0.4516,
      "num_input_tokens_seen": 6546712,
      "step": 10010
    },
    {
      "epoch": 5.24895178197065,
      "grad_norm": 0.3886512815952301,
      "learning_rate": 4.608847153343982e-05,
      "loss": 0.4388,
      "num_input_tokens_seen": 6549944,
      "step": 10015
    },
    {
      "epoch": 5.251572327044025,
      "grad_norm": 0.4144759476184845,
      "learning_rate": 4.6082328337753874e-05,
      "loss": 0.5159,
      "num_input_tokens_seen": 6552792,
      "step": 10020
    },
    {
      "epoch": 5.2541928721174,
      "grad_norm": 0.5329417586326599,
      "learning_rate": 4.6076180731885353e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 6555288,
      "step": 10025
    },
    {
      "epoch": 5.256813417190775,
      "grad_norm": 0.3171195089817047,
      "learning_rate": 4.6070028717120286e-05,
      "loss": 0.3605,
      "num_input_tokens_seen": 6558232,
      "step": 10030
    },
    {
      "epoch": 5.259433962264151,
      "grad_norm": 0.28992730379104614,
      "learning_rate": 4.606387229474558e-05,
      "loss": 0.5582,
      "num_input_tokens_seen": 6562040,
      "step": 10035
    },
    {
      "epoch": 5.262054507337526,
      "grad_norm": 1.001344919204712,
      "learning_rate": 4.60577114660491e-05,
      "loss": 0.5809,
      "num_input_tokens_seen": 6564664,
      "step": 10040
    },
    {
      "epoch": 5.264675052410902,
      "grad_norm": 0.37931951880455017,
      "learning_rate": 4.605154623231962e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 6567608,
      "step": 10045
    },
    {
      "epoch": 5.267295597484277,
      "grad_norm": 0.5126233100891113,
      "learning_rate": 4.604537659484684e-05,
      "loss": 0.5529,
      "num_input_tokens_seen": 6570424,
      "step": 10050
    },
    {
      "epoch": 5.269916142557652,
      "grad_norm": 1.0791085958480835,
      "learning_rate": 4.603920255492138e-05,
      "loss": 0.5789,
      "num_input_tokens_seen": 6573624,
      "step": 10055
    },
    {
      "epoch": 5.272536687631027,
      "grad_norm": 0.3793744742870331,
      "learning_rate": 4.603302411383477e-05,
      "loss": 0.4524,
      "num_input_tokens_seen": 6576920,
      "step": 10060
    },
    {
      "epoch": 5.2751572327044025,
      "grad_norm": 0.35427260398864746,
      "learning_rate": 4.602684127287946e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 6580440,
      "step": 10065
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 0.38369038701057434,
      "learning_rate": 4.602065403334885e-05,
      "loss": 0.3743,
      "num_input_tokens_seen": 6583064,
      "step": 10070
    },
    {
      "epoch": 5.280398322851153,
      "grad_norm": 0.6008455753326416,
      "learning_rate": 4.601446239653724e-05,
      "loss": 0.4743,
      "num_input_tokens_seen": 6586840,
      "step": 10075
    },
    {
      "epoch": 5.283018867924528,
      "grad_norm": 0.5066109895706177,
      "learning_rate": 4.600826636373984e-05,
      "loss": 0.5638,
      "num_input_tokens_seen": 6590296,
      "step": 10080
    },
    {
      "epoch": 5.285639412997903,
      "grad_norm": 0.42096832394599915,
      "learning_rate": 4.600206593625279e-05,
      "loss": 0.4935,
      "num_input_tokens_seen": 6593368,
      "step": 10085
    },
    {
      "epoch": 5.288259958071279,
      "grad_norm": 0.3147873282432556,
      "learning_rate": 4.599586111537315e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 6596856,
      "step": 10090
    },
    {
      "epoch": 5.290880503144654,
      "grad_norm": 0.6502630114555359,
      "learning_rate": 4.59896519023989e-05,
      "loss": 0.4483,
      "num_input_tokens_seen": 6600184,
      "step": 10095
    },
    {
      "epoch": 5.29350104821803,
      "grad_norm": 0.4284690320491791,
      "learning_rate": 4.5983438298628935e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 6603608,
      "step": 10100
    },
    {
      "epoch": 5.296121593291405,
      "grad_norm": 0.5692240595817566,
      "learning_rate": 4.597722030536306e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 6606712,
      "step": 10105
    },
    {
      "epoch": 5.29874213836478,
      "grad_norm": 0.34976711869239807,
      "learning_rate": 4.597099792390202e-05,
      "loss": 0.3958,
      "num_input_tokens_seen": 6609624,
      "step": 10110
    },
    {
      "epoch": 5.301362683438155,
      "grad_norm": 0.3019692897796631,
      "learning_rate": 4.596477115554748e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 6613336,
      "step": 10115
    },
    {
      "epoch": 5.30398322851153,
      "grad_norm": 0.5959246754646301,
      "learning_rate": 4.5958540001601976e-05,
      "loss": 0.527,
      "num_input_tokens_seen": 6617560,
      "step": 10120
    },
    {
      "epoch": 5.306603773584905,
      "grad_norm": 0.40980568528175354,
      "learning_rate": 4.595230446336902e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 6621496,
      "step": 10125
    },
    {
      "epoch": 5.309224318658281,
      "grad_norm": 0.7036740183830261,
      "learning_rate": 4.5946064542153005e-05,
      "loss": 0.3909,
      "num_input_tokens_seen": 6625656,
      "step": 10130
    },
    {
      "epoch": 5.311844863731656,
      "grad_norm": 0.3640423119068146,
      "learning_rate": 4.593982023925926e-05,
      "loss": 0.3287,
      "num_input_tokens_seen": 6629240,
      "step": 10135
    },
    {
      "epoch": 5.314465408805032,
      "grad_norm": 0.3337063193321228,
      "learning_rate": 4.5933571555994e-05,
      "loss": 0.5679,
      "num_input_tokens_seen": 6633784,
      "step": 10140
    },
    {
      "epoch": 5.317085953878407,
      "grad_norm": 0.36896511912345886,
      "learning_rate": 4.59273184936644e-05,
      "loss": 0.5436,
      "num_input_tokens_seen": 6636600,
      "step": 10145
    },
    {
      "epoch": 5.319706498951782,
      "grad_norm": 0.7943920493125916,
      "learning_rate": 4.5921061053578523e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 6640088,
      "step": 10150
    },
    {
      "epoch": 5.322327044025157,
      "grad_norm": 0.4157312512397766,
      "learning_rate": 4.5914799237045355e-05,
      "loss": 0.47,
      "num_input_tokens_seen": 6642840,
      "step": 10155
    },
    {
      "epoch": 5.3249475890985325,
      "grad_norm": 0.2731707692146301,
      "learning_rate": 4.5908533045374796e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 6646136,
      "step": 10160
    },
    {
      "epoch": 5.327568134171908,
      "grad_norm": 0.876069188117981,
      "learning_rate": 4.590226247987766e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 6649560,
      "step": 10165
    },
    {
      "epoch": 5.330188679245283,
      "grad_norm": 0.313625305891037,
      "learning_rate": 4.589598754186568e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 6652888,
      "step": 10170
    },
    {
      "epoch": 5.332809224318658,
      "grad_norm": 0.6145751476287842,
      "learning_rate": 4.5889708232651494e-05,
      "loss": 0.3914,
      "num_input_tokens_seen": 6655576,
      "step": 10175
    },
    {
      "epoch": 5.335429769392033,
      "grad_norm": 0.6626905202865601,
      "learning_rate": 4.588342455354867e-05,
      "loss": 0.4642,
      "num_input_tokens_seen": 6659288,
      "step": 10180
    },
    {
      "epoch": 5.338050314465409,
      "grad_norm": 0.44043591618537903,
      "learning_rate": 4.5877136505871685e-05,
      "loss": 0.4094,
      "num_input_tokens_seen": 6662360,
      "step": 10185
    },
    {
      "epoch": 5.340670859538784,
      "grad_norm": 0.21900811791419983,
      "learning_rate": 4.5870844090935914e-05,
      "loss": 0.4829,
      "num_input_tokens_seen": 6666584,
      "step": 10190
    },
    {
      "epoch": 5.34329140461216,
      "grad_norm": 0.732434868812561,
      "learning_rate": 4.586454731005768e-05,
      "loss": 0.6577,
      "num_input_tokens_seen": 6669176,
      "step": 10195
    },
    {
      "epoch": 5.345911949685535,
      "grad_norm": 0.31307148933410645,
      "learning_rate": 4.5858246164554175e-05,
      "loss": 0.5826,
      "num_input_tokens_seen": 6672728,
      "step": 10200
    },
    {
      "epoch": 5.34853249475891,
      "grad_norm": 0.3737022876739502,
      "learning_rate": 4.5851940655743536e-05,
      "loss": 0.3512,
      "num_input_tokens_seen": 6676024,
      "step": 10205
    },
    {
      "epoch": 5.351153039832285,
      "grad_norm": 0.41535264253616333,
      "learning_rate": 4.5845630784944804e-05,
      "loss": 0.5712,
      "num_input_tokens_seen": 6679224,
      "step": 10210
    },
    {
      "epoch": 5.35377358490566,
      "grad_norm": 0.9330704808235168,
      "learning_rate": 4.583931655347793e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 6681528,
      "step": 10215
    },
    {
      "epoch": 5.356394129979035,
      "grad_norm": 0.4122796952724457,
      "learning_rate": 4.5832997962663785e-05,
      "loss": 0.3973,
      "num_input_tokens_seen": 6683992,
      "step": 10220
    },
    {
      "epoch": 5.359014675052411,
      "grad_norm": 0.48299625515937805,
      "learning_rate": 4.582667501382414e-05,
      "loss": 0.3951,
      "num_input_tokens_seen": 6687576,
      "step": 10225
    },
    {
      "epoch": 5.361635220125786,
      "grad_norm": 0.3649064004421234,
      "learning_rate": 4.582034770828169e-05,
      "loss": 0.5363,
      "num_input_tokens_seen": 6690264,
      "step": 10230
    },
    {
      "epoch": 5.364255765199162,
      "grad_norm": 0.6997343897819519,
      "learning_rate": 4.581401604736002e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 6693400,
      "step": 10235
    },
    {
      "epoch": 5.366876310272537,
      "grad_norm": 0.43956679105758667,
      "learning_rate": 4.580768003238365e-05,
      "loss": 0.526,
      "num_input_tokens_seen": 6695960,
      "step": 10240
    },
    {
      "epoch": 5.369496855345912,
      "grad_norm": 0.6008044481277466,
      "learning_rate": 4.5801339664678014e-05,
      "loss": 0.4525,
      "num_input_tokens_seen": 6699064,
      "step": 10245
    },
    {
      "epoch": 5.372117400419287,
      "grad_norm": 0.2799290120601654,
      "learning_rate": 4.579499494556943e-05,
      "loss": 0.4034,
      "num_input_tokens_seen": 6702008,
      "step": 10250
    },
    {
      "epoch": 5.3747379454926625,
      "grad_norm": 0.5732994675636292,
      "learning_rate": 4.5788645876385147e-05,
      "loss": 0.6809,
      "num_input_tokens_seen": 6705560,
      "step": 10255
    },
    {
      "epoch": 5.377358490566038,
      "grad_norm": 0.6893916726112366,
      "learning_rate": 4.578229245845331e-05,
      "loss": 0.583,
      "num_input_tokens_seen": 6709016,
      "step": 10260
    },
    {
      "epoch": 5.379979035639413,
      "grad_norm": 0.477921724319458,
      "learning_rate": 4.5775934693102994e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 6711736,
      "step": 10265
    },
    {
      "epoch": 5.382599580712788,
      "grad_norm": 0.3498796224594116,
      "learning_rate": 4.5769572581664165e-05,
      "loss": 0.4576,
      "num_input_tokens_seen": 6714104,
      "step": 10270
    },
    {
      "epoch": 5.385220125786163,
      "grad_norm": 0.5394643545150757,
      "learning_rate": 4.57632061254677e-05,
      "loss": 0.4571,
      "num_input_tokens_seen": 6717720,
      "step": 10275
    },
    {
      "epoch": 5.387840670859539,
      "grad_norm": 0.7098197340965271,
      "learning_rate": 4.5756835325845394e-05,
      "loss": 0.5521,
      "num_input_tokens_seen": 6720728,
      "step": 10280
    },
    {
      "epoch": 5.390461215932914,
      "grad_norm": 0.48805853724479675,
      "learning_rate": 4.575046018412995e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 6724344,
      "step": 10285
    },
    {
      "epoch": 5.3930817610062896,
      "grad_norm": 0.39631280303001404,
      "learning_rate": 4.5744080701654955e-05,
      "loss": 0.4726,
      "num_input_tokens_seen": 6728280,
      "step": 10290
    },
    {
      "epoch": 5.395702306079665,
      "grad_norm": 0.3819165825843811,
      "learning_rate": 4.573769687975495e-05,
      "loss": 0.4746,
      "num_input_tokens_seen": 6730936,
      "step": 10295
    },
    {
      "epoch": 5.39832285115304,
      "grad_norm": 0.36442211270332336,
      "learning_rate": 4.573130871976534e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 6733624,
      "step": 10300
    },
    {
      "epoch": 5.400943396226415,
      "grad_norm": 0.434251070022583,
      "learning_rate": 4.572491622302247e-05,
      "loss": 0.4896,
      "num_input_tokens_seen": 6736184,
      "step": 10305
    },
    {
      "epoch": 5.40356394129979,
      "grad_norm": 0.3364071249961853,
      "learning_rate": 4.571851939086356e-05,
      "loss": 0.4854,
      "num_input_tokens_seen": 6739192,
      "step": 10310
    },
    {
      "epoch": 5.406184486373165,
      "grad_norm": 0.4356286823749542,
      "learning_rate": 4.571211822462676e-05,
      "loss": 0.3569,
      "num_input_tokens_seen": 6742360,
      "step": 10315
    },
    {
      "epoch": 5.408805031446541,
      "grad_norm": 0.31352537870407104,
      "learning_rate": 4.570571272565114e-05,
      "loss": 0.4675,
      "num_input_tokens_seen": 6746200,
      "step": 10320
    },
    {
      "epoch": 5.411425576519916,
      "grad_norm": 0.26367005705833435,
      "learning_rate": 4.569930289527663e-05,
      "loss": 0.4881,
      "num_input_tokens_seen": 6750616,
      "step": 10325
    },
    {
      "epoch": 5.414046121593292,
      "grad_norm": 0.5351378917694092,
      "learning_rate": 4.56928887348441e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 6753688,
      "step": 10330
    },
    {
      "epoch": 5.416666666666667,
      "grad_norm": 0.19455653429031372,
      "learning_rate": 4.568647024569533e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 6756920,
      "step": 10335
    },
    {
      "epoch": 5.419287211740042,
      "grad_norm": 0.44748547673225403,
      "learning_rate": 4.568004742917299e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 6761048,
      "step": 10340
    },
    {
      "epoch": 5.421907756813417,
      "grad_norm": 0.3030480444431305,
      "learning_rate": 4.567362028662065e-05,
      "loss": 0.3631,
      "num_input_tokens_seen": 6764280,
      "step": 10345
    },
    {
      "epoch": 5.4245283018867925,
      "grad_norm": 1.1430102586746216,
      "learning_rate": 4.566718881938281e-05,
      "loss": 0.6166,
      "num_input_tokens_seen": 6767704,
      "step": 10350
    },
    {
      "epoch": 5.427148846960168,
      "grad_norm": 0.3414372205734253,
      "learning_rate": 4.5660753028804846e-05,
      "loss": 0.4349,
      "num_input_tokens_seen": 6772280,
      "step": 10355
    },
    {
      "epoch": 5.429769392033543,
      "grad_norm": 0.46446749567985535,
      "learning_rate": 4.5654312916233066e-05,
      "loss": 0.3928,
      "num_input_tokens_seen": 6775704,
      "step": 10360
    },
    {
      "epoch": 5.432389937106918,
      "grad_norm": 0.3485252261161804,
      "learning_rate": 4.564786848301466e-05,
      "loss": 0.458,
      "num_input_tokens_seen": 6779288,
      "step": 10365
    },
    {
      "epoch": 5.435010482180293,
      "grad_norm": 0.8290926218032837,
      "learning_rate": 4.564141973049773e-05,
      "loss": 0.4581,
      "num_input_tokens_seen": 6781848,
      "step": 10370
    },
    {
      "epoch": 5.437631027253669,
      "grad_norm": 0.5332638621330261,
      "learning_rate": 4.563496666003128e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 6784920,
      "step": 10375
    },
    {
      "epoch": 5.440251572327044,
      "grad_norm": 0.4411699175834656,
      "learning_rate": 4.562850927296523e-05,
      "loss": 0.5608,
      "num_input_tokens_seen": 6788344,
      "step": 10380
    },
    {
      "epoch": 5.4428721174004195,
      "grad_norm": 0.5792927145957947,
      "learning_rate": 4.5622047570650365e-05,
      "loss": 0.649,
      "num_input_tokens_seen": 6791000,
      "step": 10385
    },
    {
      "epoch": 5.445492662473795,
      "grad_norm": 0.4802071750164032,
      "learning_rate": 4.561558155443843e-05,
      "loss": 0.4632,
      "num_input_tokens_seen": 6793560,
      "step": 10390
    },
    {
      "epoch": 5.44811320754717,
      "grad_norm": 0.4274453818798065,
      "learning_rate": 4.560911122568202e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 6795896,
      "step": 10395
    },
    {
      "epoch": 5.450733752620545,
      "grad_norm": 0.6963222622871399,
      "learning_rate": 4.560263658573467e-05,
      "loss": 0.3264,
      "num_input_tokens_seen": 6798584,
      "step": 10400
    },
    {
      "epoch": 5.45335429769392,
      "grad_norm": 0.5303966999053955,
      "learning_rate": 4.559615763595079e-05,
      "loss": 0.4056,
      "num_input_tokens_seen": 6801816,
      "step": 10405
    },
    {
      "epoch": 5.455974842767295,
      "grad_norm": 0.5614333748817444,
      "learning_rate": 4.558967437768571e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 6805880,
      "step": 10410
    },
    {
      "epoch": 5.4585953878406706,
      "grad_norm": 0.38547855615615845,
      "learning_rate": 4.558318681229565e-05,
      "loss": 0.7166,
      "num_input_tokens_seen": 6809240,
      "step": 10415
    },
    {
      "epoch": 5.461215932914046,
      "grad_norm": 0.35511860251426697,
      "learning_rate": 4.557669494113773e-05,
      "loss": 0.3829,
      "num_input_tokens_seen": 6812344,
      "step": 10420
    },
    {
      "epoch": 5.463836477987422,
      "grad_norm": 0.2546852231025696,
      "learning_rate": 4.557019876556998e-05,
      "loss": 0.5282,
      "num_input_tokens_seen": 6815800,
      "step": 10425
    },
    {
      "epoch": 5.466457023060797,
      "grad_norm": 0.3187004327774048,
      "learning_rate": 4.556369828695133e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 6823736,
      "step": 10430
    },
    {
      "epoch": 5.469077568134172,
      "grad_norm": 0.6214509010314941,
      "learning_rate": 4.5557193506641596e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 6826520,
      "step": 10435
    },
    {
      "epoch": 5.471698113207547,
      "grad_norm": 0.6919266581535339,
      "learning_rate": 4.555068442600151e-05,
      "loss": 0.4299,
      "num_input_tokens_seen": 6829880,
      "step": 10440
    },
    {
      "epoch": 5.4743186582809225,
      "grad_norm": 0.553815484046936,
      "learning_rate": 4.5544171046392703e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 6833752,
      "step": 10445
    },
    {
      "epoch": 5.476939203354298,
      "grad_norm": 0.5863984823226929,
      "learning_rate": 4.5537653369177685e-05,
      "loss": 0.5988,
      "num_input_tokens_seen": 6837208,
      "step": 10450
    },
    {
      "epoch": 5.479559748427673,
      "grad_norm": 0.2790643870830536,
      "learning_rate": 4.5531131395719895e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 6843544,
      "step": 10455
    },
    {
      "epoch": 5.482180293501048,
      "grad_norm": 0.4448651075363159,
      "learning_rate": 4.552460512738364e-05,
      "loss": 0.6402,
      "num_input_tokens_seen": 6846488,
      "step": 10460
    },
    {
      "epoch": 5.484800838574423,
      "grad_norm": 0.49087947607040405,
      "learning_rate": 4.551807456553416e-05,
      "loss": 0.4259,
      "num_input_tokens_seen": 6850136,
      "step": 10465
    },
    {
      "epoch": 5.487421383647799,
      "grad_norm": 0.41786420345306396,
      "learning_rate": 4.551153971153755e-05,
      "loss": 0.4904,
      "num_input_tokens_seen": 6853016,
      "step": 10470
    },
    {
      "epoch": 5.490041928721174,
      "grad_norm": 0.6623417735099792,
      "learning_rate": 4.550500056676085e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 6856088,
      "step": 10475
    },
    {
      "epoch": 5.4926624737945495,
      "grad_norm": 0.38497650623321533,
      "learning_rate": 4.5498457132571966e-05,
      "loss": 0.5535,
      "num_input_tokens_seen": 6859288,
      "step": 10480
    },
    {
      "epoch": 5.495283018867925,
      "grad_norm": 0.3170578181743622,
      "learning_rate": 4.5491909410339706e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 6863480,
      "step": 10485
    },
    {
      "epoch": 5.4979035639413,
      "grad_norm": 0.46805357933044434,
      "learning_rate": 4.548535740143378e-05,
      "loss": 0.2436,
      "num_input_tokens_seen": 6869656,
      "step": 10490
    },
    {
      "epoch": 5.500524109014675,
      "grad_norm": 0.40665289759635925,
      "learning_rate": 4.54788011072248e-05,
      "loss": 0.3843,
      "num_input_tokens_seen": 6872344,
      "step": 10495
    },
    {
      "epoch": 5.50314465408805,
      "grad_norm": 0.3469765782356262,
      "learning_rate": 4.5472240529084254e-05,
      "loss": 0.428,
      "num_input_tokens_seen": 6876216,
      "step": 10500
    },
    {
      "epoch": 5.505765199161425,
      "grad_norm": 0.3752704858779907,
      "learning_rate": 4.5465675668384557e-05,
      "loss": 0.3647,
      "num_input_tokens_seen": 6879384,
      "step": 10505
    },
    {
      "epoch": 5.5083857442348005,
      "grad_norm": 1.1557300090789795,
      "learning_rate": 4.545910652649899e-05,
      "loss": 0.555,
      "num_input_tokens_seen": 6882232,
      "step": 10510
    },
    {
      "epoch": 5.511006289308176,
      "grad_norm": 0.47283339500427246,
      "learning_rate": 4.5452533104801756e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 6886072,
      "step": 10515
    },
    {
      "epoch": 5.513626834381551,
      "grad_norm": 0.38131365180015564,
      "learning_rate": 4.544595540466792e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 6889272,
      "step": 10520
    },
    {
      "epoch": 5.516247379454927,
      "grad_norm": 0.3016335666179657,
      "learning_rate": 4.5439373427473476e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 6892824,
      "step": 10525
    },
    {
      "epoch": 5.518867924528302,
      "grad_norm": 0.4809129536151886,
      "learning_rate": 4.54327871745953e-05,
      "loss": 0.4747,
      "num_input_tokens_seen": 6896248,
      "step": 10530
    },
    {
      "epoch": 5.521488469601677,
      "grad_norm": 0.3673061430454254,
      "learning_rate": 4.542619664741116e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 6899640,
      "step": 10535
    },
    {
      "epoch": 5.524109014675052,
      "grad_norm": 0.3342452347278595,
      "learning_rate": 4.5419601847299706e-05,
      "loss": 0.4319,
      "num_input_tokens_seen": 6903896,
      "step": 10540
    },
    {
      "epoch": 5.526729559748428,
      "grad_norm": 0.473952054977417,
      "learning_rate": 4.5413002775640503e-05,
      "loss": 0.5413,
      "num_input_tokens_seen": 6907448,
      "step": 10545
    },
    {
      "epoch": 5.529350104821803,
      "grad_norm": 0.3939207196235657,
      "learning_rate": 4.540639943381401e-05,
      "loss": 0.4823,
      "num_input_tokens_seen": 6910840,
      "step": 10550
    },
    {
      "epoch": 5.531970649895178,
      "grad_norm": 0.46914350986480713,
      "learning_rate": 4.539979182320157e-05,
      "loss": 0.4027,
      "num_input_tokens_seen": 6913528,
      "step": 10555
    },
    {
      "epoch": 5.534591194968553,
      "grad_norm": 0.5080075860023499,
      "learning_rate": 4.5393179945185404e-05,
      "loss": 0.4983,
      "num_input_tokens_seen": 6916312,
      "step": 10560
    },
    {
      "epoch": 5.537211740041929,
      "grad_norm": 0.4398108422756195,
      "learning_rate": 4.538656380114865e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 6919384,
      "step": 10565
    },
    {
      "epoch": 5.539832285115304,
      "grad_norm": 0.6310964226722717,
      "learning_rate": 4.537994339247533e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 6921816,
      "step": 10570
    },
    {
      "epoch": 5.5424528301886795,
      "grad_norm": 0.4419868290424347,
      "learning_rate": 4.5373318720550366e-05,
      "loss": 0.5538,
      "num_input_tokens_seen": 6924760,
      "step": 10575
    },
    {
      "epoch": 5.545073375262055,
      "grad_norm": 0.2332027405500412,
      "learning_rate": 4.536668978675956e-05,
      "loss": 0.4154,
      "num_input_tokens_seen": 6928600,
      "step": 10580
    },
    {
      "epoch": 5.54769392033543,
      "grad_norm": 0.6870108842849731,
      "learning_rate": 4.53600565924896e-05,
      "loss": 0.3637,
      "num_input_tokens_seen": 6932120,
      "step": 10585
    },
    {
      "epoch": 5.550314465408805,
      "grad_norm": 0.6183087825775146,
      "learning_rate": 4.535341913912808e-05,
      "loss": 0.4623,
      "num_input_tokens_seen": 6934968,
      "step": 10590
    },
    {
      "epoch": 5.55293501048218,
      "grad_norm": 0.3786051273345947,
      "learning_rate": 4.534677742806348e-05,
      "loss": 0.5037,
      "num_input_tokens_seen": 6938776,
      "step": 10595
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 0.7432551980018616,
      "learning_rate": 4.5340131460685175e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 6942200,
      "step": 10600
    },
    {
      "epoch": 5.5581761006289305,
      "grad_norm": 0.5384470224380493,
      "learning_rate": 4.533348123838341e-05,
      "loss": 0.7133,
      "num_input_tokens_seen": 6944888,
      "step": 10605
    },
    {
      "epoch": 5.560796645702306,
      "grad_norm": 0.4542539715766907,
      "learning_rate": 4.532682676254936e-05,
      "loss": 0.4396,
      "num_input_tokens_seen": 6947896,
      "step": 10610
    },
    {
      "epoch": 5.563417190775681,
      "grad_norm": 0.8010192513465881,
      "learning_rate": 4.5320168034575044e-05,
      "loss": 0.6348,
      "num_input_tokens_seen": 6950712,
      "step": 10615
    },
    {
      "epoch": 5.566037735849057,
      "grad_norm": 0.6095278263092041,
      "learning_rate": 4.53135050558534e-05,
      "loss": 0.6164,
      "num_input_tokens_seen": 6953496,
      "step": 10620
    },
    {
      "epoch": 5.568658280922432,
      "grad_norm": 0.4198535978794098,
      "learning_rate": 4.530683782777825e-05,
      "loss": 0.4638,
      "num_input_tokens_seen": 6956152,
      "step": 10625
    },
    {
      "epoch": 5.571278825995807,
      "grad_norm": 0.5234254002571106,
      "learning_rate": 4.53001663517443e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 6959032,
      "step": 10630
    },
    {
      "epoch": 5.573899371069182,
      "grad_norm": 0.2930322587490082,
      "learning_rate": 4.529349062914714e-05,
      "loss": 0.3822,
      "num_input_tokens_seen": 6962296,
      "step": 10635
    },
    {
      "epoch": 5.576519916142558,
      "grad_norm": 0.3348231613636017,
      "learning_rate": 4.528681066138326e-05,
      "loss": 0.5175,
      "num_input_tokens_seen": 6965784,
      "step": 10640
    },
    {
      "epoch": 5.579140461215933,
      "grad_norm": 1.163069725036621,
      "learning_rate": 4.528012644985005e-05,
      "loss": 0.5214,
      "num_input_tokens_seen": 6969112,
      "step": 10645
    },
    {
      "epoch": 5.581761006289308,
      "grad_norm": 0.5316640734672546,
      "learning_rate": 4.527343799594574e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 6972216,
      "step": 10650
    },
    {
      "epoch": 5.584381551362683,
      "grad_norm": 0.3314219117164612,
      "learning_rate": 4.5266745301069494e-05,
      "loss": 0.5658,
      "num_input_tokens_seen": 6975064,
      "step": 10655
    },
    {
      "epoch": 5.587002096436059,
      "grad_norm": 0.4424344599246979,
      "learning_rate": 4.526004836662135e-05,
      "loss": 0.4931,
      "num_input_tokens_seen": 6977688,
      "step": 10660
    },
    {
      "epoch": 5.589622641509434,
      "grad_norm": 0.44969791173934937,
      "learning_rate": 4.525334719400223e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 6981208,
      "step": 10665
    },
    {
      "epoch": 5.5922431865828095,
      "grad_norm": 0.8917309045791626,
      "learning_rate": 4.524664178461393e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 6983960,
      "step": 10670
    },
    {
      "epoch": 5.594863731656185,
      "grad_norm": 0.4538506269454956,
      "learning_rate": 4.5239932139859154e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 6987320,
      "step": 10675
    },
    {
      "epoch": 5.59748427672956,
      "grad_norm": 0.395402193069458,
      "learning_rate": 4.523321826114149e-05,
      "loss": 0.4724,
      "num_input_tokens_seen": 6991064,
      "step": 10680
    },
    {
      "epoch": 5.600104821802935,
      "grad_norm": 0.569220781326294,
      "learning_rate": 4.5226500149865395e-05,
      "loss": 0.5084,
      "num_input_tokens_seen": 6993624,
      "step": 10685
    },
    {
      "epoch": 5.60272536687631,
      "grad_norm": 0.4147249460220337,
      "learning_rate": 4.5219777807436224e-05,
      "loss": 0.4935,
      "num_input_tokens_seen": 6996344,
      "step": 10690
    },
    {
      "epoch": 5.605345911949685,
      "grad_norm": 0.4989956319332123,
      "learning_rate": 4.5213051235260225e-05,
      "loss": 0.4601,
      "num_input_tokens_seen": 6999576,
      "step": 10695
    },
    {
      "epoch": 5.6079664570230605,
      "grad_norm": 0.559284508228302,
      "learning_rate": 4.52063204347445e-05,
      "loss": 0.4999,
      "num_input_tokens_seen": 7003096,
      "step": 10700
    },
    {
      "epoch": 5.610587002096436,
      "grad_norm": 0.3757461905479431,
      "learning_rate": 4.519958540729707e-05,
      "loss": 0.4287,
      "num_input_tokens_seen": 7006328,
      "step": 10705
    },
    {
      "epoch": 5.613207547169811,
      "grad_norm": 0.43090566992759705,
      "learning_rate": 4.519284615432682e-05,
      "loss": 0.5393,
      "num_input_tokens_seen": 7009560,
      "step": 10710
    },
    {
      "epoch": 5.615828092243187,
      "grad_norm": 0.34092214703559875,
      "learning_rate": 4.518610267724352e-05,
      "loss": 0.5437,
      "num_input_tokens_seen": 7012216,
      "step": 10715
    },
    {
      "epoch": 5.618448637316562,
      "grad_norm": 0.3777334690093994,
      "learning_rate": 4.517935497745786e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 7015224,
      "step": 10720
    },
    {
      "epoch": 5.621069182389937,
      "grad_norm": 0.549567699432373,
      "learning_rate": 4.517260305638133e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 7018296,
      "step": 10725
    },
    {
      "epoch": 5.623689727463312,
      "grad_norm": 0.24914997816085815,
      "learning_rate": 4.516584691542639e-05,
      "loss": 0.4213,
      "num_input_tokens_seen": 7022424,
      "step": 10730
    },
    {
      "epoch": 5.626310272536688,
      "grad_norm": 0.3840946853160858,
      "learning_rate": 4.515908655600634e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 7026456,
      "step": 10735
    },
    {
      "epoch": 5.628930817610063,
      "grad_norm": 0.40844205021858215,
      "learning_rate": 4.515232197953537e-05,
      "loss": 0.5934,
      "num_input_tokens_seen": 7029880,
      "step": 10740
    },
    {
      "epoch": 5.631551362683438,
      "grad_norm": 0.44084104895591736,
      "learning_rate": 4.514555318742856e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 7033080,
      "step": 10745
    },
    {
      "epoch": 5.634171907756813,
      "grad_norm": 0.30793097615242004,
      "learning_rate": 4.513878018110184e-05,
      "loss": 0.509,
      "num_input_tokens_seen": 7036760,
      "step": 10750
    },
    {
      "epoch": 5.636792452830189,
      "grad_norm": 0.4326469600200653,
      "learning_rate": 4.513200296197207e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 7040920,
      "step": 10755
    },
    {
      "epoch": 5.639412997903564,
      "grad_norm": 0.442617803812027,
      "learning_rate": 4.512522153145696e-05,
      "loss": 0.4416,
      "num_input_tokens_seen": 7043416,
      "step": 10760
    },
    {
      "epoch": 5.6420335429769395,
      "grad_norm": 0.711178183555603,
      "learning_rate": 4.5118435890975106e-05,
      "loss": 0.4864,
      "num_input_tokens_seen": 7046296,
      "step": 10765
    },
    {
      "epoch": 5.644654088050315,
      "grad_norm": 0.483953595161438,
      "learning_rate": 4.511164604194598e-05,
      "loss": 0.3426,
      "num_input_tokens_seen": 7050008,
      "step": 10770
    },
    {
      "epoch": 5.64727463312369,
      "grad_norm": 0.3300990164279938,
      "learning_rate": 4.5104851985789956e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 7053240,
      "step": 10775
    },
    {
      "epoch": 5.649895178197065,
      "grad_norm": 0.6017496585845947,
      "learning_rate": 4.5098053723928254e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 7056312,
      "step": 10780
    },
    {
      "epoch": 5.65251572327044,
      "grad_norm": 0.33355095982551575,
      "learning_rate": 4.509125125778302e-05,
      "loss": 0.365,
      "num_input_tokens_seen": 7059960,
      "step": 10785
    },
    {
      "epoch": 5.655136268343815,
      "grad_norm": 0.4760131537914276,
      "learning_rate": 4.508444458877722e-05,
      "loss": 0.5054,
      "num_input_tokens_seen": 7062936,
      "step": 10790
    },
    {
      "epoch": 5.6577568134171905,
      "grad_norm": 0.7523110508918762,
      "learning_rate": 4.5077633718334755e-05,
      "loss": 0.5232,
      "num_input_tokens_seen": 7065368,
      "step": 10795
    },
    {
      "epoch": 5.660377358490566,
      "grad_norm": 0.38196101784706116,
      "learning_rate": 4.5070818647880375e-05,
      "loss": 0.574,
      "num_input_tokens_seen": 7069464,
      "step": 10800
    },
    {
      "epoch": 5.662997903563941,
      "grad_norm": 0.5940772891044617,
      "learning_rate": 4.506399937883972e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 7071768,
      "step": 10805
    },
    {
      "epoch": 5.665618448637317,
      "grad_norm": 0.3864857852458954,
      "learning_rate": 4.5057175912639286e-05,
      "loss": 0.3741,
      "num_input_tokens_seen": 7075704,
      "step": 10810
    },
    {
      "epoch": 5.668238993710692,
      "grad_norm": 0.5716235637664795,
      "learning_rate": 4.5050348250706475e-05,
      "loss": 0.47,
      "num_input_tokens_seen": 7078264,
      "step": 10815
    },
    {
      "epoch": 5.670859538784067,
      "grad_norm": 0.8541088104248047,
      "learning_rate": 4.5043516394469566e-05,
      "loss": 0.5023,
      "num_input_tokens_seen": 7082424,
      "step": 10820
    },
    {
      "epoch": 5.673480083857442,
      "grad_norm": 0.5039156079292297,
      "learning_rate": 4.503668034535768e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 7085624,
      "step": 10825
    },
    {
      "epoch": 5.676100628930818,
      "grad_norm": 0.3501913249492645,
      "learning_rate": 4.5029840104800876e-05,
      "loss": 0.4239,
      "num_input_tokens_seen": 7089112,
      "step": 10830
    },
    {
      "epoch": 5.678721174004193,
      "grad_norm": 0.38759952783584595,
      "learning_rate": 4.5022995674230025e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 7092728,
      "step": 10835
    },
    {
      "epoch": 5.681341719077568,
      "grad_norm": 0.3214082419872284,
      "learning_rate": 4.5016147055076914e-05,
      "loss": 0.6056,
      "num_input_tokens_seen": 7095544,
      "step": 10840
    },
    {
      "epoch": 5.683962264150943,
      "grad_norm": 0.469497948884964,
      "learning_rate": 4.500929424877419e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 7098264,
      "step": 10845
    },
    {
      "epoch": 5.686582809224319,
      "grad_norm": 0.37452182173728943,
      "learning_rate": 4.5002437256755394e-05,
      "loss": 0.4596,
      "num_input_tokens_seen": 7101464,
      "step": 10850
    },
    {
      "epoch": 5.689203354297694,
      "grad_norm": 0.7188599705696106,
      "learning_rate": 4.4995576080454924e-05,
      "loss": 0.5512,
      "num_input_tokens_seen": 7105080,
      "step": 10855
    },
    {
      "epoch": 5.6918238993710695,
      "grad_norm": 0.6050329804420471,
      "learning_rate": 4.4988710721308056e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 7107800,
      "step": 10860
    },
    {
      "epoch": 5.694444444444445,
      "grad_norm": 0.53925621509552,
      "learning_rate": 4.498184118075095e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 7110712,
      "step": 10865
    },
    {
      "epoch": 5.69706498951782,
      "grad_norm": 0.4286741316318512,
      "learning_rate": 4.497496746022063e-05,
      "loss": 0.6768,
      "num_input_tokens_seen": 7114520,
      "step": 10870
    },
    {
      "epoch": 5.699685534591195,
      "grad_norm": 0.5880365371704102,
      "learning_rate": 4.4968089561155015e-05,
      "loss": 0.3437,
      "num_input_tokens_seen": 7118232,
      "step": 10875
    },
    {
      "epoch": 5.70230607966457,
      "grad_norm": 0.3489993214607239,
      "learning_rate": 4.496120748499286e-05,
      "loss": 0.5484,
      "num_input_tokens_seen": 7121080,
      "step": 10880
    },
    {
      "epoch": 5.704926624737945,
      "grad_norm": 1.3373587131500244,
      "learning_rate": 4.4954321233173833e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 7124792,
      "step": 10885
    },
    {
      "epoch": 5.7075471698113205,
      "grad_norm": 0.42769965529441833,
      "learning_rate": 4.494743080713846e-05,
      "loss": 0.366,
      "num_input_tokens_seen": 7127640,
      "step": 10890
    },
    {
      "epoch": 5.710167714884696,
      "grad_norm": 0.43355387449264526,
      "learning_rate": 4.494053620832812e-05,
      "loss": 0.4443,
      "num_input_tokens_seen": 7130392,
      "step": 10895
    },
    {
      "epoch": 5.712788259958071,
      "grad_norm": 0.3760250210762024,
      "learning_rate": 4.493363743818511e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 7133912,
      "step": 10900
    },
    {
      "epoch": 5.715408805031447,
      "grad_norm": 0.2664753496646881,
      "learning_rate": 4.492673449815257e-05,
      "loss": 0.4073,
      "num_input_tokens_seen": 7137272,
      "step": 10905
    },
    {
      "epoch": 5.718029350104822,
      "grad_norm": 0.553916871547699,
      "learning_rate": 4.4919827389674496e-05,
      "loss": 0.6078,
      "num_input_tokens_seen": 7140152,
      "step": 10910
    },
    {
      "epoch": 5.720649895178197,
      "grad_norm": 0.48646363615989685,
      "learning_rate": 4.491291611419579e-05,
      "loss": 0.4581,
      "num_input_tokens_seen": 7143256,
      "step": 10915
    },
    {
      "epoch": 5.723270440251572,
      "grad_norm": 0.35198506712913513,
      "learning_rate": 4.4906000673162216e-05,
      "loss": 0.5275,
      "num_input_tokens_seen": 7146456,
      "step": 10920
    },
    {
      "epoch": 5.725890985324948,
      "grad_norm": 0.354449063539505,
      "learning_rate": 4.48990810680204e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 7149432,
      "step": 10925
    },
    {
      "epoch": 5.728511530398323,
      "grad_norm": 0.49102890491485596,
      "learning_rate": 4.4892157300217844e-05,
      "loss": 0.4443,
      "num_input_tokens_seen": 7155640,
      "step": 10930
    },
    {
      "epoch": 5.731132075471698,
      "grad_norm": 0.39199960231781006,
      "learning_rate": 4.488522937120293e-05,
      "loss": 0.4605,
      "num_input_tokens_seen": 7158136,
      "step": 10935
    },
    {
      "epoch": 5.733752620545073,
      "grad_norm": 0.9080693125724792,
      "learning_rate": 4.4878297282424886e-05,
      "loss": 0.5984,
      "num_input_tokens_seen": 7161208,
      "step": 10940
    },
    {
      "epoch": 5.736373165618449,
      "grad_norm": 0.19771555066108704,
      "learning_rate": 4.4871361035333836e-05,
      "loss": 0.3075,
      "num_input_tokens_seen": 7167160,
      "step": 10945
    },
    {
      "epoch": 5.738993710691824,
      "grad_norm": 0.5759824514389038,
      "learning_rate": 4.486442063138076e-05,
      "loss": 0.4136,
      "num_input_tokens_seen": 7170072,
      "step": 10950
    },
    {
      "epoch": 5.7416142557651995,
      "grad_norm": 0.4360141456127167,
      "learning_rate": 4.4857476072017516e-05,
      "loss": 0.4495,
      "num_input_tokens_seen": 7173368,
      "step": 10955
    },
    {
      "epoch": 5.744234800838575,
      "grad_norm": 0.4063698649406433,
      "learning_rate": 4.485052735869682e-05,
      "loss": 0.5534,
      "num_input_tokens_seen": 7176504,
      "step": 10960
    },
    {
      "epoch": 5.74685534591195,
      "grad_norm": 0.5311194062232971,
      "learning_rate": 4.484357449287228e-05,
      "loss": 0.6052,
      "num_input_tokens_seen": 7179672,
      "step": 10965
    },
    {
      "epoch": 5.749475890985325,
      "grad_norm": 0.36797598004341125,
      "learning_rate": 4.483661747599833e-05,
      "loss": 0.3535,
      "num_input_tokens_seen": 7182360,
      "step": 10970
    },
    {
      "epoch": 5.7520964360587,
      "grad_norm": 0.4299255609512329,
      "learning_rate": 4.482965630953031e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 7186040,
      "step": 10975
    },
    {
      "epoch": 5.754716981132075,
      "grad_norm": 0.4480965733528137,
      "learning_rate": 4.4822690994924425e-05,
      "loss": 0.4177,
      "num_input_tokens_seen": 7189368,
      "step": 10980
    },
    {
      "epoch": 5.7573375262054505,
      "grad_norm": 0.37792137265205383,
      "learning_rate": 4.481572153363773e-05,
      "loss": 0.4346,
      "num_input_tokens_seen": 7192664,
      "step": 10985
    },
    {
      "epoch": 5.759958071278826,
      "grad_norm": 0.3372003436088562,
      "learning_rate": 4.480874792712817e-05,
      "loss": 0.4122,
      "num_input_tokens_seen": 7196120,
      "step": 10990
    },
    {
      "epoch": 5.762578616352201,
      "grad_norm": 0.595985472202301,
      "learning_rate": 4.480177017685452e-05,
      "loss": 0.4678,
      "num_input_tokens_seen": 7199992,
      "step": 10995
    },
    {
      "epoch": 5.765199161425577,
      "grad_norm": 0.5969774723052979,
      "learning_rate": 4.479478828427647e-05,
      "loss": 0.6301,
      "num_input_tokens_seen": 7203096,
      "step": 11000
    },
    {
      "epoch": 5.767819706498952,
      "grad_norm": 0.42261478304862976,
      "learning_rate": 4.478780225085454e-05,
      "loss": 0.538,
      "num_input_tokens_seen": 7206200,
      "step": 11005
    },
    {
      "epoch": 5.770440251572327,
      "grad_norm": 0.8256618976593018,
      "learning_rate": 4.4780812078050125e-05,
      "loss": 0.5548,
      "num_input_tokens_seen": 7209656,
      "step": 11010
    },
    {
      "epoch": 5.773060796645702,
      "grad_norm": 0.4483974874019623,
      "learning_rate": 4.47738177673255e-05,
      "loss": 0.6272,
      "num_input_tokens_seen": 7213272,
      "step": 11015
    },
    {
      "epoch": 5.7756813417190775,
      "grad_norm": 0.39738985896110535,
      "learning_rate": 4.4766819320143785e-05,
      "loss": 0.4682,
      "num_input_tokens_seen": 7216728,
      "step": 11020
    },
    {
      "epoch": 5.778301886792453,
      "grad_norm": 0.749086320400238,
      "learning_rate": 4.475981673796899e-05,
      "loss": 0.4601,
      "num_input_tokens_seen": 7219480,
      "step": 11025
    },
    {
      "epoch": 5.780922431865828,
      "grad_norm": 0.26353979110717773,
      "learning_rate": 4.475281002226596e-05,
      "loss": 0.4576,
      "num_input_tokens_seen": 7224376,
      "step": 11030
    },
    {
      "epoch": 5.783542976939203,
      "grad_norm": 0.37517860531806946,
      "learning_rate": 4.4745799174500414e-05,
      "loss": 0.3533,
      "num_input_tokens_seen": 7227704,
      "step": 11035
    },
    {
      "epoch": 5.786163522012579,
      "grad_norm": 0.3634648025035858,
      "learning_rate": 4.473878419613896e-05,
      "loss": 0.4081,
      "num_input_tokens_seen": 7231928,
      "step": 11040
    },
    {
      "epoch": 5.788784067085954,
      "grad_norm": 0.43260782957077026,
      "learning_rate": 4.473176508864905e-05,
      "loss": 0.4048,
      "num_input_tokens_seen": 7236984,
      "step": 11045
    },
    {
      "epoch": 5.7914046121593294,
      "grad_norm": 0.5418317914009094,
      "learning_rate": 4.4724741853498995e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 7239704,
      "step": 11050
    },
    {
      "epoch": 5.794025157232705,
      "grad_norm": 0.6812034845352173,
      "learning_rate": 4.471771449215797e-05,
      "loss": 0.4677,
      "num_input_tokens_seen": 7242936,
      "step": 11055
    },
    {
      "epoch": 5.79664570230608,
      "grad_norm": 0.3342374265193939,
      "learning_rate": 4.471068300609603e-05,
      "loss": 0.5258,
      "num_input_tokens_seen": 7245976,
      "step": 11060
    },
    {
      "epoch": 5.799266247379455,
      "grad_norm": 1.2693040370941162,
      "learning_rate": 4.4703647396784065e-05,
      "loss": 0.5345,
      "num_input_tokens_seen": 7249304,
      "step": 11065
    },
    {
      "epoch": 5.80188679245283,
      "grad_norm": 0.5600707530975342,
      "learning_rate": 4.469660766569387e-05,
      "loss": 0.56,
      "num_input_tokens_seen": 7252344,
      "step": 11070
    },
    {
      "epoch": 5.804507337526205,
      "grad_norm": 0.3092419505119324,
      "learning_rate": 4.4689563814298057e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 7257144,
      "step": 11075
    },
    {
      "epoch": 5.8071278825995805,
      "grad_norm": 0.4816408157348633,
      "learning_rate": 4.468251584407012e-05,
      "loss": 0.6645,
      "num_input_tokens_seen": 7260376,
      "step": 11080
    },
    {
      "epoch": 5.809748427672956,
      "grad_norm": 0.45609939098358154,
      "learning_rate": 4.467546375648443e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 7263256,
      "step": 11085
    },
    {
      "epoch": 5.812368972746331,
      "grad_norm": 0.47986724972724915,
      "learning_rate": 4.4668407553016176e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 7265688,
      "step": 11090
    },
    {
      "epoch": 5.814989517819707,
      "grad_norm": 0.271295964717865,
      "learning_rate": 4.466134723514146e-05,
      "loss": 0.501,
      "num_input_tokens_seen": 7268280,
      "step": 11095
    },
    {
      "epoch": 5.817610062893082,
      "grad_norm": 0.44990140199661255,
      "learning_rate": 4.465428280433721e-05,
      "loss": 0.4038,
      "num_input_tokens_seen": 7271640,
      "step": 11100
    },
    {
      "epoch": 5.820230607966457,
      "grad_norm": 0.44239166378974915,
      "learning_rate": 4.464721426208123e-05,
      "loss": 0.2821,
      "num_input_tokens_seen": 7274040,
      "step": 11105
    },
    {
      "epoch": 5.822851153039832,
      "grad_norm": 0.3065037131309509,
      "learning_rate": 4.464014160985217e-05,
      "loss": 0.5483,
      "num_input_tokens_seen": 7276888,
      "step": 11110
    },
    {
      "epoch": 5.8254716981132075,
      "grad_norm": 0.5482978224754333,
      "learning_rate": 4.463306484912955e-05,
      "loss": 0.5737,
      "num_input_tokens_seen": 7280280,
      "step": 11115
    },
    {
      "epoch": 5.828092243186583,
      "grad_norm": 0.29657435417175293,
      "learning_rate": 4.462598398139375e-05,
      "loss": 0.633,
      "num_input_tokens_seen": 7283448,
      "step": 11120
    },
    {
      "epoch": 5.830712788259958,
      "grad_norm": 0.5055733919143677,
      "learning_rate": 4.4618899008126005e-05,
      "loss": 0.7401,
      "num_input_tokens_seen": 7286168,
      "step": 11125
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 0.4251452684402466,
      "learning_rate": 4.461180993080841e-05,
      "loss": 0.5213,
      "num_input_tokens_seen": 7289688,
      "step": 11130
    },
    {
      "epoch": 5.835953878406709,
      "grad_norm": 0.4356450140476227,
      "learning_rate": 4.460471675092393e-05,
      "loss": 0.5424,
      "num_input_tokens_seen": 7292664,
      "step": 11135
    },
    {
      "epoch": 5.838574423480084,
      "grad_norm": 0.5462637543678284,
      "learning_rate": 4.4597619469956356e-05,
      "loss": 0.4512,
      "num_input_tokens_seen": 7295928,
      "step": 11140
    },
    {
      "epoch": 5.841194968553459,
      "grad_norm": 0.39859235286712646,
      "learning_rate": 4.459051808939038e-05,
      "loss": 0.5124,
      "num_input_tokens_seen": 7298424,
      "step": 11145
    },
    {
      "epoch": 5.843815513626835,
      "grad_norm": 0.713616669178009,
      "learning_rate": 4.458341261071151e-05,
      "loss": 0.4704,
      "num_input_tokens_seen": 7301624,
      "step": 11150
    },
    {
      "epoch": 5.84643605870021,
      "grad_norm": 0.5065685510635376,
      "learning_rate": 4.457630303540615e-05,
      "loss": 0.6832,
      "num_input_tokens_seen": 7304536,
      "step": 11155
    },
    {
      "epoch": 5.849056603773585,
      "grad_norm": 0.8121477365493774,
      "learning_rate": 4.456918936496154e-05,
      "loss": 0.5103,
      "num_input_tokens_seen": 7306904,
      "step": 11160
    },
    {
      "epoch": 5.85167714884696,
      "grad_norm": 0.5886672735214233,
      "learning_rate": 4.456207160086575e-05,
      "loss": 0.4678,
      "num_input_tokens_seen": 7308952,
      "step": 11165
    },
    {
      "epoch": 5.854297693920335,
      "grad_norm": 0.3685808777809143,
      "learning_rate": 4.4554949744607775e-05,
      "loss": 0.6402,
      "num_input_tokens_seen": 7312600,
      "step": 11170
    },
    {
      "epoch": 5.8569182389937104,
      "grad_norm": 0.5016460418701172,
      "learning_rate": 4.4547823797677405e-05,
      "loss": 0.5242,
      "num_input_tokens_seen": 7315416,
      "step": 11175
    },
    {
      "epoch": 5.859538784067086,
      "grad_norm": 0.47827380895614624,
      "learning_rate": 4.45406937615653e-05,
      "loss": 0.5314,
      "num_input_tokens_seen": 7318296,
      "step": 11180
    },
    {
      "epoch": 5.862159329140461,
      "grad_norm": 0.8304495811462402,
      "learning_rate": 4.4533559637763006e-05,
      "loss": 0.4186,
      "num_input_tokens_seen": 7320728,
      "step": 11185
    },
    {
      "epoch": 5.864779874213837,
      "grad_norm": 0.34375670552253723,
      "learning_rate": 4.452642142776288e-05,
      "loss": 0.4696,
      "num_input_tokens_seen": 7323928,
      "step": 11190
    },
    {
      "epoch": 5.867400419287212,
      "grad_norm": 0.4058869183063507,
      "learning_rate": 4.451927913305816e-05,
      "loss": 0.5104,
      "num_input_tokens_seen": 7327064,
      "step": 11195
    },
    {
      "epoch": 5.870020964360587,
      "grad_norm": 0.43264442682266235,
      "learning_rate": 4.4512132755142935e-05,
      "loss": 0.4775,
      "num_input_tokens_seen": 7329656,
      "step": 11200
    },
    {
      "epoch": 5.872641509433962,
      "grad_norm": 0.43517622351646423,
      "learning_rate": 4.450498229551214e-05,
      "loss": 0.4302,
      "num_input_tokens_seen": 7332760,
      "step": 11205
    },
    {
      "epoch": 5.8752620545073375,
      "grad_norm": 0.5404523015022278,
      "learning_rate": 4.449782775566158e-05,
      "loss": 0.5854,
      "num_input_tokens_seen": 7336344,
      "step": 11210
    },
    {
      "epoch": 5.877882599580713,
      "grad_norm": 0.421774685382843,
      "learning_rate": 4.449066913708789e-05,
      "loss": 0.5659,
      "num_input_tokens_seen": 7339064,
      "step": 11215
    },
    {
      "epoch": 5.880503144654088,
      "grad_norm": 0.7698339819908142,
      "learning_rate": 4.448350644128858e-05,
      "loss": 0.6764,
      "num_input_tokens_seen": 7341848,
      "step": 11220
    },
    {
      "epoch": 5.883123689727463,
      "grad_norm": 0.5428721904754639,
      "learning_rate": 4.4476339669762e-05,
      "loss": 0.5879,
      "num_input_tokens_seen": 7344632,
      "step": 11225
    },
    {
      "epoch": 5.885744234800838,
      "grad_norm": 0.42930057644844055,
      "learning_rate": 4.446916882400737e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 7347512,
      "step": 11230
    },
    {
      "epoch": 5.888364779874214,
      "grad_norm": 0.7533403038978577,
      "learning_rate": 4.446199390552472e-05,
      "loss": 0.5396,
      "num_input_tokens_seen": 7350424,
      "step": 11235
    },
    {
      "epoch": 5.890985324947589,
      "grad_norm": 0.47250157594680786,
      "learning_rate": 4.445481491581499e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 7354232,
      "step": 11240
    },
    {
      "epoch": 5.893605870020965,
      "grad_norm": 0.7710446715354919,
      "learning_rate": 4.444763185637993e-05,
      "loss": 0.4622,
      "num_input_tokens_seen": 7356632,
      "step": 11245
    },
    {
      "epoch": 5.89622641509434,
      "grad_norm": 0.42072299122810364,
      "learning_rate": 4.4440444728722166e-05,
      "loss": 0.408,
      "num_input_tokens_seen": 7359256,
      "step": 11250
    },
    {
      "epoch": 5.898846960167715,
      "grad_norm": 0.6182384490966797,
      "learning_rate": 4.4433253534345136e-05,
      "loss": 0.3704,
      "num_input_tokens_seen": 7362200,
      "step": 11255
    },
    {
      "epoch": 5.90146750524109,
      "grad_norm": 0.4108121991157532,
      "learning_rate": 4.442605827475318e-05,
      "loss": 0.3432,
      "num_input_tokens_seen": 7364664,
      "step": 11260
    },
    {
      "epoch": 5.904088050314465,
      "grad_norm": 0.5526174306869507,
      "learning_rate": 4.4418858951451457e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 7368248,
      "step": 11265
    },
    {
      "epoch": 5.90670859538784,
      "grad_norm": 0.4993804395198822,
      "learning_rate": 4.441165556594599e-05,
      "loss": 0.4294,
      "num_input_tokens_seen": 7371064,
      "step": 11270
    },
    {
      "epoch": 5.909329140461216,
      "grad_norm": 0.5207759737968445,
      "learning_rate": 4.440444811974363e-05,
      "loss": 0.4605,
      "num_input_tokens_seen": 7374168,
      "step": 11275
    },
    {
      "epoch": 5.911949685534591,
      "grad_norm": 0.31509870290756226,
      "learning_rate": 4.43972366143521e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 7377336,
      "step": 11280
    },
    {
      "epoch": 5.914570230607967,
      "grad_norm": 1.4684604406356812,
      "learning_rate": 4.4390021051279975e-05,
      "loss": 0.5157,
      "num_input_tokens_seen": 7380280,
      "step": 11285
    },
    {
      "epoch": 5.917190775681342,
      "grad_norm": 0.40282660722732544,
      "learning_rate": 4.438280143203665e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 7383864,
      "step": 11290
    },
    {
      "epoch": 5.919811320754717,
      "grad_norm": 0.30428698658943176,
      "learning_rate": 4.43755777581324e-05,
      "loss": 0.319,
      "num_input_tokens_seen": 7387256,
      "step": 11295
    },
    {
      "epoch": 5.922431865828092,
      "grad_norm": 0.29507485032081604,
      "learning_rate": 4.4368350031078334e-05,
      "loss": 0.3362,
      "num_input_tokens_seen": 7391032,
      "step": 11300
    },
    {
      "epoch": 5.9250524109014675,
      "grad_norm": 0.310884952545166,
      "learning_rate": 4.43611182523864e-05,
      "loss": 0.4217,
      "num_input_tokens_seen": 7393848,
      "step": 11305
    },
    {
      "epoch": 5.927672955974843,
      "grad_norm": 0.6772649884223938,
      "learning_rate": 4.4353882423569416e-05,
      "loss": 0.4843,
      "num_input_tokens_seen": 7397688,
      "step": 11310
    },
    {
      "epoch": 5.930293501048218,
      "grad_norm": 0.47944989800453186,
      "learning_rate": 4.434664254614103e-05,
      "loss": 0.4747,
      "num_input_tokens_seen": 7400664,
      "step": 11315
    },
    {
      "epoch": 5.932914046121593,
      "grad_norm": 0.3841644525527954,
      "learning_rate": 4.4339398621615734e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 7403832,
      "step": 11320
    },
    {
      "epoch": 5.935534591194968,
      "grad_norm": 0.4565750062465668,
      "learning_rate": 4.433215065150889e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 7406744,
      "step": 11325
    },
    {
      "epoch": 5.938155136268344,
      "grad_norm": 0.5946252346038818,
      "learning_rate": 4.4324898637336674e-05,
      "loss": 0.5803,
      "num_input_tokens_seen": 7410232,
      "step": 11330
    },
    {
      "epoch": 5.940775681341719,
      "grad_norm": 0.5361924767494202,
      "learning_rate": 4.431764258061615e-05,
      "loss": 0.5145,
      "num_input_tokens_seen": 7413688,
      "step": 11335
    },
    {
      "epoch": 5.943396226415095,
      "grad_norm": 0.37693172693252563,
      "learning_rate": 4.431038248286517e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 7417112,
      "step": 11340
    },
    {
      "epoch": 5.94601677148847,
      "grad_norm": 0.321346253156662,
      "learning_rate": 4.430311834560248e-05,
      "loss": 0.6123,
      "num_input_tokens_seen": 7419576,
      "step": 11345
    },
    {
      "epoch": 5.948637316561845,
      "grad_norm": 0.3893888294696808,
      "learning_rate": 4.4295850170347664e-05,
      "loss": 0.4349,
      "num_input_tokens_seen": 7423256,
      "step": 11350
    },
    {
      "epoch": 5.95125786163522,
      "grad_norm": 0.5521588325500488,
      "learning_rate": 4.428857795862113e-05,
      "loss": 0.3446,
      "num_input_tokens_seen": 7428792,
      "step": 11355
    },
    {
      "epoch": 5.953878406708595,
      "grad_norm": 0.41720810532569885,
      "learning_rate": 4.428130171194414e-05,
      "loss": 0.4255,
      "num_input_tokens_seen": 7432376,
      "step": 11360
    },
    {
      "epoch": 5.95649895178197,
      "grad_norm": 0.3210456669330597,
      "learning_rate": 4.427402143183879e-05,
      "loss": 0.5763,
      "num_input_tokens_seen": 7435992,
      "step": 11365
    },
    {
      "epoch": 5.959119496855346,
      "grad_norm": 0.5896318554878235,
      "learning_rate": 4.426673711982806e-05,
      "loss": 0.4422,
      "num_input_tokens_seen": 7438872,
      "step": 11370
    },
    {
      "epoch": 5.961740041928721,
      "grad_norm": 0.8563124537467957,
      "learning_rate": 4.425944877743573e-05,
      "loss": 0.4693,
      "num_input_tokens_seen": 7442008,
      "step": 11375
    },
    {
      "epoch": 5.964360587002097,
      "grad_norm": 1.1850183010101318,
      "learning_rate": 4.425215640618644e-05,
      "loss": 0.5112,
      "num_input_tokens_seen": 7444408,
      "step": 11380
    },
    {
      "epoch": 5.966981132075472,
      "grad_norm": 0.5447959899902344,
      "learning_rate": 4.424486000760567e-05,
      "loss": 0.381,
      "num_input_tokens_seen": 7448440,
      "step": 11385
    },
    {
      "epoch": 5.969601677148847,
      "grad_norm": 1.188124656677246,
      "learning_rate": 4.423755958321974e-05,
      "loss": 0.4339,
      "num_input_tokens_seen": 7451480,
      "step": 11390
    },
    {
      "epoch": 5.972222222222222,
      "grad_norm": 0.3472577929496765,
      "learning_rate": 4.423025513455582e-05,
      "loss": 0.4592,
      "num_input_tokens_seen": 7454424,
      "step": 11395
    },
    {
      "epoch": 5.9748427672955975,
      "grad_norm": 0.7813378572463989,
      "learning_rate": 4.422294666314192e-05,
      "loss": 0.5038,
      "num_input_tokens_seen": 7456824,
      "step": 11400
    },
    {
      "epoch": 5.977463312368973,
      "grad_norm": 0.3198428750038147,
      "learning_rate": 4.421563417050688e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 7459608,
      "step": 11405
    },
    {
      "epoch": 5.980083857442348,
      "grad_norm": 0.5285534858703613,
      "learning_rate": 4.42083176581804e-05,
      "loss": 0.5234,
      "num_input_tokens_seen": 7462840,
      "step": 11410
    },
    {
      "epoch": 5.982704402515723,
      "grad_norm": 0.33984988927841187,
      "learning_rate": 4.4200997127693e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 7466456,
      "step": 11415
    },
    {
      "epoch": 5.985324947589098,
      "grad_norm": 0.34621986746788025,
      "learning_rate": 4.419367258057606e-05,
      "loss": 0.3189,
      "num_input_tokens_seen": 7469304,
      "step": 11420
    },
    {
      "epoch": 5.987945492662474,
      "grad_norm": 0.8445386290550232,
      "learning_rate": 4.418634401836179e-05,
      "loss": 0.5182,
      "num_input_tokens_seen": 7472248,
      "step": 11425
    },
    {
      "epoch": 5.990566037735849,
      "grad_norm": 0.5526623129844666,
      "learning_rate": 4.417901144258324e-05,
      "loss": 0.552,
      "num_input_tokens_seen": 7475032,
      "step": 11430
    },
    {
      "epoch": 5.993186582809225,
      "grad_norm": 0.3556753695011139,
      "learning_rate": 4.4171674854774305e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 7478008,
      "step": 11435
    },
    {
      "epoch": 5.9958071278826,
      "grad_norm": 0.5246224403381348,
      "learning_rate": 4.416433425646971e-05,
      "loss": 0.4358,
      "num_input_tokens_seen": 7481112,
      "step": 11440
    },
    {
      "epoch": 5.998427672955975,
      "grad_norm": 0.3436727523803711,
      "learning_rate": 4.415698964920503e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 7484024,
      "step": 11445
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.45426803827285767,
      "eval_runtime": 9.6492,
      "eval_samples_per_second": 87.883,
      "eval_steps_per_second": 21.971,
      "num_input_tokens_seen": 7485248,
      "step": 11448
    },
    {
      "epoch": 6.00104821802935,
      "grad_norm": 0.3767646849155426,
      "learning_rate": 4.414964103451667e-05,
      "loss": 0.4058,
      "num_input_tokens_seen": 7486304,
      "step": 11450
    },
    {
      "epoch": 6.003668763102725,
      "grad_norm": 0.318769246339798,
      "learning_rate": 4.414228841394187e-05,
      "loss": 0.5194,
      "num_input_tokens_seen": 7489504,
      "step": 11455
    },
    {
      "epoch": 6.0062893081761,
      "grad_norm": 0.9719060659408569,
      "learning_rate": 4.413493178901872e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 7492896,
      "step": 11460
    },
    {
      "epoch": 6.008909853249476,
      "grad_norm": 0.3405599892139435,
      "learning_rate": 4.412757116128614e-05,
      "loss": 0.4792,
      "num_input_tokens_seen": 7495584,
      "step": 11465
    },
    {
      "epoch": 6.011530398322851,
      "grad_norm": 0.3980178236961365,
      "learning_rate": 4.41202065322839e-05,
      "loss": 0.5403,
      "num_input_tokens_seen": 7498208,
      "step": 11470
    },
    {
      "epoch": 6.014150943396227,
      "grad_norm": 0.20364169776439667,
      "learning_rate": 4.411283790355259e-05,
      "loss": 0.4045,
      "num_input_tokens_seen": 7501664,
      "step": 11475
    },
    {
      "epoch": 6.016771488469602,
      "grad_norm": 0.3299148976802826,
      "learning_rate": 4.410546527663364e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 7504416,
      "step": 11480
    },
    {
      "epoch": 6.019392033542977,
      "grad_norm": 0.5081472396850586,
      "learning_rate": 4.409808865306931e-05,
      "loss": 0.4799,
      "num_input_tokens_seen": 7507264,
      "step": 11485
    },
    {
      "epoch": 6.022012578616352,
      "grad_norm": 0.6329009532928467,
      "learning_rate": 4.409070803440273e-05,
      "loss": 0.5714,
      "num_input_tokens_seen": 7510432,
      "step": 11490
    },
    {
      "epoch": 6.0246331236897275,
      "grad_norm": 0.39774155616760254,
      "learning_rate": 4.4083323422177824e-05,
      "loss": 0.5378,
      "num_input_tokens_seen": 7513856,
      "step": 11495
    },
    {
      "epoch": 6.027253668763103,
      "grad_norm": 0.6385399699211121,
      "learning_rate": 4.407593481793937e-05,
      "loss": 0.6133,
      "num_input_tokens_seen": 7516736,
      "step": 11500
    },
    {
      "epoch": 6.029874213836478,
      "grad_norm": 0.5972583293914795,
      "learning_rate": 4.406854222323298e-05,
      "loss": 0.4469,
      "num_input_tokens_seen": 7519968,
      "step": 11505
    },
    {
      "epoch": 6.032494758909853,
      "grad_norm": 0.18275637924671173,
      "learning_rate": 4.406114563960511e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 7523040,
      "step": 11510
    },
    {
      "epoch": 6.035115303983228,
      "grad_norm": 0.3340458571910858,
      "learning_rate": 4.4053745068603026e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 7526464,
      "step": 11515
    },
    {
      "epoch": 6.037735849056604,
      "grad_norm": 0.38656386733055115,
      "learning_rate": 4.404634051177485e-05,
      "loss": 0.337,
      "num_input_tokens_seen": 7529376,
      "step": 11520
    },
    {
      "epoch": 6.040356394129979,
      "grad_norm": 0.3338925540447235,
      "learning_rate": 4.403893197066953e-05,
      "loss": 0.3986,
      "num_input_tokens_seen": 7535904,
      "step": 11525
    },
    {
      "epoch": 6.0429769392033545,
      "grad_norm": 0.3445969820022583,
      "learning_rate": 4.4031519446836855e-05,
      "loss": 0.3311,
      "num_input_tokens_seen": 7538784,
      "step": 11530
    },
    {
      "epoch": 6.04559748427673,
      "grad_norm": 0.5163747072219849,
      "learning_rate": 4.402410294182743e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 7542368,
      "step": 11535
    },
    {
      "epoch": 6.048218029350105,
      "grad_norm": 0.3496477007865906,
      "learning_rate": 4.4016682457192714e-05,
      "loss": 0.3365,
      "num_input_tokens_seen": 7545248,
      "step": 11540
    },
    {
      "epoch": 6.05083857442348,
      "grad_norm": 0.5343130230903625,
      "learning_rate": 4.400925799448497e-05,
      "loss": 0.4172,
      "num_input_tokens_seen": 7548160,
      "step": 11545
    },
    {
      "epoch": 6.053459119496855,
      "grad_norm": 0.26198530197143555,
      "learning_rate": 4.400182955525733e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 7551008,
      "step": 11550
    },
    {
      "epoch": 6.05607966457023,
      "grad_norm": 0.7174700498580933,
      "learning_rate": 4.3994397141063734e-05,
      "loss": 0.5378,
      "num_input_tokens_seen": 7554016,
      "step": 11555
    },
    {
      "epoch": 6.058700209643606,
      "grad_norm": 0.27194225788116455,
      "learning_rate": 4.398696075345896e-05,
      "loss": 0.4745,
      "num_input_tokens_seen": 7557376,
      "step": 11560
    },
    {
      "epoch": 6.061320754716981,
      "grad_norm": 0.4912940263748169,
      "learning_rate": 4.39795203939986e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 7561184,
      "step": 11565
    },
    {
      "epoch": 6.063941299790357,
      "grad_norm": 0.32263651490211487,
      "learning_rate": 4.397207606423911e-05,
      "loss": 0.3928,
      "num_input_tokens_seen": 7564832,
      "step": 11570
    },
    {
      "epoch": 6.066561844863732,
      "grad_norm": 0.361705482006073,
      "learning_rate": 4.396462776573775e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 7569152,
      "step": 11575
    },
    {
      "epoch": 6.069182389937107,
      "grad_norm": 0.6187112927436829,
      "learning_rate": 4.395717550005262e-05,
      "loss": 0.3391,
      "num_input_tokens_seen": 7572352,
      "step": 11580
    },
    {
      "epoch": 6.071802935010482,
      "grad_norm": 0.6365352272987366,
      "learning_rate": 4.394971926874265e-05,
      "loss": 0.4062,
      "num_input_tokens_seen": 7574944,
      "step": 11585
    },
    {
      "epoch": 6.0744234800838575,
      "grad_norm": 0.432871013879776,
      "learning_rate": 4.3942259073367604e-05,
      "loss": 0.435,
      "num_input_tokens_seen": 7577632,
      "step": 11590
    },
    {
      "epoch": 6.077044025157233,
      "grad_norm": 0.4565064013004303,
      "learning_rate": 4.3934794915488065e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 7580512,
      "step": 11595
    },
    {
      "epoch": 6.079664570230608,
      "grad_norm": 0.46940556168556213,
      "learning_rate": 4.392732679666544e-05,
      "loss": 0.4679,
      "num_input_tokens_seen": 7583584,
      "step": 11600
    },
    {
      "epoch": 6.082285115303983,
      "grad_norm": 0.3808470070362091,
      "learning_rate": 4.3919854718461986e-05,
      "loss": 0.5259,
      "num_input_tokens_seen": 7586688,
      "step": 11605
    },
    {
      "epoch": 6.084905660377358,
      "grad_norm": 0.8017213344573975,
      "learning_rate": 4.391237868244078e-05,
      "loss": 0.4508,
      "num_input_tokens_seen": 7589792,
      "step": 11610
    },
    {
      "epoch": 6.087526205450734,
      "grad_norm": 0.3486974537372589,
      "learning_rate": 4.390489869016571e-05,
      "loss": 0.4545,
      "num_input_tokens_seen": 7593440,
      "step": 11615
    },
    {
      "epoch": 6.090146750524109,
      "grad_norm": 0.4177249073982239,
      "learning_rate": 4.389741474320152e-05,
      "loss": 0.5826,
      "num_input_tokens_seen": 7596544,
      "step": 11620
    },
    {
      "epoch": 6.0927672955974845,
      "grad_norm": 0.6154914498329163,
      "learning_rate": 4.388992684311376e-05,
      "loss": 0.4055,
      "num_input_tokens_seen": 7602816,
      "step": 11625
    },
    {
      "epoch": 6.09538784067086,
      "grad_norm": 0.5861823558807373,
      "learning_rate": 4.388243499146881e-05,
      "loss": 0.5101,
      "num_input_tokens_seen": 7605248,
      "step": 11630
    },
    {
      "epoch": 6.098008385744235,
      "grad_norm": 0.3399370014667511,
      "learning_rate": 4.3874939189833876e-05,
      "loss": 0.5597,
      "num_input_tokens_seen": 7608640,
      "step": 11635
    },
    {
      "epoch": 6.10062893081761,
      "grad_norm": 0.4133242666721344,
      "learning_rate": 4.3867439439777005e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 7612224,
      "step": 11640
    },
    {
      "epoch": 6.103249475890985,
      "grad_norm": 0.6039639115333557,
      "learning_rate": 4.385993574286705e-05,
      "loss": 0.5072,
      "num_input_tokens_seen": 7614688,
      "step": 11645
    },
    {
      "epoch": 6.10587002096436,
      "grad_norm": 0.3493906855583191,
      "learning_rate": 4.3852428100673705e-05,
      "loss": 0.3631,
      "num_input_tokens_seen": 7618624,
      "step": 11650
    },
    {
      "epoch": 6.1084905660377355,
      "grad_norm": 0.5545619130134583,
      "learning_rate": 4.3844916514767484e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 7621568,
      "step": 11655
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 1.3029568195343018,
      "learning_rate": 4.383740098671972e-05,
      "loss": 0.6256,
      "num_input_tokens_seen": 7623712,
      "step": 11660
    },
    {
      "epoch": 6.113731656184487,
      "grad_norm": 0.38328444957733154,
      "learning_rate": 4.3829881518102576e-05,
      "loss": 0.524,
      "num_input_tokens_seen": 7626944,
      "step": 11665
    },
    {
      "epoch": 6.116352201257862,
      "grad_norm": 0.42457371950149536,
      "learning_rate": 4.3822358110489045e-05,
      "loss": 0.5199,
      "num_input_tokens_seen": 7629504,
      "step": 11670
    },
    {
      "epoch": 6.118972746331237,
      "grad_norm": 0.7274208068847656,
      "learning_rate": 4.381483076545294e-05,
      "loss": 0.3462,
      "num_input_tokens_seen": 7632320,
      "step": 11675
    },
    {
      "epoch": 6.121593291404612,
      "grad_norm": 0.5396454334259033,
      "learning_rate": 4.380729948456888e-05,
      "loss": 0.3979,
      "num_input_tokens_seen": 7634944,
      "step": 11680
    },
    {
      "epoch": 6.1242138364779874,
      "grad_norm": 0.3918992877006531,
      "learning_rate": 4.3799764269412336e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 7638752,
      "step": 11685
    },
    {
      "epoch": 6.126834381551363,
      "grad_norm": 0.46667852997779846,
      "learning_rate": 4.379222512155958e-05,
      "loss": 0.4757,
      "num_input_tokens_seen": 7641920,
      "step": 11690
    },
    {
      "epoch": 6.129454926624738,
      "grad_norm": 0.44954943656921387,
      "learning_rate": 4.378468204258773e-05,
      "loss": 0.5226,
      "num_input_tokens_seen": 7645248,
      "step": 11695
    },
    {
      "epoch": 6.132075471698113,
      "grad_norm": 0.66866135597229,
      "learning_rate": 4.37771350340747e-05,
      "loss": 0.5843,
      "num_input_tokens_seen": 7652256,
      "step": 11700
    },
    {
      "epoch": 6.134696016771488,
      "grad_norm": 0.6028106212615967,
      "learning_rate": 4.376958409759925e-05,
      "loss": 0.383,
      "num_input_tokens_seen": 7654976,
      "step": 11705
    },
    {
      "epoch": 6.137316561844864,
      "grad_norm": 0.2239263504743576,
      "learning_rate": 4.376202923474093e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 7658944,
      "step": 11710
    },
    {
      "epoch": 6.139937106918239,
      "grad_norm": 0.35968640446662903,
      "learning_rate": 4.3754470447080145e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 7662112,
      "step": 11715
    },
    {
      "epoch": 6.1425576519916145,
      "grad_norm": 0.5810678601264954,
      "learning_rate": 4.374690773619811e-05,
      "loss": 0.5194,
      "num_input_tokens_seen": 7664480,
      "step": 11720
    },
    {
      "epoch": 6.14517819706499,
      "grad_norm": 0.2392897605895996,
      "learning_rate": 4.373934110367685e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 7667520,
      "step": 11725
    },
    {
      "epoch": 6.147798742138365,
      "grad_norm": 0.3252517580986023,
      "learning_rate": 4.373177055109922e-05,
      "loss": 0.4699,
      "num_input_tokens_seen": 7670880,
      "step": 11730
    },
    {
      "epoch": 6.15041928721174,
      "grad_norm": 0.5820589661598206,
      "learning_rate": 4.3724196080048896e-05,
      "loss": 0.4553,
      "num_input_tokens_seen": 7673856,
      "step": 11735
    },
    {
      "epoch": 6.153039832285115,
      "grad_norm": 0.5057092905044556,
      "learning_rate": 4.3716617692110364e-05,
      "loss": 0.5021,
      "num_input_tokens_seen": 7676864,
      "step": 11740
    },
    {
      "epoch": 6.15566037735849,
      "grad_norm": 0.3522721827030182,
      "learning_rate": 4.3709035388868943e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 7679296,
      "step": 11745
    },
    {
      "epoch": 6.1582809224318655,
      "grad_norm": 0.6985883116722107,
      "learning_rate": 4.3701449171910767e-05,
      "loss": 0.4861,
      "num_input_tokens_seen": 7682912,
      "step": 11750
    },
    {
      "epoch": 6.160901467505241,
      "grad_norm": 0.2836645245552063,
      "learning_rate": 4.3693859042822774e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 7686592,
      "step": 11755
    },
    {
      "epoch": 6.163522012578617,
      "grad_norm": 0.4054059684276581,
      "learning_rate": 4.368626500319275e-05,
      "loss": 0.4633,
      "num_input_tokens_seen": 7689440,
      "step": 11760
    },
    {
      "epoch": 6.166142557651992,
      "grad_norm": 0.22466039657592773,
      "learning_rate": 4.367866705460927e-05,
      "loss": 0.5376,
      "num_input_tokens_seen": 7693184,
      "step": 11765
    },
    {
      "epoch": 6.168763102725367,
      "grad_norm": 0.5027027726173401,
      "learning_rate": 4.367106519866173e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 7696672,
      "step": 11770
    },
    {
      "epoch": 6.171383647798742,
      "grad_norm": 0.5119404792785645,
      "learning_rate": 4.366345943694037e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 7700352,
      "step": 11775
    },
    {
      "epoch": 6.174004192872117,
      "grad_norm": 0.5319349765777588,
      "learning_rate": 4.365584977103622e-05,
      "loss": 0.563,
      "num_input_tokens_seen": 7703008,
      "step": 11780
    },
    {
      "epoch": 6.176624737945493,
      "grad_norm": 0.38847246766090393,
      "learning_rate": 4.364823620254113e-05,
      "loss": 0.468,
      "num_input_tokens_seen": 7705824,
      "step": 11785
    },
    {
      "epoch": 6.179245283018868,
      "grad_norm": 0.4602726697921753,
      "learning_rate": 4.364061873304778e-05,
      "loss": 0.3923,
      "num_input_tokens_seen": 7708672,
      "step": 11790
    },
    {
      "epoch": 6.181865828092243,
      "grad_norm": 0.30768021941185,
      "learning_rate": 4.363299736414965e-05,
      "loss": 0.4042,
      "num_input_tokens_seen": 7712480,
      "step": 11795
    },
    {
      "epoch": 6.184486373165618,
      "grad_norm": 0.5875333547592163,
      "learning_rate": 4.3625372097441045e-05,
      "loss": 0.3248,
      "num_input_tokens_seen": 7715296,
      "step": 11800
    },
    {
      "epoch": 6.187106918238993,
      "grad_norm": 0.3698466718196869,
      "learning_rate": 4.3617742934517094e-05,
      "loss": 0.4863,
      "num_input_tokens_seen": 7719808,
      "step": 11805
    },
    {
      "epoch": 6.189727463312369,
      "grad_norm": 0.355432391166687,
      "learning_rate": 4.361010987697373e-05,
      "loss": 0.4855,
      "num_input_tokens_seen": 7722208,
      "step": 11810
    },
    {
      "epoch": 6.1923480083857445,
      "grad_norm": 0.8574251532554626,
      "learning_rate": 4.360247292640768e-05,
      "loss": 0.3047,
      "num_input_tokens_seen": 7726816,
      "step": 11815
    },
    {
      "epoch": 6.19496855345912,
      "grad_norm": 0.46685826778411865,
      "learning_rate": 4.359483208441653e-05,
      "loss": 0.5027,
      "num_input_tokens_seen": 7729504,
      "step": 11820
    },
    {
      "epoch": 6.197589098532495,
      "grad_norm": 0.5855221152305603,
      "learning_rate": 4.3587187352598646e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 7732096,
      "step": 11825
    },
    {
      "epoch": 6.20020964360587,
      "grad_norm": 0.3987751305103302,
      "learning_rate": 4.357953873255323e-05,
      "loss": 0.3757,
      "num_input_tokens_seen": 7735008,
      "step": 11830
    },
    {
      "epoch": 6.202830188679245,
      "grad_norm": 0.4124700725078583,
      "learning_rate": 4.357188622588027e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 7738432,
      "step": 11835
    },
    {
      "epoch": 6.20545073375262,
      "grad_norm": 0.5850765109062195,
      "learning_rate": 4.356422983418059e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 7741280,
      "step": 11840
    },
    {
      "epoch": 6.2080712788259955,
      "grad_norm": 0.4105289578437805,
      "learning_rate": 4.3556569559055824e-05,
      "loss": 0.5005,
      "num_input_tokens_seen": 7744512,
      "step": 11845
    },
    {
      "epoch": 6.210691823899371,
      "grad_norm": 0.4238388240337372,
      "learning_rate": 4.3548905402108406e-05,
      "loss": 0.4975,
      "num_input_tokens_seen": 7747904,
      "step": 11850
    },
    {
      "epoch": 6.213312368972747,
      "grad_norm": 0.5524535775184631,
      "learning_rate": 4.35412373649416e-05,
      "loss": 0.4606,
      "num_input_tokens_seen": 7751168,
      "step": 11855
    },
    {
      "epoch": 6.215932914046122,
      "grad_norm": 0.8872973918914795,
      "learning_rate": 4.353356544915947e-05,
      "loss": 0.4145,
      "num_input_tokens_seen": 7754368,
      "step": 11860
    },
    {
      "epoch": 6.218553459119497,
      "grad_norm": 0.5983509421348572,
      "learning_rate": 4.352588965636689e-05,
      "loss": 0.5351,
      "num_input_tokens_seen": 7757056,
      "step": 11865
    },
    {
      "epoch": 6.221174004192872,
      "grad_norm": 0.240596741437912,
      "learning_rate": 4.351820998816954e-05,
      "loss": 0.3093,
      "num_input_tokens_seen": 7759904,
      "step": 11870
    },
    {
      "epoch": 6.223794549266247,
      "grad_norm": 0.4175744354724884,
      "learning_rate": 4.3510526446173935e-05,
      "loss": 0.5827,
      "num_input_tokens_seen": 7764768,
      "step": 11875
    },
    {
      "epoch": 6.226415094339623,
      "grad_norm": 0.4583416283130646,
      "learning_rate": 4.350283903198737e-05,
      "loss": 0.5023,
      "num_input_tokens_seen": 7768416,
      "step": 11880
    },
    {
      "epoch": 6.229035639412998,
      "grad_norm": 0.5267279148101807,
      "learning_rate": 4.349514774721797e-05,
      "loss": 0.4013,
      "num_input_tokens_seen": 7770912,
      "step": 11885
    },
    {
      "epoch": 6.231656184486373,
      "grad_norm": 0.31787607073783875,
      "learning_rate": 4.348745259347467e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 7774240,
      "step": 11890
    },
    {
      "epoch": 6.234276729559748,
      "grad_norm": 0.3771696984767914,
      "learning_rate": 4.347975357236721e-05,
      "loss": 0.586,
      "num_input_tokens_seen": 7777728,
      "step": 11895
    },
    {
      "epoch": 6.236897274633123,
      "grad_norm": 0.6205476522445679,
      "learning_rate": 4.347205068550612e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 7780512,
      "step": 11900
    },
    {
      "epoch": 6.239517819706499,
      "grad_norm": 0.5953615307807922,
      "learning_rate": 4.3464343934502774e-05,
      "loss": 0.7631,
      "num_input_tokens_seen": 7783104,
      "step": 11905
    },
    {
      "epoch": 6.2421383647798745,
      "grad_norm": 0.3839385509490967,
      "learning_rate": 4.345663332096932e-05,
      "loss": 0.4194,
      "num_input_tokens_seen": 7785920,
      "step": 11910
    },
    {
      "epoch": 6.24475890985325,
      "grad_norm": 0.6708263158798218,
      "learning_rate": 4.344891884651875e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 7789024,
      "step": 11915
    },
    {
      "epoch": 6.247379454926625,
      "grad_norm": 0.3374168872833252,
      "learning_rate": 4.344120051276483e-05,
      "loss": 0.4751,
      "num_input_tokens_seen": 7792640,
      "step": 11920
    },
    {
      "epoch": 6.25,
      "grad_norm": 0.28744545578956604,
      "learning_rate": 4.3433478321322155e-05,
      "loss": 0.4395,
      "num_input_tokens_seen": 7796576,
      "step": 11925
    },
    {
      "epoch": 6.252620545073375,
      "grad_norm": 0.3283081352710724,
      "learning_rate": 4.3425752273806106e-05,
      "loss": 0.4486,
      "num_input_tokens_seen": 7799264,
      "step": 11930
    },
    {
      "epoch": 6.25524109014675,
      "grad_norm": 0.43708157539367676,
      "learning_rate": 4.34180223718329e-05,
      "loss": 0.4144,
      "num_input_tokens_seen": 7802016,
      "step": 11935
    },
    {
      "epoch": 6.2578616352201255,
      "grad_norm": 0.6060647964477539,
      "learning_rate": 4.341028861701953e-05,
      "loss": 0.5373,
      "num_input_tokens_seen": 7804320,
      "step": 11940
    },
    {
      "epoch": 6.260482180293501,
      "grad_norm": 0.27665817737579346,
      "learning_rate": 4.3402551010983825e-05,
      "loss": 0.5396,
      "num_input_tokens_seen": 7807648,
      "step": 11945
    },
    {
      "epoch": 6.263102725366876,
      "grad_norm": 0.7241175770759583,
      "learning_rate": 4.33948095553444e-05,
      "loss": 0.337,
      "num_input_tokens_seen": 7815680,
      "step": 11950
    },
    {
      "epoch": 6.265723270440252,
      "grad_norm": 0.42659229040145874,
      "learning_rate": 4.338706425172067e-05,
      "loss": 0.4094,
      "num_input_tokens_seen": 7818528,
      "step": 11955
    },
    {
      "epoch": 6.268343815513627,
      "grad_norm": 0.47423097491264343,
      "learning_rate": 4.3379315101732866e-05,
      "loss": 0.3485,
      "num_input_tokens_seen": 7821728,
      "step": 11960
    },
    {
      "epoch": 6.270964360587002,
      "grad_norm": 0.7126997709274292,
      "learning_rate": 4.3371562107002026e-05,
      "loss": 0.4951,
      "num_input_tokens_seen": 7825440,
      "step": 11965
    },
    {
      "epoch": 6.273584905660377,
      "grad_norm": 0.45591410994529724,
      "learning_rate": 4.3363805269149984e-05,
      "loss": 0.4037,
      "num_input_tokens_seen": 7828192,
      "step": 11970
    },
    {
      "epoch": 6.276205450733753,
      "grad_norm": 1.075126051902771,
      "learning_rate": 4.335604458979939e-05,
      "loss": 0.4255,
      "num_input_tokens_seen": 7831040,
      "step": 11975
    },
    {
      "epoch": 6.278825995807128,
      "grad_norm": 0.5738202929496765,
      "learning_rate": 4.3348280070573686e-05,
      "loss": 0.5884,
      "num_input_tokens_seen": 7833888,
      "step": 11980
    },
    {
      "epoch": 6.281446540880503,
      "grad_norm": 0.6552482843399048,
      "learning_rate": 4.334051171309712e-05,
      "loss": 0.4084,
      "num_input_tokens_seen": 7836928,
      "step": 11985
    },
    {
      "epoch": 6.284067085953878,
      "grad_norm": 0.3480679392814636,
      "learning_rate": 4.333273951899474e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 7839808,
      "step": 11990
    },
    {
      "epoch": 6.286687631027253,
      "grad_norm": 0.7090106010437012,
      "learning_rate": 4.33249634898924e-05,
      "loss": 0.569,
      "num_input_tokens_seen": 7842656,
      "step": 11995
    },
    {
      "epoch": 6.289308176100629,
      "grad_norm": 0.5546310544013977,
      "learning_rate": 4.331718362741677e-05,
      "loss": 0.4743,
      "num_input_tokens_seen": 7845536,
      "step": 12000
    },
    {
      "epoch": 6.2919287211740045,
      "grad_norm": 0.44575196504592896,
      "learning_rate": 4.330939993319529e-05,
      "loss": 0.5051,
      "num_input_tokens_seen": 7849632,
      "step": 12005
    },
    {
      "epoch": 6.29454926624738,
      "grad_norm": 0.26016107201576233,
      "learning_rate": 4.330161240885623e-05,
      "loss": 0.3747,
      "num_input_tokens_seen": 7852480,
      "step": 12010
    },
    {
      "epoch": 6.297169811320755,
      "grad_norm": 0.5884751677513123,
      "learning_rate": 4.3293821056028654e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 7856192,
      "step": 12015
    },
    {
      "epoch": 6.29979035639413,
      "grad_norm": 0.33961910009384155,
      "learning_rate": 4.3286025876342415e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 7859744,
      "step": 12020
    },
    {
      "epoch": 6.302410901467505,
      "grad_norm": 0.45559167861938477,
      "learning_rate": 4.327822687142819e-05,
      "loss": 0.5524,
      "num_input_tokens_seen": 7862848,
      "step": 12025
    },
    {
      "epoch": 6.30503144654088,
      "grad_norm": 0.5529646277427673,
      "learning_rate": 4.327042404291742e-05,
      "loss": 0.5546,
      "num_input_tokens_seen": 7865408,
      "step": 12030
    },
    {
      "epoch": 6.3076519916142555,
      "grad_norm": 0.4436994194984436,
      "learning_rate": 4.326261739244239e-05,
      "loss": 0.3491,
      "num_input_tokens_seen": 7868064,
      "step": 12035
    },
    {
      "epoch": 6.310272536687631,
      "grad_norm": 0.5294432640075684,
      "learning_rate": 4.325480692163615e-05,
      "loss": 0.5068,
      "num_input_tokens_seen": 7870400,
      "step": 12040
    },
    {
      "epoch": 6.312893081761006,
      "grad_norm": 0.37990301847457886,
      "learning_rate": 4.324699263213256e-05,
      "loss": 0.3713,
      "num_input_tokens_seen": 7875008,
      "step": 12045
    },
    {
      "epoch": 6.315513626834382,
      "grad_norm": 0.3697409927845001,
      "learning_rate": 4.3239174525566285e-05,
      "loss": 0.4967,
      "num_input_tokens_seen": 7878560,
      "step": 12050
    },
    {
      "epoch": 6.318134171907757,
      "grad_norm": 0.3842674493789673,
      "learning_rate": 4.3231352603572784e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 7881984,
      "step": 12055
    },
    {
      "epoch": 6.320754716981132,
      "grad_norm": 0.3327830135822296,
      "learning_rate": 4.322352686778832e-05,
      "loss": 0.4516,
      "num_input_tokens_seen": 7885056,
      "step": 12060
    },
    {
      "epoch": 6.323375262054507,
      "grad_norm": 0.2291073203086853,
      "learning_rate": 4.3215697319849926e-05,
      "loss": 0.5292,
      "num_input_tokens_seen": 7888928,
      "step": 12065
    },
    {
      "epoch": 6.325995807127883,
      "grad_norm": 0.46736660599708557,
      "learning_rate": 4.3207863961395476e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 7891648,
      "step": 12070
    },
    {
      "epoch": 6.328616352201258,
      "grad_norm": 0.4993501305580139,
      "learning_rate": 4.320002679406361e-05,
      "loss": 0.3949,
      "num_input_tokens_seen": 7894432,
      "step": 12075
    },
    {
      "epoch": 6.331236897274633,
      "grad_norm": 0.6646471619606018,
      "learning_rate": 4.319218581949378e-05,
      "loss": 0.425,
      "num_input_tokens_seen": 7896608,
      "step": 12080
    },
    {
      "epoch": 6.333857442348008,
      "grad_norm": 0.6341317296028137,
      "learning_rate": 4.318434103932622e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 7900000,
      "step": 12085
    },
    {
      "epoch": 6.336477987421383,
      "grad_norm": 0.7452888488769531,
      "learning_rate": 4.317649245520197e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 7902848,
      "step": 12090
    },
    {
      "epoch": 6.339098532494759,
      "grad_norm": 0.3435037136077881,
      "learning_rate": 4.3168640068762875e-05,
      "loss": 0.3385,
      "num_input_tokens_seen": 7905760,
      "step": 12095
    },
    {
      "epoch": 6.3417190775681345,
      "grad_norm": 0.37447890639305115,
      "learning_rate": 4.316078388165155e-05,
      "loss": 0.5192,
      "num_input_tokens_seen": 7909408,
      "step": 12100
    },
    {
      "epoch": 6.34433962264151,
      "grad_norm": 0.4202411472797394,
      "learning_rate": 4.315292389551143e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 7912768,
      "step": 12105
    },
    {
      "epoch": 6.346960167714885,
      "grad_norm": 0.5739821791648865,
      "learning_rate": 4.3145060111986726e-05,
      "loss": 0.5553,
      "num_input_tokens_seen": 7915488,
      "step": 12110
    },
    {
      "epoch": 6.34958071278826,
      "grad_norm": 0.37764161825180054,
      "learning_rate": 4.313719253272246e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 7918592,
      "step": 12115
    },
    {
      "epoch": 6.352201257861635,
      "grad_norm": 0.5081576704978943,
      "learning_rate": 4.3129321159364436e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 7921952,
      "step": 12120
    },
    {
      "epoch": 6.35482180293501,
      "grad_norm": 0.3740832805633545,
      "learning_rate": 4.312144599355925e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 7925632,
      "step": 12125
    },
    {
      "epoch": 6.3574423480083855,
      "grad_norm": 0.41060754656791687,
      "learning_rate": 4.31135670369543e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 7928896,
      "step": 12130
    },
    {
      "epoch": 6.360062893081761,
      "grad_norm": 0.29243889451026917,
      "learning_rate": 4.310568429119779e-05,
      "loss": 0.4721,
      "num_input_tokens_seen": 7932064,
      "step": 12135
    },
    {
      "epoch": 6.362683438155136,
      "grad_norm": 0.45657598972320557,
      "learning_rate": 4.309779775793868e-05,
      "loss": 0.517,
      "num_input_tokens_seen": 7934688,
      "step": 12140
    },
    {
      "epoch": 6.365303983228512,
      "grad_norm": 0.6227694749832153,
      "learning_rate": 4.308990743882675e-05,
      "loss": 0.529,
      "num_input_tokens_seen": 7937536,
      "step": 12145
    },
    {
      "epoch": 6.367924528301887,
      "grad_norm": 0.9094924926757812,
      "learning_rate": 4.308201333551256e-05,
      "loss": 0.3715,
      "num_input_tokens_seen": 7940832,
      "step": 12150
    },
    {
      "epoch": 6.370545073375262,
      "grad_norm": 0.3993486166000366,
      "learning_rate": 4.3074115449647476e-05,
      "loss": 0.3794,
      "num_input_tokens_seen": 7943392,
      "step": 12155
    },
    {
      "epoch": 6.373165618448637,
      "grad_norm": 0.8138371706008911,
      "learning_rate": 4.306621378288364e-05,
      "loss": 0.4202,
      "num_input_tokens_seen": 7946688,
      "step": 12160
    },
    {
      "epoch": 6.3757861635220126,
      "grad_norm": 0.4832499325275421,
      "learning_rate": 4.305830833687399e-05,
      "loss": 0.5631,
      "num_input_tokens_seen": 7948960,
      "step": 12165
    },
    {
      "epoch": 6.378406708595388,
      "grad_norm": 0.5960817337036133,
      "learning_rate": 4.305039911327225e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 7952832,
      "step": 12170
    },
    {
      "epoch": 6.381027253668763,
      "grad_norm": 0.6064738035202026,
      "learning_rate": 4.304248611373296e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 7955616,
      "step": 12175
    },
    {
      "epoch": 6.383647798742138,
      "grad_norm": 0.29312312602996826,
      "learning_rate": 4.30345693399114e-05,
      "loss": 0.489,
      "num_input_tokens_seen": 7959456,
      "step": 12180
    },
    {
      "epoch": 6.386268343815513,
      "grad_norm": 0.7405017614364624,
      "learning_rate": 4.30266487934637e-05,
      "loss": 0.6667,
      "num_input_tokens_seen": 7964128,
      "step": 12185
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 0.623728334903717,
      "learning_rate": 4.3018724476046714e-05,
      "loss": 0.4395,
      "num_input_tokens_seen": 7966816,
      "step": 12190
    },
    {
      "epoch": 6.3915094339622645,
      "grad_norm": 0.39225178956985474,
      "learning_rate": 4.301079638931814e-05,
      "loss": 0.657,
      "num_input_tokens_seen": 7969600,
      "step": 12195
    },
    {
      "epoch": 6.39412997903564,
      "grad_norm": 0.5166759490966797,
      "learning_rate": 4.3002864534936444e-05,
      "loss": 0.4564,
      "num_input_tokens_seen": 7972672,
      "step": 12200
    },
    {
      "epoch": 6.396750524109015,
      "grad_norm": 0.6144890785217285,
      "learning_rate": 4.2994928914560874e-05,
      "loss": 0.5008,
      "num_input_tokens_seen": 7975744,
      "step": 12205
    },
    {
      "epoch": 6.39937106918239,
      "grad_norm": 0.4611741900444031,
      "learning_rate": 4.298698952985147e-05,
      "loss": 0.551,
      "num_input_tokens_seen": 7978816,
      "step": 12210
    },
    {
      "epoch": 6.401991614255765,
      "grad_norm": 0.8045106530189514,
      "learning_rate": 4.297904638246906e-05,
      "loss": 0.5168,
      "num_input_tokens_seen": 7981504,
      "step": 12215
    },
    {
      "epoch": 6.40461215932914,
      "grad_norm": 0.5577869415283203,
      "learning_rate": 4.297109947407527e-05,
      "loss": 0.5206,
      "num_input_tokens_seen": 7984000,
      "step": 12220
    },
    {
      "epoch": 6.4072327044025155,
      "grad_norm": 0.4853866696357727,
      "learning_rate": 4.2963148806332486e-05,
      "loss": 0.4366,
      "num_input_tokens_seen": 7986624,
      "step": 12225
    },
    {
      "epoch": 6.409853249475891,
      "grad_norm": 0.5199059844017029,
      "learning_rate": 4.2955194380903915e-05,
      "loss": 0.4003,
      "num_input_tokens_seen": 7989728,
      "step": 12230
    },
    {
      "epoch": 6.412473794549266,
      "grad_norm": 0.45977094769477844,
      "learning_rate": 4.2947236199453524e-05,
      "loss": 0.5616,
      "num_input_tokens_seen": 7993696,
      "step": 12235
    },
    {
      "epoch": 6.415094339622642,
      "grad_norm": 0.4362972378730774,
      "learning_rate": 4.2939274263646065e-05,
      "loss": 0.4161,
      "num_input_tokens_seen": 7996192,
      "step": 12240
    },
    {
      "epoch": 6.417714884696017,
      "grad_norm": 0.47784948348999023,
      "learning_rate": 4.29313085751471e-05,
      "loss": 0.4967,
      "num_input_tokens_seen": 8000192,
      "step": 12245
    },
    {
      "epoch": 6.420335429769392,
      "grad_norm": 0.43283161520957947,
      "learning_rate": 4.292333913562294e-05,
      "loss": 0.5588,
      "num_input_tokens_seen": 8004256,
      "step": 12250
    },
    {
      "epoch": 6.422955974842767,
      "grad_norm": 0.5583256483078003,
      "learning_rate": 4.291536594674073e-05,
      "loss": 0.6285,
      "num_input_tokens_seen": 8007200,
      "step": 12255
    },
    {
      "epoch": 6.4255765199161425,
      "grad_norm": 0.8206962943077087,
      "learning_rate": 4.2907389010168344e-05,
      "loss": 0.4539,
      "num_input_tokens_seen": 8011328,
      "step": 12260
    },
    {
      "epoch": 6.428197064989518,
      "grad_norm": 0.5774081945419312,
      "learning_rate": 4.2899408327574475e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 8014240,
      "step": 12265
    },
    {
      "epoch": 6.430817610062893,
      "grad_norm": 0.3741203546524048,
      "learning_rate": 4.289142390062859e-05,
      "loss": 0.5167,
      "num_input_tokens_seen": 8017600,
      "step": 12270
    },
    {
      "epoch": 6.433438155136268,
      "grad_norm": 0.3942808210849762,
      "learning_rate": 4.2883435731000946e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 8020416,
      "step": 12275
    },
    {
      "epoch": 6.436058700209643,
      "grad_norm": 0.33124682307243347,
      "learning_rate": 4.287544382036256e-05,
      "loss": 0.4965,
      "num_input_tokens_seen": 8023328,
      "step": 12280
    },
    {
      "epoch": 6.438679245283019,
      "grad_norm": 0.3069738745689392,
      "learning_rate": 4.286744817038527e-05,
      "loss": 0.5929,
      "num_input_tokens_seen": 8026336,
      "step": 12285
    },
    {
      "epoch": 6.441299790356394,
      "grad_norm": 0.6146788001060486,
      "learning_rate": 4.2859448782741666e-05,
      "loss": 0.5673,
      "num_input_tokens_seen": 8029728,
      "step": 12290
    },
    {
      "epoch": 6.44392033542977,
      "grad_norm": 0.42158517241477966,
      "learning_rate": 4.2851445659105126e-05,
      "loss": 0.4642,
      "num_input_tokens_seen": 8033088,
      "step": 12295
    },
    {
      "epoch": 6.446540880503145,
      "grad_norm": 0.35719403624534607,
      "learning_rate": 4.284343880114982e-05,
      "loss": 0.4237,
      "num_input_tokens_seen": 8036352,
      "step": 12300
    },
    {
      "epoch": 6.44916142557652,
      "grad_norm": 0.42154020071029663,
      "learning_rate": 4.2835428210550676e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 8039616,
      "step": 12305
    },
    {
      "epoch": 6.451781970649895,
      "grad_norm": 1.534941554069519,
      "learning_rate": 4.282741388898344e-05,
      "loss": 0.5504,
      "num_input_tokens_seen": 8042240,
      "step": 12310
    },
    {
      "epoch": 6.45440251572327,
      "grad_norm": 0.37755224108695984,
      "learning_rate": 4.281939583812459e-05,
      "loss": 0.5278,
      "num_input_tokens_seen": 8046816,
      "step": 12315
    },
    {
      "epoch": 6.4570230607966455,
      "grad_norm": 0.25412362813949585,
      "learning_rate": 4.281137405965143e-05,
      "loss": 0.4881,
      "num_input_tokens_seen": 8050528,
      "step": 12320
    },
    {
      "epoch": 6.459643605870021,
      "grad_norm": 0.4351080656051636,
      "learning_rate": 4.280334855524203e-05,
      "loss": 0.5956,
      "num_input_tokens_seen": 8054336,
      "step": 12325
    },
    {
      "epoch": 6.462264150943396,
      "grad_norm": 0.5647839903831482,
      "learning_rate": 4.279531932657521e-05,
      "loss": 0.5229,
      "num_input_tokens_seen": 8056832,
      "step": 12330
    },
    {
      "epoch": 6.464884696016772,
      "grad_norm": 0.9804746508598328,
      "learning_rate": 4.278728637533062e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 8059616,
      "step": 12335
    },
    {
      "epoch": 6.467505241090147,
      "grad_norm": 0.5200055837631226,
      "learning_rate": 4.2779249703188626e-05,
      "loss": 0.489,
      "num_input_tokens_seen": 8062400,
      "step": 12340
    },
    {
      "epoch": 6.470125786163522,
      "grad_norm": 0.42712199687957764,
      "learning_rate": 4.277120931183045e-05,
      "loss": 0.4711,
      "num_input_tokens_seen": 8065824,
      "step": 12345
    },
    {
      "epoch": 6.472746331236897,
      "grad_norm": 0.623309850692749,
      "learning_rate": 4.276316520293803e-05,
      "loss": 0.3521,
      "num_input_tokens_seen": 8068544,
      "step": 12350
    },
    {
      "epoch": 6.4753668763102725,
      "grad_norm": 0.6922481060028076,
      "learning_rate": 4.27551173781941e-05,
      "loss": 0.5058,
      "num_input_tokens_seen": 8071712,
      "step": 12355
    },
    {
      "epoch": 6.477987421383648,
      "grad_norm": 0.5865228176116943,
      "learning_rate": 4.274706583928216e-05,
      "loss": 0.5019,
      "num_input_tokens_seen": 8075360,
      "step": 12360
    },
    {
      "epoch": 6.480607966457023,
      "grad_norm": 0.6781355142593384,
      "learning_rate": 4.273901058788654e-05,
      "loss": 0.4595,
      "num_input_tokens_seen": 8078720,
      "step": 12365
    },
    {
      "epoch": 6.483228511530398,
      "grad_norm": 0.2868400514125824,
      "learning_rate": 4.2730951625692275e-05,
      "loss": 0.4287,
      "num_input_tokens_seen": 8081504,
      "step": 12370
    },
    {
      "epoch": 6.485849056603773,
      "grad_norm": 0.550747811794281,
      "learning_rate": 4.27228889543852e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 8084608,
      "step": 12375
    },
    {
      "epoch": 6.488469601677149,
      "grad_norm": 0.42509719729423523,
      "learning_rate": 4.2714822575651966e-05,
      "loss": 0.5296,
      "num_input_tokens_seen": 8087936,
      "step": 12380
    },
    {
      "epoch": 6.491090146750524,
      "grad_norm": 0.5323606133460999,
      "learning_rate": 4.270675249117994e-05,
      "loss": 0.3155,
      "num_input_tokens_seen": 8091648,
      "step": 12385
    },
    {
      "epoch": 6.4937106918239,
      "grad_norm": 0.27809759974479675,
      "learning_rate": 4.269867870265731e-05,
      "loss": 0.5062,
      "num_input_tokens_seen": 8094848,
      "step": 12390
    },
    {
      "epoch": 6.496331236897275,
      "grad_norm": 0.2983499765396118,
      "learning_rate": 4.269060121177301e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 8097952,
      "step": 12395
    },
    {
      "epoch": 6.49895178197065,
      "grad_norm": 0.658173680305481,
      "learning_rate": 4.2682520020216764e-05,
      "loss": 0.4858,
      "num_input_tokens_seen": 8102016,
      "step": 12400
    },
    {
      "epoch": 6.501572327044025,
      "grad_norm": 0.3633674383163452,
      "learning_rate": 4.267443512967906e-05,
      "loss": 0.7151,
      "num_input_tokens_seen": 8105568,
      "step": 12405
    },
    {
      "epoch": 6.5041928721174,
      "grad_norm": 0.48467445373535156,
      "learning_rate": 4.2666346541851165e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 8108864,
      "step": 12410
    },
    {
      "epoch": 6.506813417190775,
      "grad_norm": 0.36526572704315186,
      "learning_rate": 4.265825425842512e-05,
      "loss": 0.4078,
      "num_input_tokens_seen": 8112032,
      "step": 12415
    },
    {
      "epoch": 6.509433962264151,
      "grad_norm": 0.4363238215446472,
      "learning_rate": 4.265015828109375e-05,
      "loss": 0.7571,
      "num_input_tokens_seen": 8115136,
      "step": 12420
    },
    {
      "epoch": 6.512054507337526,
      "grad_norm": 0.37377047538757324,
      "learning_rate": 4.264205861155063e-05,
      "loss": 0.5054,
      "num_input_tokens_seen": 8120704,
      "step": 12425
    },
    {
      "epoch": 6.514675052410902,
      "grad_norm": 0.5135635137557983,
      "learning_rate": 4.263395525149011e-05,
      "loss": 0.4273,
      "num_input_tokens_seen": 8123904,
      "step": 12430
    },
    {
      "epoch": 6.517295597484277,
      "grad_norm": 0.27415794134140015,
      "learning_rate": 4.262584820260734e-05,
      "loss": 0.5588,
      "num_input_tokens_seen": 8127264,
      "step": 12435
    },
    {
      "epoch": 6.519916142557652,
      "grad_norm": 0.5269688963890076,
      "learning_rate": 4.261773746659821e-05,
      "loss": 0.5337,
      "num_input_tokens_seen": 8130048,
      "step": 12440
    },
    {
      "epoch": 6.522536687631027,
      "grad_norm": 0.3518599271774292,
      "learning_rate": 4.260962304515939e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 8133440,
      "step": 12445
    },
    {
      "epoch": 6.5251572327044025,
      "grad_norm": 0.5788725018501282,
      "learning_rate": 4.2601504939988335e-05,
      "loss": 0.5462,
      "num_input_tokens_seen": 8137216,
      "step": 12450
    },
    {
      "epoch": 6.527777777777778,
      "grad_norm": 0.4075828790664673,
      "learning_rate": 4.259338315278325e-05,
      "loss": 0.6057,
      "num_input_tokens_seen": 8141408,
      "step": 12455
    },
    {
      "epoch": 6.530398322851153,
      "grad_norm": 0.43422451615333557,
      "learning_rate": 4.258525768524313e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 8144832,
      "step": 12460
    },
    {
      "epoch": 6.533018867924528,
      "grad_norm": 0.5795870423316956,
      "learning_rate": 4.2577128539067725e-05,
      "loss": 0.4448,
      "num_input_tokens_seen": 8147584,
      "step": 12465
    },
    {
      "epoch": 6.535639412997903,
      "grad_norm": 0.5948998928070068,
      "learning_rate": 4.256899571595756e-05,
      "loss": 0.4904,
      "num_input_tokens_seen": 8150528,
      "step": 12470
    },
    {
      "epoch": 6.538259958071279,
      "grad_norm": 0.3242965042591095,
      "learning_rate": 4.256085921761393e-05,
      "loss": 0.4973,
      "num_input_tokens_seen": 8154304,
      "step": 12475
    },
    {
      "epoch": 6.540880503144654,
      "grad_norm": 0.37611308693885803,
      "learning_rate": 4.25527190457389e-05,
      "loss": 0.6003,
      "num_input_tokens_seen": 8156832,
      "step": 12480
    },
    {
      "epoch": 6.54350104821803,
      "grad_norm": 0.539097785949707,
      "learning_rate": 4.2544575202035285e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 8159456,
      "step": 12485
    },
    {
      "epoch": 6.546121593291405,
      "grad_norm": 0.7888802289962769,
      "learning_rate": 4.253642768820671e-05,
      "loss": 0.4733,
      "num_input_tokens_seen": 8162720,
      "step": 12490
    },
    {
      "epoch": 6.54874213836478,
      "grad_norm": 0.283356249332428,
      "learning_rate": 4.252827650595752e-05,
      "loss": 0.4363,
      "num_input_tokens_seen": 8165824,
      "step": 12495
    },
    {
      "epoch": 6.551362683438155,
      "grad_norm": 0.42965567111968994,
      "learning_rate": 4.252012165699286e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 8168928,
      "step": 12500
    },
    {
      "epoch": 6.55398322851153,
      "grad_norm": 0.46688878536224365,
      "learning_rate": 4.2511963143018634e-05,
      "loss": 0.4397,
      "num_input_tokens_seen": 8172320,
      "step": 12505
    },
    {
      "epoch": 6.556603773584905,
      "grad_norm": 0.383841872215271,
      "learning_rate": 4.25038009657415e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 8175360,
      "step": 12510
    },
    {
      "epoch": 6.559224318658281,
      "grad_norm": 0.34547287225723267,
      "learning_rate": 4.24956351268689e-05,
      "loss": 0.6285,
      "num_input_tokens_seen": 8178464,
      "step": 12515
    },
    {
      "epoch": 6.561844863731656,
      "grad_norm": 0.5943605899810791,
      "learning_rate": 4.2487465628109036e-05,
      "loss": 0.4305,
      "num_input_tokens_seen": 8181536,
      "step": 12520
    },
    {
      "epoch": 6.564465408805032,
      "grad_norm": 0.6868793368339539,
      "learning_rate": 4.247929247117087e-05,
      "loss": 0.5906,
      "num_input_tokens_seen": 8187328,
      "step": 12525
    },
    {
      "epoch": 6.567085953878407,
      "grad_norm": 0.5562564134597778,
      "learning_rate": 4.247111565776413e-05,
      "loss": 0.5176,
      "num_input_tokens_seen": 8190400,
      "step": 12530
    },
    {
      "epoch": 6.569706498951782,
      "grad_norm": 0.3329634666442871,
      "learning_rate": 4.246293518959932e-05,
      "loss": 0.3696,
      "num_input_tokens_seen": 8193568,
      "step": 12535
    },
    {
      "epoch": 6.572327044025157,
      "grad_norm": 0.4837157726287842,
      "learning_rate": 4.245475106838769e-05,
      "loss": 0.5714,
      "num_input_tokens_seen": 8196672,
      "step": 12540
    },
    {
      "epoch": 6.5749475890985325,
      "grad_norm": 0.7143033742904663,
      "learning_rate": 4.2446563295841266e-05,
      "loss": 0.5804,
      "num_input_tokens_seen": 8199744,
      "step": 12545
    },
    {
      "epoch": 6.577568134171908,
      "grad_norm": 0.36781030893325806,
      "learning_rate": 4.243837187367285e-05,
      "loss": 0.4669,
      "num_input_tokens_seen": 8201984,
      "step": 12550
    },
    {
      "epoch": 6.580188679245283,
      "grad_norm": 0.43630194664001465,
      "learning_rate": 4.2430176803595987e-05,
      "loss": 0.4322,
      "num_input_tokens_seen": 8205344,
      "step": 12555
    },
    {
      "epoch": 6.582809224318658,
      "grad_norm": 0.4157339334487915,
      "learning_rate": 4.2421978087324985e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 8208736,
      "step": 12560
    },
    {
      "epoch": 6.585429769392033,
      "grad_norm": 0.3472878038883209,
      "learning_rate": 4.241377572657493e-05,
      "loss": 0.5549,
      "num_input_tokens_seen": 8212832,
      "step": 12565
    },
    {
      "epoch": 6.588050314465409,
      "grad_norm": 0.3712708353996277,
      "learning_rate": 4.2405569723061654e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 8216512,
      "step": 12570
    },
    {
      "epoch": 6.590670859538784,
      "grad_norm": 0.4845564663410187,
      "learning_rate": 4.239736007850178e-05,
      "loss": 0.5647,
      "num_input_tokens_seen": 8220000,
      "step": 12575
    },
    {
      "epoch": 6.59329140461216,
      "grad_norm": 0.6971351504325867,
      "learning_rate": 4.238914679461264e-05,
      "loss": 0.3604,
      "num_input_tokens_seen": 8223840,
      "step": 12580
    },
    {
      "epoch": 6.595911949685535,
      "grad_norm": 0.9981552958488464,
      "learning_rate": 4.2380929873112386e-05,
      "loss": 0.4717,
      "num_input_tokens_seen": 8226688,
      "step": 12585
    },
    {
      "epoch": 6.59853249475891,
      "grad_norm": 0.26223647594451904,
      "learning_rate": 4.2372709315719896e-05,
      "loss": 0.4185,
      "num_input_tokens_seen": 8230208,
      "step": 12590
    },
    {
      "epoch": 6.601153039832285,
      "grad_norm": 0.5683720111846924,
      "learning_rate": 4.2364485124154806e-05,
      "loss": 0.3788,
      "num_input_tokens_seen": 8232448,
      "step": 12595
    },
    {
      "epoch": 6.60377358490566,
      "grad_norm": 0.7034668326377869,
      "learning_rate": 4.235625730013754e-05,
      "loss": 0.4533,
      "num_input_tokens_seen": 8235136,
      "step": 12600
    },
    {
      "epoch": 6.606394129979035,
      "grad_norm": 0.5330551266670227,
      "learning_rate": 4.234802584538926e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 8237856,
      "step": 12605
    },
    {
      "epoch": 6.609014675052411,
      "grad_norm": 0.5154452323913574,
      "learning_rate": 4.2339790761631884e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 8241472,
      "step": 12610
    },
    {
      "epoch": 6.611635220125786,
      "grad_norm": 0.36310112476348877,
      "learning_rate": 4.233155205058811e-05,
      "loss": 0.545,
      "num_input_tokens_seen": 8245504,
      "step": 12615
    },
    {
      "epoch": 6.614255765199162,
      "grad_norm": 0.39839431643486023,
      "learning_rate": 4.232330971398137e-05,
      "loss": 0.5023,
      "num_input_tokens_seen": 8248480,
      "step": 12620
    },
    {
      "epoch": 6.616876310272537,
      "grad_norm": 0.4917641282081604,
      "learning_rate": 4.2315063753535875e-05,
      "loss": 0.598,
      "num_input_tokens_seen": 8251968,
      "step": 12625
    },
    {
      "epoch": 6.619496855345912,
      "grad_norm": 0.48836633563041687,
      "learning_rate": 4.230681417097658e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 8254496,
      "step": 12630
    },
    {
      "epoch": 6.622117400419287,
      "grad_norm": 0.5536395907402039,
      "learning_rate": 4.229856096802921e-05,
      "loss": 0.6629,
      "num_input_tokens_seen": 8257472,
      "step": 12635
    },
    {
      "epoch": 6.6247379454926625,
      "grad_norm": 0.41888904571533203,
      "learning_rate": 4.229030414642024e-05,
      "loss": 0.5182,
      "num_input_tokens_seen": 8260768,
      "step": 12640
    },
    {
      "epoch": 6.627358490566038,
      "grad_norm": 0.3350510001182556,
      "learning_rate": 4.2282043707876896e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 8264672,
      "step": 12645
    },
    {
      "epoch": 6.629979035639413,
      "grad_norm": 0.3027591109275818,
      "learning_rate": 4.2273779654127175e-05,
      "loss": 0.3961,
      "num_input_tokens_seen": 8267424,
      "step": 12650
    },
    {
      "epoch": 6.632599580712788,
      "grad_norm": 0.4842663109302521,
      "learning_rate": 4.226551198689982e-05,
      "loss": 0.5341,
      "num_input_tokens_seen": 8270912,
      "step": 12655
    },
    {
      "epoch": 6.635220125786163,
      "grad_norm": 0.5942329168319702,
      "learning_rate": 4.225724070792434e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 8274080,
      "step": 12660
    },
    {
      "epoch": 6.637840670859539,
      "grad_norm": 0.5457887649536133,
      "learning_rate": 4.2248965818930975e-05,
      "loss": 0.3712,
      "num_input_tokens_seen": 8278656,
      "step": 12665
    },
    {
      "epoch": 6.640461215932914,
      "grad_norm": 0.47778719663619995,
      "learning_rate": 4.224068732165075e-05,
      "loss": 0.3411,
      "num_input_tokens_seen": 8282496,
      "step": 12670
    },
    {
      "epoch": 6.6430817610062896,
      "grad_norm": 0.8323506116867065,
      "learning_rate": 4.223240521781543e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 8285504,
      "step": 12675
    },
    {
      "epoch": 6.645702306079665,
      "grad_norm": 0.5531857013702393,
      "learning_rate": 4.2224119509157535e-05,
      "loss": 0.5069,
      "num_input_tokens_seen": 8288448,
      "step": 12680
    },
    {
      "epoch": 6.64832285115304,
      "grad_norm": 0.8166529536247253,
      "learning_rate": 4.221583019741033e-05,
      "loss": 0.3817,
      "num_input_tokens_seen": 8291072,
      "step": 12685
    },
    {
      "epoch": 6.650943396226415,
      "grad_norm": 0.4773726463317871,
      "learning_rate": 4.2207537284307876e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 8294144,
      "step": 12690
    },
    {
      "epoch": 6.65356394129979,
      "grad_norm": 0.25023823976516724,
      "learning_rate": 4.219924077158493e-05,
      "loss": 0.6041,
      "num_input_tokens_seen": 8297408,
      "step": 12695
    },
    {
      "epoch": 6.656184486373165,
      "grad_norm": 0.48606017231941223,
      "learning_rate": 4.219094066097703e-05,
      "loss": 0.4652,
      "num_input_tokens_seen": 8300768,
      "step": 12700
    },
    {
      "epoch": 6.658805031446541,
      "grad_norm": 0.30092233419418335,
      "learning_rate": 4.218263695422047e-05,
      "loss": 0.4241,
      "num_input_tokens_seen": 8303616,
      "step": 12705
    },
    {
      "epoch": 6.661425576519916,
      "grad_norm": 0.27950841188430786,
      "learning_rate": 4.217432965305229e-05,
      "loss": 0.551,
      "num_input_tokens_seen": 8307424,
      "step": 12710
    },
    {
      "epoch": 6.664046121593291,
      "grad_norm": 0.4649941921234131,
      "learning_rate": 4.2166018759210286e-05,
      "loss": 0.4607,
      "num_input_tokens_seen": 8310560,
      "step": 12715
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 0.555636465549469,
      "learning_rate": 4.215770427443299e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 8312992,
      "step": 12720
    },
    {
      "epoch": 6.669287211740042,
      "grad_norm": 0.5793598890304565,
      "learning_rate": 4.214938620045971e-05,
      "loss": 0.3745,
      "num_input_tokens_seen": 8316544,
      "step": 12725
    },
    {
      "epoch": 6.671907756813417,
      "grad_norm": 0.5198521614074707,
      "learning_rate": 4.214106453903048e-05,
      "loss": 0.4152,
      "num_input_tokens_seen": 8319520,
      "step": 12730
    },
    {
      "epoch": 6.6745283018867925,
      "grad_norm": 0.46971505880355835,
      "learning_rate": 4.2132739291886114e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 8323872,
      "step": 12735
    },
    {
      "epoch": 6.677148846960168,
      "grad_norm": 0.40418311953544617,
      "learning_rate": 4.212441046076814e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 8326976,
      "step": 12740
    },
    {
      "epoch": 6.679769392033543,
      "grad_norm": 0.2858564257621765,
      "learning_rate": 4.211607804741887e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 8330304,
      "step": 12745
    },
    {
      "epoch": 6.682389937106918,
      "grad_norm": 0.6359819173812866,
      "learning_rate": 4.2107742053581344e-05,
      "loss": 0.3983,
      "num_input_tokens_seen": 8333632,
      "step": 12750
    },
    {
      "epoch": 6.685010482180293,
      "grad_norm": 0.7013576626777649,
      "learning_rate": 4.209940248099935e-05,
      "loss": 0.4363,
      "num_input_tokens_seen": 8336896,
      "step": 12755
    },
    {
      "epoch": 6.687631027253669,
      "grad_norm": 0.2481996864080429,
      "learning_rate": 4.2091059331417445e-05,
      "loss": 0.4517,
      "num_input_tokens_seen": 8340512,
      "step": 12760
    },
    {
      "epoch": 6.690251572327044,
      "grad_norm": 0.31639236211776733,
      "learning_rate": 4.208271260658091e-05,
      "loss": 0.5018,
      "num_input_tokens_seen": 8343680,
      "step": 12765
    },
    {
      "epoch": 6.6928721174004195,
      "grad_norm": 0.28688347339630127,
      "learning_rate": 4.2074362308235785e-05,
      "loss": 0.4697,
      "num_input_tokens_seen": 8347456,
      "step": 12770
    },
    {
      "epoch": 6.695492662473795,
      "grad_norm": 0.2463976889848709,
      "learning_rate": 4.206600843812886e-05,
      "loss": 0.5545,
      "num_input_tokens_seen": 8351200,
      "step": 12775
    },
    {
      "epoch": 6.69811320754717,
      "grad_norm": 0.74269038438797,
      "learning_rate": 4.205765099800767e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 8354240,
      "step": 12780
    },
    {
      "epoch": 6.700733752620545,
      "grad_norm": 0.3328041136264801,
      "learning_rate": 4.204928998962049e-05,
      "loss": 0.513,
      "num_input_tokens_seen": 8357856,
      "step": 12785
    },
    {
      "epoch": 6.70335429769392,
      "grad_norm": 0.3759252727031708,
      "learning_rate": 4.204092541471636e-05,
      "loss": 0.451,
      "num_input_tokens_seen": 8361184,
      "step": 12790
    },
    {
      "epoch": 6.705974842767295,
      "grad_norm": 0.5586721301078796,
      "learning_rate": 4.203255727504504e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 8364256,
      "step": 12795
    },
    {
      "epoch": 6.7085953878406706,
      "grad_norm": 0.615740954875946,
      "learning_rate": 4.202418557235706e-05,
      "loss": 0.4499,
      "num_input_tokens_seen": 8367232,
      "step": 12800
    },
    {
      "epoch": 6.711215932914046,
      "grad_norm": 0.5355754494667053,
      "learning_rate": 4.201581030840368e-05,
      "loss": 0.5645,
      "num_input_tokens_seen": 8370688,
      "step": 12805
    },
    {
      "epoch": 6.713836477987421,
      "grad_norm": 0.25881463289260864,
      "learning_rate": 4.200743148493691e-05,
      "loss": 0.4843,
      "num_input_tokens_seen": 8374464,
      "step": 12810
    },
    {
      "epoch": 6.716457023060797,
      "grad_norm": 0.31435465812683105,
      "learning_rate": 4.199904910370951e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 8382368,
      "step": 12815
    },
    {
      "epoch": 6.719077568134172,
      "grad_norm": 0.4889432489871979,
      "learning_rate": 4.1990663166474964e-05,
      "loss": 0.4922,
      "num_input_tokens_seen": 8385248,
      "step": 12820
    },
    {
      "epoch": 6.721698113207547,
      "grad_norm": 0.6755163073539734,
      "learning_rate": 4.1982273674987524e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 8388128,
      "step": 12825
    },
    {
      "epoch": 6.7243186582809225,
      "grad_norm": 0.29584646224975586,
      "learning_rate": 4.197388063100218e-05,
      "loss": 0.4632,
      "num_input_tokens_seen": 8391008,
      "step": 12830
    },
    {
      "epoch": 6.726939203354298,
      "grad_norm": 0.36426135897636414,
      "learning_rate": 4.196548403627465e-05,
      "loss": 0.3769,
      "num_input_tokens_seen": 8393952,
      "step": 12835
    },
    {
      "epoch": 6.729559748427673,
      "grad_norm": 0.725246250629425,
      "learning_rate": 4.195708389256142e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 8396704,
      "step": 12840
    },
    {
      "epoch": 6.732180293501048,
      "grad_norm": 0.4858284592628479,
      "learning_rate": 4.194868020161969e-05,
      "loss": 0.4984,
      "num_input_tokens_seen": 8400160,
      "step": 12845
    },
    {
      "epoch": 6.734800838574423,
      "grad_norm": 0.29437750577926636,
      "learning_rate": 4.1940272965207425e-05,
      "loss": 0.4085,
      "num_input_tokens_seen": 8403168,
      "step": 12850
    },
    {
      "epoch": 6.737421383647799,
      "grad_norm": 0.359252005815506,
      "learning_rate": 4.1931862185083326e-05,
      "loss": 0.3253,
      "num_input_tokens_seen": 8405952,
      "step": 12855
    },
    {
      "epoch": 6.740041928721174,
      "grad_norm": 0.6292282342910767,
      "learning_rate": 4.192344786300682e-05,
      "loss": 0.4248,
      "num_input_tokens_seen": 8408192,
      "step": 12860
    },
    {
      "epoch": 6.7426624737945495,
      "grad_norm": 0.5106409788131714,
      "learning_rate": 4.19150300007381e-05,
      "loss": 0.3878,
      "num_input_tokens_seen": 8410688,
      "step": 12865
    },
    {
      "epoch": 6.745283018867925,
      "grad_norm": 0.4285902678966522,
      "learning_rate": 4.190660860003806e-05,
      "loss": 0.3769,
      "num_input_tokens_seen": 8414688,
      "step": 12870
    },
    {
      "epoch": 6.7479035639413,
      "grad_norm": 0.31100398302078247,
      "learning_rate": 4.189818366266841e-05,
      "loss": 0.4089,
      "num_input_tokens_seen": 8418240,
      "step": 12875
    },
    {
      "epoch": 6.750524109014675,
      "grad_norm": 1.0397324562072754,
      "learning_rate": 4.188975519039151e-05,
      "loss": 0.5209,
      "num_input_tokens_seen": 8421888,
      "step": 12880
    },
    {
      "epoch": 6.75314465408805,
      "grad_norm": 0.5613752007484436,
      "learning_rate": 4.188132318497051e-05,
      "loss": 0.5687,
      "num_input_tokens_seen": 8424768,
      "step": 12885
    },
    {
      "epoch": 6.755765199161425,
      "grad_norm": 0.6260790228843689,
      "learning_rate": 4.1872887648169304e-05,
      "loss": 0.3405,
      "num_input_tokens_seen": 8428576,
      "step": 12890
    },
    {
      "epoch": 6.7583857442348005,
      "grad_norm": 0.35087326169013977,
      "learning_rate": 4.186444858175248e-05,
      "loss": 0.4871,
      "num_input_tokens_seen": 8431744,
      "step": 12895
    },
    {
      "epoch": 6.761006289308176,
      "grad_norm": 0.8429364562034607,
      "learning_rate": 4.1856005987485424e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 8435712,
      "step": 12900
    },
    {
      "epoch": 6.763626834381551,
      "grad_norm": 0.5411970615386963,
      "learning_rate": 4.1847559867134215e-05,
      "loss": 0.3972,
      "num_input_tokens_seen": 8439072,
      "step": 12905
    },
    {
      "epoch": 6.766247379454927,
      "grad_norm": 0.4610627293586731,
      "learning_rate": 4.1839110222465693e-05,
      "loss": 0.3448,
      "num_input_tokens_seen": 8443392,
      "step": 12910
    },
    {
      "epoch": 6.768867924528302,
      "grad_norm": 0.2572508156299591,
      "learning_rate": 4.183065705524742e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 8446592,
      "step": 12915
    },
    {
      "epoch": 6.771488469601677,
      "grad_norm": 0.4965554475784302,
      "learning_rate": 4.1822200367247696e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 8449632,
      "step": 12920
    },
    {
      "epoch": 6.774109014675052,
      "grad_norm": 0.34385889768600464,
      "learning_rate": 4.181374016023559e-05,
      "loss": 0.4853,
      "num_input_tokens_seen": 8452608,
      "step": 12925
    },
    {
      "epoch": 6.776729559748428,
      "grad_norm": 0.4675944149494171,
      "learning_rate": 4.1805276435980854e-05,
      "loss": 0.4307,
      "num_input_tokens_seen": 8454880,
      "step": 12930
    },
    {
      "epoch": 6.779350104821803,
      "grad_norm": 0.490119069814682,
      "learning_rate": 4.1796809196254014e-05,
      "loss": 0.4932,
      "num_input_tokens_seen": 8457696,
      "step": 12935
    },
    {
      "epoch": 6.781970649895178,
      "grad_norm": 0.2684031128883362,
      "learning_rate": 4.178833844282632e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 8461088,
      "step": 12940
    },
    {
      "epoch": 6.784591194968553,
      "grad_norm": 0.8924992084503174,
      "learning_rate": 4.177986417746976e-05,
      "loss": 0.4601,
      "num_input_tokens_seen": 8464416,
      "step": 12945
    },
    {
      "epoch": 6.787211740041929,
      "grad_norm": 0.2642418444156647,
      "learning_rate": 4.177138640195704e-05,
      "loss": 0.5398,
      "num_input_tokens_seen": 8468512,
      "step": 12950
    },
    {
      "epoch": 6.789832285115304,
      "grad_norm": 0.27279579639434814,
      "learning_rate": 4.176290511806162e-05,
      "loss": 0.3987,
      "num_input_tokens_seen": 8472256,
      "step": 12955
    },
    {
      "epoch": 6.7924528301886795,
      "grad_norm": 0.33590173721313477,
      "learning_rate": 4.1754420327557706e-05,
      "loss": 0.4099,
      "num_input_tokens_seen": 8476416,
      "step": 12960
    },
    {
      "epoch": 6.795073375262055,
      "grad_norm": 0.421237975358963,
      "learning_rate": 4.1745932032220195e-05,
      "loss": 0.3801,
      "num_input_tokens_seen": 8481184,
      "step": 12965
    },
    {
      "epoch": 6.79769392033543,
      "grad_norm": 0.5090124607086182,
      "learning_rate": 4.173744023382474e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 8485472,
      "step": 12970
    },
    {
      "epoch": 6.800314465408805,
      "grad_norm": 0.5016621351242065,
      "learning_rate": 4.172894493414776e-05,
      "loss": 0.5575,
      "num_input_tokens_seen": 8488544,
      "step": 12975
    },
    {
      "epoch": 6.80293501048218,
      "grad_norm": 0.9129032492637634,
      "learning_rate": 4.172044613496634e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 8491712,
      "step": 12980
    },
    {
      "epoch": 6.805555555555555,
      "grad_norm": 0.3808901607990265,
      "learning_rate": 4.171194383805835e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 8495008,
      "step": 12985
    },
    {
      "epoch": 6.8081761006289305,
      "grad_norm": 0.28898516297340393,
      "learning_rate": 4.170343804520237e-05,
      "loss": 0.3706,
      "num_input_tokens_seen": 8498464,
      "step": 12990
    },
    {
      "epoch": 6.810796645702306,
      "grad_norm": 0.4356305003166199,
      "learning_rate": 4.169492875817771e-05,
      "loss": 0.4356,
      "num_input_tokens_seen": 8501696,
      "step": 12995
    },
    {
      "epoch": 6.813417190775681,
      "grad_norm": 0.6014792919158936,
      "learning_rate": 4.1686415978764415e-05,
      "loss": 0.5182,
      "num_input_tokens_seen": 8504288,
      "step": 13000
    },
    {
      "epoch": 6.816037735849057,
      "grad_norm": 0.2678202688694,
      "learning_rate": 4.167789970874327e-05,
      "loss": 0.4603,
      "num_input_tokens_seen": 8507840,
      "step": 13005
    },
    {
      "epoch": 6.818658280922432,
      "grad_norm": 0.5419698357582092,
      "learning_rate": 4.166937994989578e-05,
      "loss": 0.4751,
      "num_input_tokens_seen": 8510944,
      "step": 13010
    },
    {
      "epoch": 6.821278825995807,
      "grad_norm": 0.4368855357170105,
      "learning_rate": 4.166085670400418e-05,
      "loss": 0.5448,
      "num_input_tokens_seen": 8514944,
      "step": 13015
    },
    {
      "epoch": 6.823899371069182,
      "grad_norm": 0.3848039209842682,
      "learning_rate": 4.165232997285143e-05,
      "loss": 0.5973,
      "num_input_tokens_seen": 8517920,
      "step": 13020
    },
    {
      "epoch": 6.826519916142558,
      "grad_norm": 0.3299594521522522,
      "learning_rate": 4.164379975822124e-05,
      "loss": 0.5043,
      "num_input_tokens_seen": 8520448,
      "step": 13025
    },
    {
      "epoch": 6.829140461215933,
      "grad_norm": 0.36507442593574524,
      "learning_rate": 4.1635266061898015e-05,
      "loss": 0.4771,
      "num_input_tokens_seen": 8524000,
      "step": 13030
    },
    {
      "epoch": 6.831761006289308,
      "grad_norm": 0.40572819113731384,
      "learning_rate": 4.162672888566691e-05,
      "loss": 0.4199,
      "num_input_tokens_seen": 8526624,
      "step": 13035
    },
    {
      "epoch": 6.834381551362683,
      "grad_norm": 0.3040654957294464,
      "learning_rate": 4.1618188231313816e-05,
      "loss": 0.4705,
      "num_input_tokens_seen": 8529920,
      "step": 13040
    },
    {
      "epoch": 6.837002096436059,
      "grad_norm": 0.4758315086364746,
      "learning_rate": 4.1609644100625335e-05,
      "loss": 0.4545,
      "num_input_tokens_seen": 8532672,
      "step": 13045
    },
    {
      "epoch": 6.839622641509434,
      "grad_norm": 0.49306756258010864,
      "learning_rate": 4.16010964953888e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 8535520,
      "step": 13050
    },
    {
      "epoch": 6.8422431865828095,
      "grad_norm": 0.6254735589027405,
      "learning_rate": 4.159254541739227e-05,
      "loss": 0.5239,
      "num_input_tokens_seen": 8537920,
      "step": 13055
    },
    {
      "epoch": 6.844863731656185,
      "grad_norm": 0.3862360715866089,
      "learning_rate": 4.158399086842454e-05,
      "loss": 0.4605,
      "num_input_tokens_seen": 8541984,
      "step": 13060
    },
    {
      "epoch": 6.84748427672956,
      "grad_norm": 0.4251181185245514,
      "learning_rate": 4.157543285027512e-05,
      "loss": 0.4727,
      "num_input_tokens_seen": 8546272,
      "step": 13065
    },
    {
      "epoch": 6.850104821802935,
      "grad_norm": 1.2567882537841797,
      "learning_rate": 4.156687136473425e-05,
      "loss": 0.6023,
      "num_input_tokens_seen": 8549280,
      "step": 13070
    },
    {
      "epoch": 6.85272536687631,
      "grad_norm": 0.4443606436252594,
      "learning_rate": 4.1558306413592894e-05,
      "loss": 0.3808,
      "num_input_tokens_seen": 8552064,
      "step": 13075
    },
    {
      "epoch": 6.855345911949685,
      "grad_norm": 0.5890039205551147,
      "learning_rate": 4.1549737998642745e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 8555712,
      "step": 13080
    },
    {
      "epoch": 6.8579664570230605,
      "grad_norm": 0.33610621094703674,
      "learning_rate": 4.154116612167622e-05,
      "loss": 0.5098,
      "num_input_tokens_seen": 8559328,
      "step": 13085
    },
    {
      "epoch": 6.860587002096436,
      "grad_norm": 0.40969422459602356,
      "learning_rate": 4.1532590784486444e-05,
      "loss": 0.5762,
      "num_input_tokens_seen": 8561856,
      "step": 13090
    },
    {
      "epoch": 6.863207547169811,
      "grad_norm": 0.5193758606910706,
      "learning_rate": 4.1524011988867294e-05,
      "loss": 0.5724,
      "num_input_tokens_seen": 8564960,
      "step": 13095
    },
    {
      "epoch": 6.865828092243187,
      "grad_norm": 0.32004639506340027,
      "learning_rate": 4.151542973661336e-05,
      "loss": 0.5311,
      "num_input_tokens_seen": 8568640,
      "step": 13100
    },
    {
      "epoch": 6.868448637316562,
      "grad_norm": 0.47200512886047363,
      "learning_rate": 4.150684402951994e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 8572064,
      "step": 13105
    },
    {
      "epoch": 6.871069182389937,
      "grad_norm": 0.4141511917114258,
      "learning_rate": 4.149825486938306e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 8575104,
      "step": 13110
    },
    {
      "epoch": 6.873689727463312,
      "grad_norm": 0.6127253174781799,
      "learning_rate": 4.148966225799949e-05,
      "loss": 0.5673,
      "num_input_tokens_seen": 8577888,
      "step": 13115
    },
    {
      "epoch": 6.876310272536688,
      "grad_norm": 0.33945250511169434,
      "learning_rate": 4.14810661971667e-05,
      "loss": 0.5436,
      "num_input_tokens_seen": 8580896,
      "step": 13120
    },
    {
      "epoch": 6.878930817610063,
      "grad_norm": 0.49277815222740173,
      "learning_rate": 4.147246668868289e-05,
      "loss": 0.5088,
      "num_input_tokens_seen": 8584864,
      "step": 13125
    },
    {
      "epoch": 6.881551362683438,
      "grad_norm": 0.3690752685070038,
      "learning_rate": 4.1463863734346964e-05,
      "loss": 0.7094,
      "num_input_tokens_seen": 8587936,
      "step": 13130
    },
    {
      "epoch": 6.884171907756813,
      "grad_norm": 0.3935890793800354,
      "learning_rate": 4.145525733595859e-05,
      "loss": 0.4157,
      "num_input_tokens_seen": 8592064,
      "step": 13135
    },
    {
      "epoch": 6.886792452830189,
      "grad_norm": 1.285722255706787,
      "learning_rate": 4.14466474953181e-05,
      "loss": 0.5132,
      "num_input_tokens_seen": 8595136,
      "step": 13140
    },
    {
      "epoch": 6.889412997903564,
      "grad_norm": 0.3107511103153229,
      "learning_rate": 4.14380342142266e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 8597472,
      "step": 13145
    },
    {
      "epoch": 6.8920335429769395,
      "grad_norm": 0.9542084336280823,
      "learning_rate": 4.142941749448587e-05,
      "loss": 0.5274,
      "num_input_tokens_seen": 8600000,
      "step": 13150
    },
    {
      "epoch": 6.894654088050315,
      "grad_norm": 0.21988241374492645,
      "learning_rate": 4.142079733789843e-05,
      "loss": 0.3674,
      "num_input_tokens_seen": 8604960,
      "step": 13155
    },
    {
      "epoch": 6.89727463312369,
      "grad_norm": 0.2696070671081543,
      "learning_rate": 4.141217374626753e-05,
      "loss": 0.4581,
      "num_input_tokens_seen": 8607776,
      "step": 13160
    },
    {
      "epoch": 6.899895178197065,
      "grad_norm": 0.5154539346694946,
      "learning_rate": 4.140354672139713e-05,
      "loss": 0.414,
      "num_input_tokens_seen": 8610944,
      "step": 13165
    },
    {
      "epoch": 6.90251572327044,
      "grad_norm": 0.5695741176605225,
      "learning_rate": 4.1394916265091886e-05,
      "loss": 0.538,
      "num_input_tokens_seen": 8614368,
      "step": 13170
    },
    {
      "epoch": 6.905136268343815,
      "grad_norm": 0.4227425158023834,
      "learning_rate": 4.1386282379157214e-05,
      "loss": 0.4419,
      "num_input_tokens_seen": 8618272,
      "step": 13175
    },
    {
      "epoch": 6.9077568134171905,
      "grad_norm": 0.5793626308441162,
      "learning_rate": 4.137764506539921e-05,
      "loss": 0.3116,
      "num_input_tokens_seen": 8620640,
      "step": 13180
    },
    {
      "epoch": 6.910377358490566,
      "grad_norm": 0.4016636312007904,
      "learning_rate": 4.136900432562471e-05,
      "loss": 0.397,
      "num_input_tokens_seen": 8624192,
      "step": 13185
    },
    {
      "epoch": 6.912997903563941,
      "grad_norm": 0.6693088412284851,
      "learning_rate": 4.136036016164123e-05,
      "loss": 0.357,
      "num_input_tokens_seen": 8627072,
      "step": 13190
    },
    {
      "epoch": 6.915618448637317,
      "grad_norm": 0.5196492671966553,
      "learning_rate": 4.135171257525707e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 8631232,
      "step": 13195
    },
    {
      "epoch": 6.918238993710692,
      "grad_norm": 0.4262711703777313,
      "learning_rate": 4.1343061568281195e-05,
      "loss": 0.5934,
      "num_input_tokens_seen": 8633696,
      "step": 13200
    },
    {
      "epoch": 6.920859538784067,
      "grad_norm": 0.5294492840766907,
      "learning_rate": 4.133440714252328e-05,
      "loss": 0.3438,
      "num_input_tokens_seen": 8636800,
      "step": 13205
    },
    {
      "epoch": 6.923480083857442,
      "grad_norm": 0.38058239221572876,
      "learning_rate": 4.132574929979375e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 8639968,
      "step": 13210
    },
    {
      "epoch": 6.926100628930818,
      "grad_norm": 0.2675808072090149,
      "learning_rate": 4.1317088041903714e-05,
      "loss": 0.4687,
      "num_input_tokens_seen": 8643328,
      "step": 13215
    },
    {
      "epoch": 6.928721174004193,
      "grad_norm": 0.3908812403678894,
      "learning_rate": 4.1308423370665015e-05,
      "loss": 0.4075,
      "num_input_tokens_seen": 8646144,
      "step": 13220
    },
    {
      "epoch": 6.931341719077568,
      "grad_norm": 0.3616536557674408,
      "learning_rate": 4.1299755287890206e-05,
      "loss": 0.4941,
      "num_input_tokens_seen": 8649536,
      "step": 13225
    },
    {
      "epoch": 6.933962264150943,
      "grad_norm": 0.3255927562713623,
      "learning_rate": 4.129108379539255e-05,
      "loss": 0.3388,
      "num_input_tokens_seen": 8652992,
      "step": 13230
    },
    {
      "epoch": 6.936582809224319,
      "grad_norm": 0.6499491333961487,
      "learning_rate": 4.128240889498602e-05,
      "loss": 0.4138,
      "num_input_tokens_seen": 8656608,
      "step": 13235
    },
    {
      "epoch": 6.939203354297694,
      "grad_norm": 0.41107627749443054,
      "learning_rate": 4.127373058848531e-05,
      "loss": 0.64,
      "num_input_tokens_seen": 8659744,
      "step": 13240
    },
    {
      "epoch": 6.9418238993710695,
      "grad_norm": 0.32004880905151367,
      "learning_rate": 4.126504887770582e-05,
      "loss": 0.5052,
      "num_input_tokens_seen": 8662176,
      "step": 13245
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 0.3783542811870575,
      "learning_rate": 4.1256363764463666e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 8665184,
      "step": 13250
    },
    {
      "epoch": 6.94706498951782,
      "grad_norm": 1.1477725505828857,
      "learning_rate": 4.1247675250575676e-05,
      "loss": 0.4452,
      "num_input_tokens_seen": 8668096,
      "step": 13255
    },
    {
      "epoch": 6.949685534591195,
      "grad_norm": 0.3218315541744232,
      "learning_rate": 4.1238983337859396e-05,
      "loss": 0.4207,
      "num_input_tokens_seen": 8671392,
      "step": 13260
    },
    {
      "epoch": 6.95230607966457,
      "grad_norm": 0.3972824513912201,
      "learning_rate": 4.123028802813305e-05,
      "loss": 0.6558,
      "num_input_tokens_seen": 8675360,
      "step": 13265
    },
    {
      "epoch": 6.954926624737945,
      "grad_norm": 0.4460081160068512,
      "learning_rate": 4.1221589323215625e-05,
      "loss": 0.3511,
      "num_input_tokens_seen": 8678880,
      "step": 13270
    },
    {
      "epoch": 6.9575471698113205,
      "grad_norm": 0.8376274704933167,
      "learning_rate": 4.121288722492678e-05,
      "loss": 0.3533,
      "num_input_tokens_seen": 8682176,
      "step": 13275
    },
    {
      "epoch": 6.960167714884696,
      "grad_norm": 0.6890677809715271,
      "learning_rate": 4.120418173508688e-05,
      "loss": 0.3759,
      "num_input_tokens_seen": 8685024,
      "step": 13280
    },
    {
      "epoch": 6.962788259958071,
      "grad_norm": 0.7143837809562683,
      "learning_rate": 4.1195472855517045e-05,
      "loss": 0.6204,
      "num_input_tokens_seen": 8688768,
      "step": 13285
    },
    {
      "epoch": 6.965408805031447,
      "grad_norm": 0.505692183971405,
      "learning_rate": 4.1186760588039045e-05,
      "loss": 0.4013,
      "num_input_tokens_seen": 8692032,
      "step": 13290
    },
    {
      "epoch": 6.968029350104822,
      "grad_norm": 0.3782356381416321,
      "learning_rate": 4.11780449344754e-05,
      "loss": 0.5036,
      "num_input_tokens_seen": 8695744,
      "step": 13295
    },
    {
      "epoch": 6.970649895178197,
      "grad_norm": 0.34468868374824524,
      "learning_rate": 4.1169325896649326e-05,
      "loss": 0.5039,
      "num_input_tokens_seen": 8699680,
      "step": 13300
    },
    {
      "epoch": 6.973270440251572,
      "grad_norm": 0.3505600392818451,
      "learning_rate": 4.116060347638474e-05,
      "loss": 0.38,
      "num_input_tokens_seen": 8702816,
      "step": 13305
    },
    {
      "epoch": 6.975890985324948,
      "grad_norm": 0.3654826879501343,
      "learning_rate": 4.115187767550628e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 8705600,
      "step": 13310
    },
    {
      "epoch": 6.978511530398323,
      "grad_norm": 0.42897719144821167,
      "learning_rate": 4.114314849583928e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 8708064,
      "step": 13315
    },
    {
      "epoch": 6.981132075471698,
      "grad_norm": 0.5387104153633118,
      "learning_rate": 4.113441593920979e-05,
      "loss": 0.38,
      "num_input_tokens_seen": 8710944,
      "step": 13320
    },
    {
      "epoch": 6.983752620545073,
      "grad_norm": 0.487385630607605,
      "learning_rate": 4.112568000744454e-05,
      "loss": 0.5703,
      "num_input_tokens_seen": 8714368,
      "step": 13325
    },
    {
      "epoch": 6.986373165618449,
      "grad_norm": 0.3707428276538849,
      "learning_rate": 4.1116940702371025e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 8717824,
      "step": 13330
    },
    {
      "epoch": 6.988993710691824,
      "grad_norm": 0.4461338520050049,
      "learning_rate": 4.1108198025817376e-05,
      "loss": 0.4483,
      "num_input_tokens_seen": 8720928,
      "step": 13335
    },
    {
      "epoch": 6.9916142557651995,
      "grad_norm": 0.8770213723182678,
      "learning_rate": 4.1099451979612476e-05,
      "loss": 0.494,
      "num_input_tokens_seen": 8724704,
      "step": 13340
    },
    {
      "epoch": 6.994234800838575,
      "grad_norm": 0.4893170893192291,
      "learning_rate": 4.109070256558589e-05,
      "loss": 0.4165,
      "num_input_tokens_seen": 8727232,
      "step": 13345
    },
    {
      "epoch": 6.99685534591195,
      "grad_norm": 0.5730828046798706,
      "learning_rate": 4.108194978556791e-05,
      "loss": 0.4815,
      "num_input_tokens_seen": 8729920,
      "step": 13350
    },
    {
      "epoch": 6.999475890985325,
      "grad_norm": 0.5157949924468994,
      "learning_rate": 4.1073193641389495e-05,
      "loss": 0.6624,
      "num_input_tokens_seen": 8732928,
      "step": 13355
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.45220306515693665,
      "eval_runtime": 9.6549,
      "eval_samples_per_second": 87.831,
      "eval_steps_per_second": 21.958,
      "num_input_tokens_seen": 8733024,
      "step": 13356
    },
    {
      "epoch": 7.0020964360587,
      "grad_norm": 0.3366113305091858,
      "learning_rate": 4.1064434134882345e-05,
      "loss": 0.538,
      "num_input_tokens_seen": 8736160,
      "step": 13360
    },
    {
      "epoch": 7.004716981132075,
      "grad_norm": 0.3633488714694977,
      "learning_rate": 4.1055671267878847e-05,
      "loss": 0.464,
      "num_input_tokens_seen": 8738656,
      "step": 13365
    },
    {
      "epoch": 7.0073375262054505,
      "grad_norm": 0.40504297614097595,
      "learning_rate": 4.1046905042212106e-05,
      "loss": 0.6642,
      "num_input_tokens_seen": 8742240,
      "step": 13370
    },
    {
      "epoch": 7.009958071278826,
      "grad_norm": 0.6244048476219177,
      "learning_rate": 4.103813545971589e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 8745312,
      "step": 13375
    },
    {
      "epoch": 7.012578616352202,
      "grad_norm": 0.5897805094718933,
      "learning_rate": 4.10293625222247e-05,
      "loss": 0.4098,
      "num_input_tokens_seen": 8748032,
      "step": 13380
    },
    {
      "epoch": 7.015199161425577,
      "grad_norm": 0.6458056569099426,
      "learning_rate": 4.102058623157376e-05,
      "loss": 0.5478,
      "num_input_tokens_seen": 8751584,
      "step": 13385
    },
    {
      "epoch": 7.017819706498952,
      "grad_norm": 0.24352259933948517,
      "learning_rate": 4.101180658959894e-05,
      "loss": 0.3577,
      "num_input_tokens_seen": 8754432,
      "step": 13390
    },
    {
      "epoch": 7.020440251572327,
      "grad_norm": 0.482522577047348,
      "learning_rate": 4.1003023598136866e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 8757344,
      "step": 13395
    },
    {
      "epoch": 7.023060796645702,
      "grad_norm": 0.06027965620160103,
      "learning_rate": 4.099423725902482e-05,
      "loss": 0.3829,
      "num_input_tokens_seen": 8764288,
      "step": 13400
    },
    {
      "epoch": 7.0256813417190775,
      "grad_norm": 0.24504920840263367,
      "learning_rate": 4.098544757410081e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 8768160,
      "step": 13405
    },
    {
      "epoch": 7.028301886792453,
      "grad_norm": 0.4250894784927368,
      "learning_rate": 4.097665454520354e-05,
      "loss": 0.3859,
      "num_input_tokens_seen": 8771712,
      "step": 13410
    },
    {
      "epoch": 7.030922431865828,
      "grad_norm": 0.3907134234905243,
      "learning_rate": 4.096785817417241e-05,
      "loss": 0.4609,
      "num_input_tokens_seen": 8775136,
      "step": 13415
    },
    {
      "epoch": 7.033542976939203,
      "grad_norm": 1.2691316604614258,
      "learning_rate": 4.0959058462847524e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 8778304,
      "step": 13420
    },
    {
      "epoch": 7.036163522012578,
      "grad_norm": 0.5564826130867004,
      "learning_rate": 4.095025541306968e-05,
      "loss": 0.4693,
      "num_input_tokens_seen": 8781024,
      "step": 13425
    },
    {
      "epoch": 7.038784067085954,
      "grad_norm": 0.5655983090400696,
      "learning_rate": 4.0941449026680374e-05,
      "loss": 0.597,
      "num_input_tokens_seen": 8784000,
      "step": 13430
    },
    {
      "epoch": 7.0414046121593294,
      "grad_norm": 0.4186331629753113,
      "learning_rate": 4.09326393055218e-05,
      "loss": 0.5897,
      "num_input_tokens_seen": 8787424,
      "step": 13435
    },
    {
      "epoch": 7.044025157232705,
      "grad_norm": 0.3279379606246948,
      "learning_rate": 4.0923826251436845e-05,
      "loss": 0.3714,
      "num_input_tokens_seen": 8790816,
      "step": 13440
    },
    {
      "epoch": 7.04664570230608,
      "grad_norm": 0.944725751876831,
      "learning_rate": 4.0915009866269115e-05,
      "loss": 0.3698,
      "num_input_tokens_seen": 8792992,
      "step": 13445
    },
    {
      "epoch": 7.049266247379455,
      "grad_norm": 0.422562837600708,
      "learning_rate": 4.090619015186288e-05,
      "loss": 0.3258,
      "num_input_tokens_seen": 8795840,
      "step": 13450
    },
    {
      "epoch": 7.05188679245283,
      "grad_norm": 0.4998694956302643,
      "learning_rate": 4.0897367110063145e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 8798560,
      "step": 13455
    },
    {
      "epoch": 7.054507337526205,
      "grad_norm": 0.6475602984428406,
      "learning_rate": 4.088854074271557e-05,
      "loss": 0.6433,
      "num_input_tokens_seen": 8802240,
      "step": 13460
    },
    {
      "epoch": 7.0571278825995805,
      "grad_norm": 0.4071342945098877,
      "learning_rate": 4.0879711051666534e-05,
      "loss": 0.3665,
      "num_input_tokens_seen": 8804960,
      "step": 13465
    },
    {
      "epoch": 7.059748427672956,
      "grad_norm": 0.4534774720668793,
      "learning_rate": 4.087087803876313e-05,
      "loss": 0.4458,
      "num_input_tokens_seen": 8808000,
      "step": 13470
    },
    {
      "epoch": 7.062368972746331,
      "grad_norm": 0.550571084022522,
      "learning_rate": 4.086204170585308e-05,
      "loss": 0.4001,
      "num_input_tokens_seen": 8811008,
      "step": 13475
    },
    {
      "epoch": 7.064989517819707,
      "grad_norm": 0.3839147090911865,
      "learning_rate": 4.085320205478488e-05,
      "loss": 0.4453,
      "num_input_tokens_seen": 8813728,
      "step": 13480
    },
    {
      "epoch": 7.067610062893082,
      "grad_norm": 0.43673017621040344,
      "learning_rate": 4.084435908740768e-05,
      "loss": 0.5175,
      "num_input_tokens_seen": 8817824,
      "step": 13485
    },
    {
      "epoch": 7.070230607966457,
      "grad_norm": 1.1331965923309326,
      "learning_rate": 4.0835512805571306e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 8822848,
      "step": 13490
    },
    {
      "epoch": 7.072851153039832,
      "grad_norm": 0.46993252635002136,
      "learning_rate": 4.082666321112633e-05,
      "loss": 0.4688,
      "num_input_tokens_seen": 8826016,
      "step": 13495
    },
    {
      "epoch": 7.0754716981132075,
      "grad_norm": 0.5722844004631042,
      "learning_rate": 4.081781030592396e-05,
      "loss": 0.4132,
      "num_input_tokens_seen": 8829088,
      "step": 13500
    },
    {
      "epoch": 7.078092243186583,
      "grad_norm": 0.6301876902580261,
      "learning_rate": 4.080895409181613e-05,
      "loss": 0.4651,
      "num_input_tokens_seen": 8832448,
      "step": 13505
    },
    {
      "epoch": 7.080712788259958,
      "grad_norm": 0.5482256412506104,
      "learning_rate": 4.080009457065547e-05,
      "loss": 0.437,
      "num_input_tokens_seen": 8835232,
      "step": 13510
    },
    {
      "epoch": 7.083333333333333,
      "grad_norm": 0.5460415482521057,
      "learning_rate": 4.0791231744295286e-05,
      "loss": 0.3439,
      "num_input_tokens_seen": 8838176,
      "step": 13515
    },
    {
      "epoch": 7.085953878406708,
      "grad_norm": 0.28297731280326843,
      "learning_rate": 4.078236561458957e-05,
      "loss": 0.3798,
      "num_input_tokens_seen": 8841120,
      "step": 13520
    },
    {
      "epoch": 7.088574423480084,
      "grad_norm": 0.9249153733253479,
      "learning_rate": 4.077349618339302e-05,
      "loss": 0.5354,
      "num_input_tokens_seen": 8843744,
      "step": 13525
    },
    {
      "epoch": 7.091194968553459,
      "grad_norm": 0.4121631681919098,
      "learning_rate": 4.076462345256103e-05,
      "loss": 0.4704,
      "num_input_tokens_seen": 8846944,
      "step": 13530
    },
    {
      "epoch": 7.093815513626835,
      "grad_norm": 2.280163526535034,
      "learning_rate": 4.0755747423949655e-05,
      "loss": 0.4449,
      "num_input_tokens_seen": 8849888,
      "step": 13535
    },
    {
      "epoch": 7.09643605870021,
      "grad_norm": 0.669028639793396,
      "learning_rate": 4.074686809941568e-05,
      "loss": 0.4922,
      "num_input_tokens_seen": 8852608,
      "step": 13540
    },
    {
      "epoch": 7.099056603773585,
      "grad_norm": 0.5059594511985779,
      "learning_rate": 4.0737985480816546e-05,
      "loss": 0.4394,
      "num_input_tokens_seen": 8855072,
      "step": 13545
    },
    {
      "epoch": 7.10167714884696,
      "grad_norm": 0.28051459789276123,
      "learning_rate": 4.0729099570010395e-05,
      "loss": 0.3848,
      "num_input_tokens_seen": 8858656,
      "step": 13550
    },
    {
      "epoch": 7.104297693920335,
      "grad_norm": 0.5288695693016052,
      "learning_rate": 4.072021036885607e-05,
      "loss": 0.5678,
      "num_input_tokens_seen": 8861856,
      "step": 13555
    },
    {
      "epoch": 7.1069182389937104,
      "grad_norm": 0.566032350063324,
      "learning_rate": 4.0711317879213075e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 8864704,
      "step": 13560
    },
    {
      "epoch": 7.109538784067086,
      "grad_norm": 0.7312678694725037,
      "learning_rate": 4.070242210294162e-05,
      "loss": 0.4658,
      "num_input_tokens_seen": 8867776,
      "step": 13565
    },
    {
      "epoch": 7.112159329140461,
      "grad_norm": 0.4638868272304535,
      "learning_rate": 4.069352304190261e-05,
      "loss": 0.5152,
      "num_input_tokens_seen": 8870560,
      "step": 13570
    },
    {
      "epoch": 7.114779874213837,
      "grad_norm": 0.31681519746780396,
      "learning_rate": 4.068462069795761e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 8874112,
      "step": 13575
    },
    {
      "epoch": 7.117400419287212,
      "grad_norm": 0.4455295503139496,
      "learning_rate": 4.067571507296891e-05,
      "loss": 0.4855,
      "num_input_tokens_seen": 8877184,
      "step": 13580
    },
    {
      "epoch": 7.120020964360587,
      "grad_norm": 0.4041725993156433,
      "learning_rate": 4.0666806168799455e-05,
      "loss": 0.5379,
      "num_input_tokens_seen": 8880512,
      "step": 13585
    },
    {
      "epoch": 7.122641509433962,
      "grad_norm": 0.7215731143951416,
      "learning_rate": 4.0657893987312876e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 8883232,
      "step": 13590
    },
    {
      "epoch": 7.1252620545073375,
      "grad_norm": 0.47289228439331055,
      "learning_rate": 4.064897853037352e-05,
      "loss": 0.4931,
      "num_input_tokens_seen": 8887008,
      "step": 13595
    },
    {
      "epoch": 7.127882599580713,
      "grad_norm": 0.7070528864860535,
      "learning_rate": 4.064005979984639e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 8889952,
      "step": 13600
    },
    {
      "epoch": 7.130503144654088,
      "grad_norm": 0.34339794516563416,
      "learning_rate": 4.0631137797597176e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 8893408,
      "step": 13605
    },
    {
      "epoch": 7.133123689727463,
      "grad_norm": 0.5466752052307129,
      "learning_rate": 4.0622212525492266e-05,
      "loss": 0.4327,
      "num_input_tokens_seen": 8897248,
      "step": 13610
    },
    {
      "epoch": 7.135744234800838,
      "grad_norm": 0.41535845398902893,
      "learning_rate": 4.061328398539872e-05,
      "loss": 0.5062,
      "num_input_tokens_seen": 8901152,
      "step": 13615
    },
    {
      "epoch": 7.138364779874214,
      "grad_norm": 0.47767236828804016,
      "learning_rate": 4.060435217918429e-05,
      "loss": 0.5481,
      "num_input_tokens_seen": 8904800,
      "step": 13620
    },
    {
      "epoch": 7.140985324947589,
      "grad_norm": 0.39358100295066833,
      "learning_rate": 4.0595417108717405e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 8907840,
      "step": 13625
    },
    {
      "epoch": 7.143605870020965,
      "grad_norm": 0.4032335579395294,
      "learning_rate": 4.0586478775867185e-05,
      "loss": 0.3909,
      "num_input_tokens_seen": 8910976,
      "step": 13630
    },
    {
      "epoch": 7.14622641509434,
      "grad_norm": 0.3937318027019501,
      "learning_rate": 4.057753718250343e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 8913728,
      "step": 13635
    },
    {
      "epoch": 7.148846960167715,
      "grad_norm": 0.2969170808792114,
      "learning_rate": 4.056859233049661e-05,
      "loss": 0.3631,
      "num_input_tokens_seen": 8916928,
      "step": 13640
    },
    {
      "epoch": 7.15146750524109,
      "grad_norm": 0.825558066368103,
      "learning_rate": 4.05596442217179e-05,
      "loss": 0.5023,
      "num_input_tokens_seen": 8920128,
      "step": 13645
    },
    {
      "epoch": 7.154088050314465,
      "grad_norm": 0.4356265068054199,
      "learning_rate": 4.055069285803912e-05,
      "loss": 0.4356,
      "num_input_tokens_seen": 8923424,
      "step": 13650
    },
    {
      "epoch": 7.15670859538784,
      "grad_norm": 0.497408390045166,
      "learning_rate": 4.0541738241332815e-05,
      "loss": 0.3524,
      "num_input_tokens_seen": 8925856,
      "step": 13655
    },
    {
      "epoch": 7.159329140461216,
      "grad_norm": 1.2174451351165771,
      "learning_rate": 4.053278037347217e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 8928544,
      "step": 13660
    },
    {
      "epoch": 7.161949685534591,
      "grad_norm": 0.5036420226097107,
      "learning_rate": 4.052381925633109e-05,
      "loss": 0.4927,
      "num_input_tokens_seen": 8931136,
      "step": 13665
    },
    {
      "epoch": 7.164570230607967,
      "grad_norm": 0.3643576502799988,
      "learning_rate": 4.051485489178412e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 8934336,
      "step": 13670
    },
    {
      "epoch": 7.167190775681342,
      "grad_norm": 0.5427345037460327,
      "learning_rate": 4.05058872817065e-05,
      "loss": 0.5536,
      "num_input_tokens_seen": 8936896,
      "step": 13675
    },
    {
      "epoch": 7.169811320754717,
      "grad_norm": 0.46372562646865845,
      "learning_rate": 4.0496916427974177e-05,
      "loss": 0.3987,
      "num_input_tokens_seen": 8940224,
      "step": 13680
    },
    {
      "epoch": 7.172431865828092,
      "grad_norm": 0.5533252954483032,
      "learning_rate": 4.048794233246373e-05,
      "loss": 0.4702,
      "num_input_tokens_seen": 8943840,
      "step": 13685
    },
    {
      "epoch": 7.1750524109014675,
      "grad_norm": 0.8879913687705994,
      "learning_rate": 4.0478964997052435e-05,
      "loss": 0.4588,
      "num_input_tokens_seen": 8946624,
      "step": 13690
    },
    {
      "epoch": 7.177672955974843,
      "grad_norm": 0.9638756513595581,
      "learning_rate": 4.046998442361827e-05,
      "loss": 0.6308,
      "num_input_tokens_seen": 8950272,
      "step": 13695
    },
    {
      "epoch": 7.180293501048218,
      "grad_norm": 0.3783470094203949,
      "learning_rate": 4.046100061403984e-05,
      "loss": 0.4123,
      "num_input_tokens_seen": 8952608,
      "step": 13700
    },
    {
      "epoch": 7.182914046121593,
      "grad_norm": 0.4304353594779968,
      "learning_rate": 4.045201357019648e-05,
      "loss": 0.5958,
      "num_input_tokens_seen": 8955744,
      "step": 13705
    },
    {
      "epoch": 7.185534591194968,
      "grad_norm": 0.4059954583644867,
      "learning_rate": 4.044302329396816e-05,
      "loss": 0.3078,
      "num_input_tokens_seen": 8958496,
      "step": 13710
    },
    {
      "epoch": 7.188155136268344,
      "grad_norm": 0.6073222756385803,
      "learning_rate": 4.0434029787235564e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 8961856,
      "step": 13715
    },
    {
      "epoch": 7.190775681341719,
      "grad_norm": 0.30564165115356445,
      "learning_rate": 4.042503305188e-05,
      "loss": 0.4336,
      "num_input_tokens_seen": 8964608,
      "step": 13720
    },
    {
      "epoch": 7.193396226415095,
      "grad_norm": 0.28049227595329285,
      "learning_rate": 4.041603308978351e-05,
      "loss": 0.4166,
      "num_input_tokens_seen": 8970016,
      "step": 13725
    },
    {
      "epoch": 7.19601677148847,
      "grad_norm": 0.5435615181922913,
      "learning_rate": 4.040702990282878e-05,
      "loss": 0.4208,
      "num_input_tokens_seen": 8973088,
      "step": 13730
    },
    {
      "epoch": 7.198637316561845,
      "grad_norm": 0.437811940908432,
      "learning_rate": 4.039802349289914e-05,
      "loss": 0.5137,
      "num_input_tokens_seen": 8976416,
      "step": 13735
    },
    {
      "epoch": 7.20125786163522,
      "grad_norm": 0.5100035667419434,
      "learning_rate": 4.038901386187868e-05,
      "loss": 0.4119,
      "num_input_tokens_seen": 8979424,
      "step": 13740
    },
    {
      "epoch": 7.203878406708595,
      "grad_norm": 0.7149427533149719,
      "learning_rate": 4.038000101165208e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 8982944,
      "step": 13745
    },
    {
      "epoch": 7.20649895178197,
      "grad_norm": 0.7207822203636169,
      "learning_rate": 4.0370984944104725e-05,
      "loss": 0.4474,
      "num_input_tokens_seen": 8986144,
      "step": 13750
    },
    {
      "epoch": 7.209119496855346,
      "grad_norm": 0.430806964635849,
      "learning_rate": 4.0361965661122695e-05,
      "loss": 0.3705,
      "num_input_tokens_seen": 8990048,
      "step": 13755
    },
    {
      "epoch": 7.211740041928721,
      "grad_norm": 0.4299003779888153,
      "learning_rate": 4.035294316459271e-05,
      "loss": 0.5575,
      "num_input_tokens_seen": 8992288,
      "step": 13760
    },
    {
      "epoch": 7.214360587002097,
      "grad_norm": 0.46649402379989624,
      "learning_rate": 4.034391745640215e-05,
      "loss": 0.4452,
      "num_input_tokens_seen": 8995200,
      "step": 13765
    },
    {
      "epoch": 7.216981132075472,
      "grad_norm": 0.634655773639679,
      "learning_rate": 4.033488853843913e-05,
      "loss": 0.5752,
      "num_input_tokens_seen": 8998112,
      "step": 13770
    },
    {
      "epoch": 7.219601677148847,
      "grad_norm": 0.3205602467060089,
      "learning_rate": 4.0325856412592363e-05,
      "loss": 0.5463,
      "num_input_tokens_seen": 9002560,
      "step": 13775
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 0.451229065656662,
      "learning_rate": 4.031682108075129e-05,
      "loss": 0.4652,
      "num_input_tokens_seen": 9005536,
      "step": 13780
    },
    {
      "epoch": 7.2248427672955975,
      "grad_norm": 0.5941562056541443,
      "learning_rate": 4.030778254480599e-05,
      "loss": 0.4776,
      "num_input_tokens_seen": 9008480,
      "step": 13785
    },
    {
      "epoch": 7.227463312368973,
      "grad_norm": 0.7157579660415649,
      "learning_rate": 4.029874080664722e-05,
      "loss": 0.5914,
      "num_input_tokens_seen": 9011328,
      "step": 13790
    },
    {
      "epoch": 7.230083857442348,
      "grad_norm": 0.268058717250824,
      "learning_rate": 4.0289695868166414e-05,
      "loss": 0.5036,
      "num_input_tokens_seen": 9014304,
      "step": 13795
    },
    {
      "epoch": 7.232704402515723,
      "grad_norm": 0.3763597309589386,
      "learning_rate": 4.028064773125566e-05,
      "loss": 0.6458,
      "num_input_tokens_seen": 9016928,
      "step": 13800
    },
    {
      "epoch": 7.235324947589098,
      "grad_norm": 0.44919317960739136,
      "learning_rate": 4.0271596397807734e-05,
      "loss": 0.4412,
      "num_input_tokens_seen": 9019456,
      "step": 13805
    },
    {
      "epoch": 7.237945492662474,
      "grad_norm": 0.47124379873275757,
      "learning_rate": 4.026254186971607e-05,
      "loss": 0.4854,
      "num_input_tokens_seen": 9022816,
      "step": 13810
    },
    {
      "epoch": 7.240566037735849,
      "grad_norm": 0.5896731019020081,
      "learning_rate": 4.025348414887477e-05,
      "loss": 0.4057,
      "num_input_tokens_seen": 9025728,
      "step": 13815
    },
    {
      "epoch": 7.243186582809225,
      "grad_norm": 0.4109937250614166,
      "learning_rate": 4.024442323717861e-05,
      "loss": 0.4674,
      "num_input_tokens_seen": 9028864,
      "step": 13820
    },
    {
      "epoch": 7.2458071278826,
      "grad_norm": 0.9197207689285278,
      "learning_rate": 4.023535913652302e-05,
      "loss": 0.5218,
      "num_input_tokens_seen": 9032160,
      "step": 13825
    },
    {
      "epoch": 7.248427672955975,
      "grad_norm": 0.35556161403656006,
      "learning_rate": 4.022629184880412e-05,
      "loss": 0.2742,
      "num_input_tokens_seen": 9035872,
      "step": 13830
    },
    {
      "epoch": 7.25104821802935,
      "grad_norm": 0.32236728072166443,
      "learning_rate": 4.021722137591867e-05,
      "loss": 0.3667,
      "num_input_tokens_seen": 9039456,
      "step": 13835
    },
    {
      "epoch": 7.253668763102725,
      "grad_norm": 0.49908220767974854,
      "learning_rate": 4.020814771976412e-05,
      "loss": 0.5006,
      "num_input_tokens_seen": 9042304,
      "step": 13840
    },
    {
      "epoch": 7.2562893081761,
      "grad_norm": 0.65738844871521,
      "learning_rate": 4.019907088223857e-05,
      "loss": 0.5691,
      "num_input_tokens_seen": 9045760,
      "step": 13845
    },
    {
      "epoch": 7.258909853249476,
      "grad_norm": 0.3007947504520416,
      "learning_rate": 4.018999086524079e-05,
      "loss": 0.4522,
      "num_input_tokens_seen": 9048704,
      "step": 13850
    },
    {
      "epoch": 7.261530398322851,
      "grad_norm": 0.6625242233276367,
      "learning_rate": 4.018090767067022e-05,
      "loss": 0.4667,
      "num_input_tokens_seen": 9052352,
      "step": 13855
    },
    {
      "epoch": 7.264150943396227,
      "grad_norm": 1.1613649129867554,
      "learning_rate": 4.017182130042696e-05,
      "loss": 0.3705,
      "num_input_tokens_seen": 9055136,
      "step": 13860
    },
    {
      "epoch": 7.266771488469602,
      "grad_norm": 0.3755546510219574,
      "learning_rate": 4.0162731756411764e-05,
      "loss": 0.4819,
      "num_input_tokens_seen": 9057632,
      "step": 13865
    },
    {
      "epoch": 7.269392033542977,
      "grad_norm": 0.2834559381008148,
      "learning_rate": 4.015363904052608e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 9061248,
      "step": 13870
    },
    {
      "epoch": 7.272012578616352,
      "grad_norm": 0.5301706194877625,
      "learning_rate": 4.014454315467199e-05,
      "loss": 0.4197,
      "num_input_tokens_seen": 9064800,
      "step": 13875
    },
    {
      "epoch": 7.2746331236897275,
      "grad_norm": 0.6040002107620239,
      "learning_rate": 4.013544410075224e-05,
      "loss": 0.4264,
      "num_input_tokens_seen": 9068928,
      "step": 13880
    },
    {
      "epoch": 7.277253668763103,
      "grad_norm": 0.3305199146270752,
      "learning_rate": 4.0126341880670273e-05,
      "loss": 0.339,
      "num_input_tokens_seen": 9072160,
      "step": 13885
    },
    {
      "epoch": 7.279874213836478,
      "grad_norm": 0.6599375605583191,
      "learning_rate": 4.011723649633014e-05,
      "loss": 0.4101,
      "num_input_tokens_seen": 9074784,
      "step": 13890
    },
    {
      "epoch": 7.282494758909853,
      "grad_norm": 0.392557829618454,
      "learning_rate": 4.0108127949636606e-05,
      "loss": 0.4,
      "num_input_tokens_seen": 9078688,
      "step": 13895
    },
    {
      "epoch": 7.285115303983228,
      "grad_norm": 0.6911959052085876,
      "learning_rate": 4.009901624249507e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 9081184,
      "step": 13900
    },
    {
      "epoch": 7.287735849056604,
      "grad_norm": 0.31636106967926025,
      "learning_rate": 4.008990137681159e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 9085344,
      "step": 13905
    },
    {
      "epoch": 7.290356394129979,
      "grad_norm": 0.448025643825531,
      "learning_rate": 4.00807833544929e-05,
      "loss": 0.4295,
      "num_input_tokens_seen": 9088384,
      "step": 13910
    },
    {
      "epoch": 7.2929769392033545,
      "grad_norm": 0.3887183368206024,
      "learning_rate": 4.0071662177446376e-05,
      "loss": 0.4242,
      "num_input_tokens_seen": 9090848,
      "step": 13915
    },
    {
      "epoch": 7.29559748427673,
      "grad_norm": 0.41494470834732056,
      "learning_rate": 4.006253784758008e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 9095328,
      "step": 13920
    },
    {
      "epoch": 7.298218029350105,
      "grad_norm": 0.5233990550041199,
      "learning_rate": 4.0053410366802704e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 9098464,
      "step": 13925
    },
    {
      "epoch": 7.30083857442348,
      "grad_norm": 0.39727288484573364,
      "learning_rate": 4.004427973702362e-05,
      "loss": 0.2827,
      "num_input_tokens_seen": 9103360,
      "step": 13930
    },
    {
      "epoch": 7.303459119496855,
      "grad_norm": 0.3571101427078247,
      "learning_rate": 4.003514596015285e-05,
      "loss": 0.3948,
      "num_input_tokens_seen": 9107328,
      "step": 13935
    },
    {
      "epoch": 7.30607966457023,
      "grad_norm": 0.4481819272041321,
      "learning_rate": 4.002600903810108e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 9110848,
      "step": 13940
    },
    {
      "epoch": 7.308700209643606,
      "grad_norm": 0.4012252688407898,
      "learning_rate": 4.001686897277963e-05,
      "loss": 0.5231,
      "num_input_tokens_seen": 9114112,
      "step": 13945
    },
    {
      "epoch": 7.311320754716981,
      "grad_norm": 1.901537537574768,
      "learning_rate": 4.000772576610052e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 9116512,
      "step": 13950
    },
    {
      "epoch": 7.313941299790357,
      "grad_norm": 0.3407653272151947,
      "learning_rate": 3.99985794199764e-05,
      "loss": 0.4563,
      "num_input_tokens_seen": 9119584,
      "step": 13955
    },
    {
      "epoch": 7.316561844863732,
      "grad_norm": 0.6065948009490967,
      "learning_rate": 3.998942993632058e-05,
      "loss": 0.5367,
      "num_input_tokens_seen": 9122784,
      "step": 13960
    },
    {
      "epoch": 7.319182389937107,
      "grad_norm": 0.3666752278804779,
      "learning_rate": 3.998027731704702e-05,
      "loss": 0.497,
      "num_input_tokens_seen": 9126080,
      "step": 13965
    },
    {
      "epoch": 7.321802935010482,
      "grad_norm": 0.9605438709259033,
      "learning_rate": 3.9971121564070354e-05,
      "loss": 0.474,
      "num_input_tokens_seen": 9128832,
      "step": 13970
    },
    {
      "epoch": 7.3244234800838575,
      "grad_norm": 0.5117751359939575,
      "learning_rate": 3.996196267930585e-05,
      "loss": 0.5842,
      "num_input_tokens_seen": 9132256,
      "step": 13975
    },
    {
      "epoch": 7.327044025157233,
      "grad_norm": 0.7967911958694458,
      "learning_rate": 3.995280066466945e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 9135584,
      "step": 13980
    },
    {
      "epoch": 7.329664570230608,
      "grad_norm": 0.5210265517234802,
      "learning_rate": 3.9943635522077746e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 9138368,
      "step": 13985
    },
    {
      "epoch": 7.332285115303983,
      "grad_norm": 0.46898242831230164,
      "learning_rate": 3.993446725344797e-05,
      "loss": 0.4471,
      "num_input_tokens_seen": 9141248,
      "step": 13990
    },
    {
      "epoch": 7.334905660377358,
      "grad_norm": 0.3895612061023712,
      "learning_rate": 3.992529586069803e-05,
      "loss": 0.4986,
      "num_input_tokens_seen": 9145536,
      "step": 13995
    },
    {
      "epoch": 7.337526205450734,
      "grad_norm": 0.5520787239074707,
      "learning_rate": 3.991612134574648e-05,
      "loss": 0.3906,
      "num_input_tokens_seen": 9148224,
      "step": 14000
    },
    {
      "epoch": 7.340146750524109,
      "grad_norm": 0.6411945223808289,
      "learning_rate": 3.99069437105125e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 9152096,
      "step": 14005
    },
    {
      "epoch": 7.3427672955974845,
      "grad_norm": 0.33814477920532227,
      "learning_rate": 3.9897762956915976e-05,
      "loss": 0.53,
      "num_input_tokens_seen": 9155232,
      "step": 14010
    },
    {
      "epoch": 7.34538784067086,
      "grad_norm": 0.34479832649230957,
      "learning_rate": 3.9888579086877403e-05,
      "loss": 0.3936,
      "num_input_tokens_seen": 9158720,
      "step": 14015
    },
    {
      "epoch": 7.348008385744235,
      "grad_norm": 0.38786911964416504,
      "learning_rate": 3.987939210231794e-05,
      "loss": 0.3192,
      "num_input_tokens_seen": 9162560,
      "step": 14020
    },
    {
      "epoch": 7.35062893081761,
      "grad_norm": 0.309940367937088,
      "learning_rate": 3.987020200515941e-05,
      "loss": 0.496,
      "num_input_tokens_seen": 9165504,
      "step": 14025
    },
    {
      "epoch": 7.353249475890985,
      "grad_norm": 0.38071078062057495,
      "learning_rate": 3.9861008797324255e-05,
      "loss": 0.4728,
      "num_input_tokens_seen": 9168288,
      "step": 14030
    },
    {
      "epoch": 7.35587002096436,
      "grad_norm": 0.5211071372032166,
      "learning_rate": 3.9851812480735605e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 9171168,
      "step": 14035
    },
    {
      "epoch": 7.3584905660377355,
      "grad_norm": 0.36945685744285583,
      "learning_rate": 3.9842613057317235e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 9176960,
      "step": 14040
    },
    {
      "epoch": 7.361111111111111,
      "grad_norm": 0.3321371078491211,
      "learning_rate": 3.9833410528993534e-05,
      "loss": 0.4468,
      "num_input_tokens_seen": 9181760,
      "step": 14045
    },
    {
      "epoch": 7.363731656184487,
      "grad_norm": 0.6749398112297058,
      "learning_rate": 3.9824204897689585e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 9185280,
      "step": 14050
    },
    {
      "epoch": 7.366352201257862,
      "grad_norm": 0.3746694326400757,
      "learning_rate": 3.98149961653311e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 9189184,
      "step": 14055
    },
    {
      "epoch": 7.368972746331237,
      "grad_norm": 0.41766175627708435,
      "learning_rate": 3.980578433384442e-05,
      "loss": 0.5214,
      "num_input_tokens_seen": 9191904,
      "step": 14060
    },
    {
      "epoch": 7.371593291404612,
      "grad_norm": 0.5760388970375061,
      "learning_rate": 3.979656940515659e-05,
      "loss": 0.3446,
      "num_input_tokens_seen": 9194496,
      "step": 14065
    },
    {
      "epoch": 7.3742138364779874,
      "grad_norm": 0.3827909231185913,
      "learning_rate": 3.9787351381195235e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 9197184,
      "step": 14070
    },
    {
      "epoch": 7.376834381551363,
      "grad_norm": 0.41616806387901306,
      "learning_rate": 3.977813026388868e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 9204128,
      "step": 14075
    },
    {
      "epoch": 7.379454926624738,
      "grad_norm": 0.33779147267341614,
      "learning_rate": 3.976890605516587e-05,
      "loss": 0.3805,
      "num_input_tokens_seen": 9206912,
      "step": 14080
    },
    {
      "epoch": 7.382075471698113,
      "grad_norm": 0.40756991505622864,
      "learning_rate": 3.97596787569564e-05,
      "loss": 0.4527,
      "num_input_tokens_seen": 9210144,
      "step": 14085
    },
    {
      "epoch": 7.384696016771488,
      "grad_norm": 0.31286776065826416,
      "learning_rate": 3.975044837119053e-05,
      "loss": 0.4425,
      "num_input_tokens_seen": 9214336,
      "step": 14090
    },
    {
      "epoch": 7.387316561844864,
      "grad_norm": 0.3536500334739685,
      "learning_rate": 3.974121489979914e-05,
      "loss": 0.4475,
      "num_input_tokens_seen": 9217728,
      "step": 14095
    },
    {
      "epoch": 7.389937106918239,
      "grad_norm": 0.5921961665153503,
      "learning_rate": 3.973197834471377e-05,
      "loss": 0.5007,
      "num_input_tokens_seen": 9221280,
      "step": 14100
    },
    {
      "epoch": 7.3925576519916145,
      "grad_norm": 0.39556992053985596,
      "learning_rate": 3.9722738707866595e-05,
      "loss": 0.4794,
      "num_input_tokens_seen": 9225376,
      "step": 14105
    },
    {
      "epoch": 7.39517819706499,
      "grad_norm": 0.36463698744773865,
      "learning_rate": 3.971349599119046e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 9229088,
      "step": 14110
    },
    {
      "epoch": 7.397798742138365,
      "grad_norm": 0.40332403779029846,
      "learning_rate": 3.970425019661881e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 9231840,
      "step": 14115
    },
    {
      "epoch": 7.40041928721174,
      "grad_norm": 0.5036874413490295,
      "learning_rate": 3.9695001326085785e-05,
      "loss": 0.475,
      "num_input_tokens_seen": 9235264,
      "step": 14120
    },
    {
      "epoch": 7.403039832285115,
      "grad_norm": 0.3207394480705261,
      "learning_rate": 3.968574938152613e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 9238656,
      "step": 14125
    },
    {
      "epoch": 7.40566037735849,
      "grad_norm": 0.34966662526130676,
      "learning_rate": 3.9676494364875247e-05,
      "loss": 0.5014,
      "num_input_tokens_seen": 9242144,
      "step": 14130
    },
    {
      "epoch": 7.4082809224318655,
      "grad_norm": 0.6809329986572266,
      "learning_rate": 3.966723627806918e-05,
      "loss": 0.5203,
      "num_input_tokens_seen": 9245024,
      "step": 14135
    },
    {
      "epoch": 7.410901467505241,
      "grad_norm": 0.5016713738441467,
      "learning_rate": 3.9657975123044624e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 9248032,
      "step": 14140
    },
    {
      "epoch": 7.413522012578617,
      "grad_norm": 0.6394686102867126,
      "learning_rate": 3.964871090173888e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 9250240,
      "step": 14145
    },
    {
      "epoch": 7.416142557651992,
      "grad_norm": 0.381326287984848,
      "learning_rate": 3.963944361608995e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 9253408,
      "step": 14150
    },
    {
      "epoch": 7.418763102725367,
      "grad_norm": 0.6397131681442261,
      "learning_rate": 3.963017326803644e-05,
      "loss": 0.4389,
      "num_input_tokens_seen": 9256320,
      "step": 14155
    },
    {
      "epoch": 7.421383647798742,
      "grad_norm": 0.2071445733308792,
      "learning_rate": 3.962089985951758e-05,
      "loss": 0.4536,
      "num_input_tokens_seen": 9261312,
      "step": 14160
    },
    {
      "epoch": 7.424004192872117,
      "grad_norm": 0.4355472922325134,
      "learning_rate": 3.961162339247327e-05,
      "loss": 0.552,
      "num_input_tokens_seen": 9264544,
      "step": 14165
    },
    {
      "epoch": 7.426624737945493,
      "grad_norm": 0.4019078314304352,
      "learning_rate": 3.9602343868844055e-05,
      "loss": 0.5218,
      "num_input_tokens_seen": 9267584,
      "step": 14170
    },
    {
      "epoch": 7.429245283018868,
      "grad_norm": 0.5952529311180115,
      "learning_rate": 3.959306129057109e-05,
      "loss": 0.4875,
      "num_input_tokens_seen": 9270784,
      "step": 14175
    },
    {
      "epoch": 7.431865828092243,
      "grad_norm": 0.3417656719684601,
      "learning_rate": 3.958377565959619e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 9274048,
      "step": 14180
    },
    {
      "epoch": 7.434486373165618,
      "grad_norm": 0.2150396704673767,
      "learning_rate": 3.95744869778618e-05,
      "loss": 0.3794,
      "num_input_tokens_seen": 9277760,
      "step": 14185
    },
    {
      "epoch": 7.437106918238994,
      "grad_norm": 0.743678092956543,
      "learning_rate": 3.956519524731101e-05,
      "loss": 0.5024,
      "num_input_tokens_seen": 9281088,
      "step": 14190
    },
    {
      "epoch": 7.439727463312369,
      "grad_norm": 0.3675294518470764,
      "learning_rate": 3.9555900469887555e-05,
      "loss": 0.5161,
      "num_input_tokens_seen": 9283968,
      "step": 14195
    },
    {
      "epoch": 7.4423480083857445,
      "grad_norm": 0.6466695666313171,
      "learning_rate": 3.954660264753577e-05,
      "loss": 0.5198,
      "num_input_tokens_seen": 9286816,
      "step": 14200
    },
    {
      "epoch": 7.44496855345912,
      "grad_norm": 0.38020631670951843,
      "learning_rate": 3.953730178220067e-05,
      "loss": 0.5264,
      "num_input_tokens_seen": 9290272,
      "step": 14205
    },
    {
      "epoch": 7.447589098532495,
      "grad_norm": 0.6451031565666199,
      "learning_rate": 3.952799787582789e-05,
      "loss": 0.5807,
      "num_input_tokens_seen": 9293568,
      "step": 14210
    },
    {
      "epoch": 7.45020964360587,
      "grad_norm": 0.999002993106842,
      "learning_rate": 3.9518690930363694e-05,
      "loss": 0.327,
      "num_input_tokens_seen": 9295968,
      "step": 14215
    },
    {
      "epoch": 7.452830188679245,
      "grad_norm": 0.4617979824542999,
      "learning_rate": 3.9509380947754994e-05,
      "loss": 0.3759,
      "num_input_tokens_seen": 9300128,
      "step": 14220
    },
    {
      "epoch": 7.45545073375262,
      "grad_norm": 0.9538396000862122,
      "learning_rate": 3.9500067929949336e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 9302816,
      "step": 14225
    },
    {
      "epoch": 7.4580712788259955,
      "grad_norm": 0.5107448101043701,
      "learning_rate": 3.9490751878894895e-05,
      "loss": 0.5442,
      "num_input_tokens_seen": 9305632,
      "step": 14230
    },
    {
      "epoch": 7.460691823899371,
      "grad_norm": 1.132587194442749,
      "learning_rate": 3.948143279654047e-05,
      "loss": 0.5715,
      "num_input_tokens_seen": 9308224,
      "step": 14235
    },
    {
      "epoch": 7.463312368972747,
      "grad_norm": 0.4526172876358032,
      "learning_rate": 3.947211068483553e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 9311008,
      "step": 14240
    },
    {
      "epoch": 7.465932914046122,
      "grad_norm": 0.4054047763347626,
      "learning_rate": 3.946278554573013e-05,
      "loss": 0.3765,
      "num_input_tokens_seen": 9314272,
      "step": 14245
    },
    {
      "epoch": 7.468553459119497,
      "grad_norm": 0.522346556186676,
      "learning_rate": 3.9453457381175e-05,
      "loss": 0.7299,
      "num_input_tokens_seen": 9316896,
      "step": 14250
    },
    {
      "epoch": 7.471174004192872,
      "grad_norm": 0.44756513833999634,
      "learning_rate": 3.9444126193121474e-05,
      "loss": 0.3222,
      "num_input_tokens_seen": 9320896,
      "step": 14255
    },
    {
      "epoch": 7.473794549266247,
      "grad_norm": 0.3845694959163666,
      "learning_rate": 3.943479198352153e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 9324832,
      "step": 14260
    },
    {
      "epoch": 7.476415094339623,
      "grad_norm": 0.4911576509475708,
      "learning_rate": 3.94254547543278e-05,
      "loss": 0.4042,
      "num_input_tokens_seen": 9328000,
      "step": 14265
    },
    {
      "epoch": 7.479035639412998,
      "grad_norm": 0.3706408143043518,
      "learning_rate": 3.9416114507493504e-05,
      "loss": 0.4233,
      "num_input_tokens_seen": 9330400,
      "step": 14270
    },
    {
      "epoch": 7.481656184486373,
      "grad_norm": 0.5956059694290161,
      "learning_rate": 3.940677124497252e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 9332992,
      "step": 14275
    },
    {
      "epoch": 7.484276729559748,
      "grad_norm": 0.5288172364234924,
      "learning_rate": 3.939742496871935e-05,
      "loss": 0.5414,
      "num_input_tokens_seen": 9335968,
      "step": 14280
    },
    {
      "epoch": 7.486897274633124,
      "grad_norm": 0.44435954093933105,
      "learning_rate": 3.9388075680689136e-05,
      "loss": 0.332,
      "num_input_tokens_seen": 9339456,
      "step": 14285
    },
    {
      "epoch": 7.489517819706499,
      "grad_norm": 0.5592369437217712,
      "learning_rate": 3.9378723382837634e-05,
      "loss": 0.442,
      "num_input_tokens_seen": 9343520,
      "step": 14290
    },
    {
      "epoch": 7.4921383647798745,
      "grad_norm": 0.5647506713867188,
      "learning_rate": 3.936936807712124e-05,
      "loss": 0.5173,
      "num_input_tokens_seen": 9346816,
      "step": 14295
    },
    {
      "epoch": 7.49475890985325,
      "grad_norm": 0.5726174712181091,
      "learning_rate": 3.936000976549698e-05,
      "loss": 0.3944,
      "num_input_tokens_seen": 9349856,
      "step": 14300
    },
    {
      "epoch": 7.497379454926625,
      "grad_norm": 0.5471824407577515,
      "learning_rate": 3.935064844992251e-05,
      "loss": 0.5211,
      "num_input_tokens_seen": 9353504,
      "step": 14305
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.36847367882728577,
      "learning_rate": 3.93412841323561e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 9357504,
      "step": 14310
    },
    {
      "epoch": 7.502620545073375,
      "grad_norm": 0.7131635546684265,
      "learning_rate": 3.933191681475666e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 9361504,
      "step": 14315
    },
    {
      "epoch": 7.50524109014675,
      "grad_norm": 0.34764373302459717,
      "learning_rate": 3.932254649908373e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 9365760,
      "step": 14320
    },
    {
      "epoch": 7.5078616352201255,
      "grad_norm": 0.5188412666320801,
      "learning_rate": 3.931317318729748e-05,
      "loss": 0.5637,
      "num_input_tokens_seen": 9368384,
      "step": 14325
    },
    {
      "epoch": 7.510482180293501,
      "grad_norm": 0.7124032378196716,
      "learning_rate": 3.930379688135869e-05,
      "loss": 0.5268,
      "num_input_tokens_seen": 9371360,
      "step": 14330
    },
    {
      "epoch": 7.513102725366876,
      "grad_norm": 0.48690274357795715,
      "learning_rate": 3.929441758322878e-05,
      "loss": 0.5561,
      "num_input_tokens_seen": 9374624,
      "step": 14335
    },
    {
      "epoch": 7.515723270440252,
      "grad_norm": 0.30943161249160767,
      "learning_rate": 3.928503529486979e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 9378048,
      "step": 14340
    },
    {
      "epoch": 7.518343815513627,
      "grad_norm": 0.3844745457172394,
      "learning_rate": 3.927565001824438e-05,
      "loss": 0.4742,
      "num_input_tokens_seen": 9381120,
      "step": 14345
    },
    {
      "epoch": 7.520964360587002,
      "grad_norm": 0.4417717456817627,
      "learning_rate": 3.926626175531586e-05,
      "loss": 0.3672,
      "num_input_tokens_seen": 9384736,
      "step": 14350
    },
    {
      "epoch": 7.523584905660377,
      "grad_norm": 0.32679319381713867,
      "learning_rate": 3.925687050804814e-05,
      "loss": 0.3689,
      "num_input_tokens_seen": 9388192,
      "step": 14355
    },
    {
      "epoch": 7.526205450733753,
      "grad_norm": 0.4886597990989685,
      "learning_rate": 3.924747627840575e-05,
      "loss": 0.3427,
      "num_input_tokens_seen": 9391168,
      "step": 14360
    },
    {
      "epoch": 7.528825995807128,
      "grad_norm": 0.5815616250038147,
      "learning_rate": 3.923807906835388e-05,
      "loss": 0.5382,
      "num_input_tokens_seen": 9394528,
      "step": 14365
    },
    {
      "epoch": 7.531446540880503,
      "grad_norm": 0.42802223563194275,
      "learning_rate": 3.92286788798583e-05,
      "loss": 0.448,
      "num_input_tokens_seen": 9398784,
      "step": 14370
    },
    {
      "epoch": 7.534067085953878,
      "grad_norm": 0.5502181053161621,
      "learning_rate": 3.921927571488543e-05,
      "loss": 0.5554,
      "num_input_tokens_seen": 9402176,
      "step": 14375
    },
    {
      "epoch": 7.536687631027254,
      "grad_norm": 0.8524311780929565,
      "learning_rate": 3.9209869575402296e-05,
      "loss": 0.364,
      "num_input_tokens_seen": 9405344,
      "step": 14380
    },
    {
      "epoch": 7.539308176100629,
      "grad_norm": 0.3381570279598236,
      "learning_rate": 3.9200460463376566e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 9408384,
      "step": 14385
    },
    {
      "epoch": 7.5419287211740045,
      "grad_norm": 0.5258088707923889,
      "learning_rate": 3.9191048380776515e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 9410976,
      "step": 14390
    },
    {
      "epoch": 7.54454926624738,
      "grad_norm": 0.3178010880947113,
      "learning_rate": 3.918163332957104e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 9413632,
      "step": 14395
    },
    {
      "epoch": 7.547169811320755,
      "grad_norm": 0.7016047239303589,
      "learning_rate": 3.9172215311729665e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 9416640,
      "step": 14400
    },
    {
      "epoch": 7.54979035639413,
      "grad_norm": 0.3524203598499298,
      "learning_rate": 3.916279432922254e-05,
      "loss": 0.5267,
      "num_input_tokens_seen": 9419584,
      "step": 14405
    },
    {
      "epoch": 7.552410901467505,
      "grad_norm": 0.5389264822006226,
      "learning_rate": 3.915337038402041e-05,
      "loss": 0.4617,
      "num_input_tokens_seen": 9422624,
      "step": 14410
    },
    {
      "epoch": 7.55503144654088,
      "grad_norm": 0.6892464756965637,
      "learning_rate": 3.914394347809467e-05,
      "loss": 0.3296,
      "num_input_tokens_seen": 9424832,
      "step": 14415
    },
    {
      "epoch": 7.5576519916142555,
      "grad_norm": 0.5611845254898071,
      "learning_rate": 3.9134513613417315e-05,
      "loss": 0.6292,
      "num_input_tokens_seen": 9427360,
      "step": 14420
    },
    {
      "epoch": 7.560272536687631,
      "grad_norm": 0.4848403036594391,
      "learning_rate": 3.912508079196097e-05,
      "loss": 0.5427,
      "num_input_tokens_seen": 9430112,
      "step": 14425
    },
    {
      "epoch": 7.562893081761006,
      "grad_norm": 0.38263750076293945,
      "learning_rate": 3.9115645015698866e-05,
      "loss": 0.4466,
      "num_input_tokens_seen": 9434560,
      "step": 14430
    },
    {
      "epoch": 7.565513626834382,
      "grad_norm": 0.9663940668106079,
      "learning_rate": 3.9106206286604865e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 9437664,
      "step": 14435
    },
    {
      "epoch": 7.568134171907757,
      "grad_norm": 0.2697952389717102,
      "learning_rate": 3.909676460665345e-05,
      "loss": 0.2801,
      "num_input_tokens_seen": 9441472,
      "step": 14440
    },
    {
      "epoch": 7.570754716981132,
      "grad_norm": 0.4262256324291229,
      "learning_rate": 3.908731997781969e-05,
      "loss": 0.3373,
      "num_input_tokens_seen": 9444288,
      "step": 14445
    },
    {
      "epoch": 7.573375262054507,
      "grad_norm": 0.4511813223361969,
      "learning_rate": 3.907787240207932e-05,
      "loss": 0.4363,
      "num_input_tokens_seen": 9448256,
      "step": 14450
    },
    {
      "epoch": 7.575995807127883,
      "grad_norm": 0.258451372385025,
      "learning_rate": 3.9068421881408645e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 9452768,
      "step": 14455
    },
    {
      "epoch": 7.578616352201258,
      "grad_norm": 0.341372013092041,
      "learning_rate": 3.9058968417784614e-05,
      "loss": 0.5637,
      "num_input_tokens_seen": 9456288,
      "step": 14460
    },
    {
      "epoch": 7.581236897274633,
      "grad_norm": 0.9688933491706848,
      "learning_rate": 3.9049512013184785e-05,
      "loss": 0.5307,
      "num_input_tokens_seen": 9459104,
      "step": 14465
    },
    {
      "epoch": 7.583857442348008,
      "grad_norm": 0.627156138420105,
      "learning_rate": 3.9040052669587325e-05,
      "loss": 0.4347,
      "num_input_tokens_seen": 9462688,
      "step": 14470
    },
    {
      "epoch": 7.586477987421384,
      "grad_norm": 0.39816558361053467,
      "learning_rate": 3.9030590388971026e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 9465440,
      "step": 14475
    },
    {
      "epoch": 7.589098532494759,
      "grad_norm": 0.6572660803794861,
      "learning_rate": 3.902112517331529e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 9469184,
      "step": 14480
    },
    {
      "epoch": 7.5917190775681345,
      "grad_norm": 0.31143879890441895,
      "learning_rate": 3.901165702460012e-05,
      "loss": 0.5577,
      "num_input_tokens_seen": 9472704,
      "step": 14485
    },
    {
      "epoch": 7.59433962264151,
      "grad_norm": 1.190742015838623,
      "learning_rate": 3.900218594480616e-05,
      "loss": 0.5751,
      "num_input_tokens_seen": 9476224,
      "step": 14490
    },
    {
      "epoch": 7.596960167714885,
      "grad_norm": 0.44911229610443115,
      "learning_rate": 3.899271193591465e-05,
      "loss": 0.5458,
      "num_input_tokens_seen": 9479264,
      "step": 14495
    },
    {
      "epoch": 7.59958071278826,
      "grad_norm": 0.4940272271633148,
      "learning_rate": 3.8983234999907435e-05,
      "loss": 0.4795,
      "num_input_tokens_seen": 9481824,
      "step": 14500
    },
    {
      "epoch": 7.602201257861635,
      "grad_norm": 0.4764665961265564,
      "learning_rate": 3.897375513876699e-05,
      "loss": 0.3557,
      "num_input_tokens_seen": 9485376,
      "step": 14505
    },
    {
      "epoch": 7.60482180293501,
      "grad_norm": 0.5390456318855286,
      "learning_rate": 3.8964272354476396e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 9488672,
      "step": 14510
    },
    {
      "epoch": 7.6074423480083855,
      "grad_norm": 0.3999389708042145,
      "learning_rate": 3.895478664901933e-05,
      "loss": 0.5806,
      "num_input_tokens_seen": 9491776,
      "step": 14515
    },
    {
      "epoch": 7.610062893081761,
      "grad_norm": 0.6546087265014648,
      "learning_rate": 3.89452980243801e-05,
      "loss": 0.4186,
      "num_input_tokens_seen": 9494624,
      "step": 14520
    },
    {
      "epoch": 7.612683438155136,
      "grad_norm": 0.49795615673065186,
      "learning_rate": 3.893580648254362e-05,
      "loss": 0.5698,
      "num_input_tokens_seen": 9497472,
      "step": 14525
    },
    {
      "epoch": 7.615303983228512,
      "grad_norm": 0.6930887699127197,
      "learning_rate": 3.8926312025495414e-05,
      "loss": 0.4641,
      "num_input_tokens_seen": 9500928,
      "step": 14530
    },
    {
      "epoch": 7.617924528301887,
      "grad_norm": 0.33526426553726196,
      "learning_rate": 3.891681465522161e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 9504384,
      "step": 14535
    },
    {
      "epoch": 7.620545073375262,
      "grad_norm": 0.44556236267089844,
      "learning_rate": 3.890731437370896e-05,
      "loss": 0.4578,
      "num_input_tokens_seen": 9506912,
      "step": 14540
    },
    {
      "epoch": 7.623165618448637,
      "grad_norm": 0.4793689548969269,
      "learning_rate": 3.88978111829448e-05,
      "loss": 0.3474,
      "num_input_tokens_seen": 9509632,
      "step": 14545
    },
    {
      "epoch": 7.6257861635220126,
      "grad_norm": 0.4744497239589691,
      "learning_rate": 3.88883050849171e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 9513024,
      "step": 14550
    },
    {
      "epoch": 7.628406708595388,
      "grad_norm": 0.5521525144577026,
      "learning_rate": 3.8878796081614417e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 9515936,
      "step": 14555
    },
    {
      "epoch": 7.631027253668763,
      "grad_norm": 0.43281784653663635,
      "learning_rate": 3.8869284175025935e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 9519744,
      "step": 14560
    },
    {
      "epoch": 7.633647798742138,
      "grad_norm": 0.7168855667114258,
      "learning_rate": 3.885976936714143e-05,
      "loss": 0.5592,
      "num_input_tokens_seen": 9522464,
      "step": 14565
    },
    {
      "epoch": 7.636268343815514,
      "grad_norm": 0.7753410339355469,
      "learning_rate": 3.885025165995129e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 9527200,
      "step": 14570
    },
    {
      "epoch": 7.638888888888889,
      "grad_norm": 0.5954751372337341,
      "learning_rate": 3.884073105544652e-05,
      "loss": 0.6222,
      "num_input_tokens_seen": 9530336,
      "step": 14575
    },
    {
      "epoch": 7.6415094339622645,
      "grad_norm": 0.473769873380661,
      "learning_rate": 3.8831207555618723e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 9533728,
      "step": 14580
    },
    {
      "epoch": 7.64412997903564,
      "grad_norm": 0.7033950090408325,
      "learning_rate": 3.882168116246009e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 9536576,
      "step": 14585
    },
    {
      "epoch": 7.646750524109015,
      "grad_norm": 0.37104693055152893,
      "learning_rate": 3.881215187796344e-05,
      "loss": 0.3615,
      "num_input_tokens_seen": 9540160,
      "step": 14590
    },
    {
      "epoch": 7.64937106918239,
      "grad_norm": 0.47889062762260437,
      "learning_rate": 3.8802619704122204e-05,
      "loss": 0.4763,
      "num_input_tokens_seen": 9542976,
      "step": 14595
    },
    {
      "epoch": 7.651991614255765,
      "grad_norm": 0.39240339398384094,
      "learning_rate": 3.8793084642930387e-05,
      "loss": 0.4236,
      "num_input_tokens_seen": 9546240,
      "step": 14600
    },
    {
      "epoch": 7.65461215932914,
      "grad_norm": 0.5070850253105164,
      "learning_rate": 3.878354669638263e-05,
      "loss": 0.5251,
      "num_input_tokens_seen": 9549024,
      "step": 14605
    },
    {
      "epoch": 7.6572327044025155,
      "grad_norm": 0.6203898191452026,
      "learning_rate": 3.8774005866474155e-05,
      "loss": 0.5716,
      "num_input_tokens_seen": 9551328,
      "step": 14610
    },
    {
      "epoch": 7.659853249475891,
      "grad_norm": 0.26767459511756897,
      "learning_rate": 3.876446215520079e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 9554880,
      "step": 14615
    },
    {
      "epoch": 7.662473794549266,
      "grad_norm": 0.8840593695640564,
      "learning_rate": 3.875491556455898e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 9557536,
      "step": 14620
    },
    {
      "epoch": 7.665094339622642,
      "grad_norm": 0.439301073551178,
      "learning_rate": 3.8745366096545754e-05,
      "loss": 0.3575,
      "num_input_tokens_seen": 9560544,
      "step": 14625
    },
    {
      "epoch": 7.667714884696017,
      "grad_norm": 0.5005175471305847,
      "learning_rate": 3.873581375315876e-05,
      "loss": 0.6454,
      "num_input_tokens_seen": 9563392,
      "step": 14630
    },
    {
      "epoch": 7.670335429769392,
      "grad_norm": 0.77032470703125,
      "learning_rate": 3.872625853639623e-05,
      "loss": 0.3993,
      "num_input_tokens_seen": 9565984,
      "step": 14635
    },
    {
      "epoch": 7.672955974842767,
      "grad_norm": 0.3351028263568878,
      "learning_rate": 3.871670044825701e-05,
      "loss": 0.3914,
      "num_input_tokens_seen": 9569504,
      "step": 14640
    },
    {
      "epoch": 7.6755765199161425,
      "grad_norm": 0.4099217355251312,
      "learning_rate": 3.870713949074056e-05,
      "loss": 0.6095,
      "num_input_tokens_seen": 9572832,
      "step": 14645
    },
    {
      "epoch": 7.678197064989518,
      "grad_norm": 0.5691353678703308,
      "learning_rate": 3.86975756658469e-05,
      "loss": 0.5294,
      "num_input_tokens_seen": 9575488,
      "step": 14650
    },
    {
      "epoch": 7.680817610062893,
      "grad_norm": 0.5227708220481873,
      "learning_rate": 3.868800897557667e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 9578304,
      "step": 14655
    },
    {
      "epoch": 7.683438155136268,
      "grad_norm": 0.40216535329818726,
      "learning_rate": 3.867843942193113e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 9581440,
      "step": 14660
    },
    {
      "epoch": 7.686058700209644,
      "grad_norm": 0.38867443799972534,
      "learning_rate": 3.866886700691212e-05,
      "loss": 0.4807,
      "num_input_tokens_seen": 9584032,
      "step": 14665
    },
    {
      "epoch": 7.688679245283019,
      "grad_norm": 0.6475823521614075,
      "learning_rate": 3.8659291732522075e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 9588384,
      "step": 14670
    },
    {
      "epoch": 7.691299790356394,
      "grad_norm": 0.32028457522392273,
      "learning_rate": 3.8649713600764026e-05,
      "loss": 0.4504,
      "num_input_tokens_seen": 9593120,
      "step": 14675
    },
    {
      "epoch": 7.69392033542977,
      "grad_norm": 0.273234099149704,
      "learning_rate": 3.864013261364162e-05,
      "loss": 0.4962,
      "num_input_tokens_seen": 9596480,
      "step": 14680
    },
    {
      "epoch": 7.696540880503145,
      "grad_norm": 0.4422357380390167,
      "learning_rate": 3.8630548773159084e-05,
      "loss": 0.4483,
      "num_input_tokens_seen": 9599424,
      "step": 14685
    },
    {
      "epoch": 7.69916142557652,
      "grad_norm": 0.7365807294845581,
      "learning_rate": 3.862096208132125e-05,
      "loss": 0.3917,
      "num_input_tokens_seen": 9603648,
      "step": 14690
    },
    {
      "epoch": 7.701781970649895,
      "grad_norm": 0.49353963136672974,
      "learning_rate": 3.861137254013355e-05,
      "loss": 0.3942,
      "num_input_tokens_seen": 9606208,
      "step": 14695
    },
    {
      "epoch": 7.70440251572327,
      "grad_norm": 0.4634001851081848,
      "learning_rate": 3.8601780151601996e-05,
      "loss": 0.4233,
      "num_input_tokens_seen": 9610176,
      "step": 14700
    },
    {
      "epoch": 7.7070230607966455,
      "grad_norm": 0.5747643709182739,
      "learning_rate": 3.859218491773321e-05,
      "loss": 0.5202,
      "num_input_tokens_seen": 9613888,
      "step": 14705
    },
    {
      "epoch": 7.709643605870021,
      "grad_norm": 0.44985851645469666,
      "learning_rate": 3.858258684053441e-05,
      "loss": 0.3489,
      "num_input_tokens_seen": 9617792,
      "step": 14710
    },
    {
      "epoch": 7.712264150943396,
      "grad_norm": 0.5226224660873413,
      "learning_rate": 3.8572985922013405e-05,
      "loss": 0.4087,
      "num_input_tokens_seen": 9620832,
      "step": 14715
    },
    {
      "epoch": 7.714884696016772,
      "grad_norm": 0.35755351185798645,
      "learning_rate": 3.856338216417859e-05,
      "loss": 0.5031,
      "num_input_tokens_seen": 9623744,
      "step": 14720
    },
    {
      "epoch": 7.717505241090147,
      "grad_norm": 0.4812719225883484,
      "learning_rate": 3.855377556903897e-05,
      "loss": 0.5438,
      "num_input_tokens_seen": 9627136,
      "step": 14725
    },
    {
      "epoch": 7.720125786163522,
      "grad_norm": 1.241239309310913,
      "learning_rate": 3.8544166138604124e-05,
      "loss": 0.5648,
      "num_input_tokens_seen": 9630592,
      "step": 14730
    },
    {
      "epoch": 7.722746331236897,
      "grad_norm": 0.585163950920105,
      "learning_rate": 3.8534553874884244e-05,
      "loss": 0.4916,
      "num_input_tokens_seen": 9634240,
      "step": 14735
    },
    {
      "epoch": 7.7253668763102725,
      "grad_norm": 0.6955058574676514,
      "learning_rate": 3.852493877989011e-05,
      "loss": 0.6118,
      "num_input_tokens_seen": 9637120,
      "step": 14740
    },
    {
      "epoch": 7.727987421383648,
      "grad_norm": 1.0382753610610962,
      "learning_rate": 3.851532085563308e-05,
      "loss": 0.3935,
      "num_input_tokens_seen": 9639808,
      "step": 14745
    },
    {
      "epoch": 7.730607966457023,
      "grad_norm": 0.29686224460601807,
      "learning_rate": 3.850570010412511e-05,
      "loss": 0.6181,
      "num_input_tokens_seen": 9644384,
      "step": 14750
    },
    {
      "epoch": 7.733228511530398,
      "grad_norm": 0.29330015182495117,
      "learning_rate": 3.8496076527378766e-05,
      "loss": 0.4512,
      "num_input_tokens_seen": 9648288,
      "step": 14755
    },
    {
      "epoch": 7.735849056603773,
      "grad_norm": 0.3393269181251526,
      "learning_rate": 3.848645012740718e-05,
      "loss": 0.4567,
      "num_input_tokens_seen": 9652032,
      "step": 14760
    },
    {
      "epoch": 7.738469601677149,
      "grad_norm": 0.3665097951889038,
      "learning_rate": 3.8476820906224096e-05,
      "loss": 0.5025,
      "num_input_tokens_seen": 9655296,
      "step": 14765
    },
    {
      "epoch": 7.741090146750524,
      "grad_norm": 0.6118259429931641,
      "learning_rate": 3.846718886584382e-05,
      "loss": 0.4532,
      "num_input_tokens_seen": 9657856,
      "step": 14770
    },
    {
      "epoch": 7.7437106918239,
      "grad_norm": 0.4120965003967285,
      "learning_rate": 3.845755400828128e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 9660576,
      "step": 14775
    },
    {
      "epoch": 7.746331236897275,
      "grad_norm": 0.5894432663917542,
      "learning_rate": 3.844791633555198e-05,
      "loss": 0.6182,
      "num_input_tokens_seen": 9663360,
      "step": 14780
    },
    {
      "epoch": 7.74895178197065,
      "grad_norm": 0.3038877546787262,
      "learning_rate": 3.8438275849671996e-05,
      "loss": 0.4291,
      "num_input_tokens_seen": 9667200,
      "step": 14785
    },
    {
      "epoch": 7.751572327044025,
      "grad_norm": 0.2849988341331482,
      "learning_rate": 3.842863255265802e-05,
      "loss": 0.5226,
      "num_input_tokens_seen": 9671136,
      "step": 14790
    },
    {
      "epoch": 7.7541928721174,
      "grad_norm": 0.7787024974822998,
      "learning_rate": 3.841898644652732e-05,
      "loss": 0.4866,
      "num_input_tokens_seen": 9674208,
      "step": 14795
    },
    {
      "epoch": 7.756813417190775,
      "grad_norm": 0.5664849877357483,
      "learning_rate": 3.840933753329774e-05,
      "loss": 0.4695,
      "num_input_tokens_seen": 9677280,
      "step": 14800
    },
    {
      "epoch": 7.759433962264151,
      "grad_norm": 0.5628212690353394,
      "learning_rate": 3.839968581498773e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 9680320,
      "step": 14805
    },
    {
      "epoch": 7.762054507337526,
      "grad_norm": 0.5209387540817261,
      "learning_rate": 3.839003129361632e-05,
      "loss": 0.6292,
      "num_input_tokens_seen": 9682848,
      "step": 14810
    },
    {
      "epoch": 7.764675052410902,
      "grad_norm": 0.5061444044113159,
      "learning_rate": 3.8380373971203134e-05,
      "loss": 0.4845,
      "num_input_tokens_seen": 9686112,
      "step": 14815
    },
    {
      "epoch": 7.767295597484277,
      "grad_norm": 0.36020100116729736,
      "learning_rate": 3.8370713849768346e-05,
      "loss": 0.5742,
      "num_input_tokens_seen": 9689088,
      "step": 14820
    },
    {
      "epoch": 7.769916142557652,
      "grad_norm": 2.049767017364502,
      "learning_rate": 3.8361050931332774e-05,
      "loss": 0.5339,
      "num_input_tokens_seen": 9693280,
      "step": 14825
    },
    {
      "epoch": 7.772536687631027,
      "grad_norm": 0.4293545186519623,
      "learning_rate": 3.835138521791778e-05,
      "loss": 0.421,
      "num_input_tokens_seen": 9695872,
      "step": 14830
    },
    {
      "epoch": 7.7751572327044025,
      "grad_norm": 0.8911715745925903,
      "learning_rate": 3.8341716711545314e-05,
      "loss": 0.5731,
      "num_input_tokens_seen": 9698272,
      "step": 14835
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 0.8842140436172485,
      "learning_rate": 3.833204541423793e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 9701568,
      "step": 14840
    },
    {
      "epoch": 7.780398322851153,
      "grad_norm": 0.34613099694252014,
      "learning_rate": 3.832237132801874e-05,
      "loss": 0.4312,
      "num_input_tokens_seen": 9704832,
      "step": 14845
    },
    {
      "epoch": 7.783018867924528,
      "grad_norm": 0.5741600394248962,
      "learning_rate": 3.831269445491146e-05,
      "loss": 0.623,
      "num_input_tokens_seen": 9707616,
      "step": 14850
    },
    {
      "epoch": 7.785639412997903,
      "grad_norm": 0.41018351912498474,
      "learning_rate": 3.8303014796940375e-05,
      "loss": 0.468,
      "num_input_tokens_seen": 9710560,
      "step": 14855
    },
    {
      "epoch": 7.788259958071279,
      "grad_norm": 0.4752562642097473,
      "learning_rate": 3.829333235613036e-05,
      "loss": 0.5725,
      "num_input_tokens_seen": 9713984,
      "step": 14860
    },
    {
      "epoch": 7.790880503144654,
      "grad_norm": 0.5118677616119385,
      "learning_rate": 3.828364713450688e-05,
      "loss": 0.3199,
      "num_input_tokens_seen": 9716832,
      "step": 14865
    },
    {
      "epoch": 7.79350104821803,
      "grad_norm": 0.4803861081600189,
      "learning_rate": 3.827395913409597e-05,
      "loss": 0.4631,
      "num_input_tokens_seen": 9720544,
      "step": 14870
    },
    {
      "epoch": 7.796121593291405,
      "grad_norm": 0.3846943974494934,
      "learning_rate": 3.826426835692424e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 9723392,
      "step": 14875
    },
    {
      "epoch": 7.79874213836478,
      "grad_norm": 0.46275296807289124,
      "learning_rate": 3.825457480501891e-05,
      "loss": 0.4072,
      "num_input_tokens_seen": 9726464,
      "step": 14880
    },
    {
      "epoch": 7.801362683438155,
      "grad_norm": 0.5331300497055054,
      "learning_rate": 3.824487848040773e-05,
      "loss": 0.3973,
      "num_input_tokens_seen": 9729952,
      "step": 14885
    },
    {
      "epoch": 7.80398322851153,
      "grad_norm": 0.5549814105033875,
      "learning_rate": 3.823517938511908e-05,
      "loss": 0.5781,
      "num_input_tokens_seen": 9732832,
      "step": 14890
    },
    {
      "epoch": 7.806603773584905,
      "grad_norm": 0.409917950630188,
      "learning_rate": 3.8225477521181906e-05,
      "loss": 0.516,
      "num_input_tokens_seen": 9736736,
      "step": 14895
    },
    {
      "epoch": 7.809224318658281,
      "grad_norm": 0.5010854601860046,
      "learning_rate": 3.821577289062571e-05,
      "loss": 0.4232,
      "num_input_tokens_seen": 9739968,
      "step": 14900
    },
    {
      "epoch": 7.811844863731656,
      "grad_norm": 0.5093492865562439,
      "learning_rate": 3.82060654954806e-05,
      "loss": 0.3613,
      "num_input_tokens_seen": 9742528,
      "step": 14905
    },
    {
      "epoch": 7.814465408805032,
      "grad_norm": 0.42307791113853455,
      "learning_rate": 3.819635533777726e-05,
      "loss": 0.442,
      "num_input_tokens_seen": 9746752,
      "step": 14910
    },
    {
      "epoch": 7.817085953878407,
      "grad_norm": 0.4585382640361786,
      "learning_rate": 3.8186642419546924e-05,
      "loss": 0.5096,
      "num_input_tokens_seen": 9750176,
      "step": 14915
    },
    {
      "epoch": 7.819706498951782,
      "grad_norm": 0.30700135231018066,
      "learning_rate": 3.817692674282145e-05,
      "loss": 0.5179,
      "num_input_tokens_seen": 9752768,
      "step": 14920
    },
    {
      "epoch": 7.822327044025157,
      "grad_norm": 0.5549469590187073,
      "learning_rate": 3.816720830963322e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 9756352,
      "step": 14925
    },
    {
      "epoch": 7.8249475890985325,
      "grad_norm": 0.4852515757083893,
      "learning_rate": 3.815748712201523e-05,
      "loss": 0.4257,
      "num_input_tokens_seen": 9759264,
      "step": 14930
    },
    {
      "epoch": 7.827568134171908,
      "grad_norm": 0.6642870903015137,
      "learning_rate": 3.814776318200105e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 9766976,
      "step": 14935
    },
    {
      "epoch": 7.830188679245283,
      "grad_norm": 0.5315698385238647,
      "learning_rate": 3.8138036491624805e-05,
      "loss": 0.6175,
      "num_input_tokens_seen": 9769984,
      "step": 14940
    },
    {
      "epoch": 7.832809224318658,
      "grad_norm": 0.3450002670288086,
      "learning_rate": 3.812830705292122e-05,
      "loss": 0.5828,
      "num_input_tokens_seen": 9772864,
      "step": 14945
    },
    {
      "epoch": 7.835429769392033,
      "grad_norm": 0.6863869428634644,
      "learning_rate": 3.8118574867925564e-05,
      "loss": 0.374,
      "num_input_tokens_seen": 9775264,
      "step": 14950
    },
    {
      "epoch": 7.838050314465409,
      "grad_norm": 0.28562769293785095,
      "learning_rate": 3.810883993867371e-05,
      "loss": 0.3819,
      "num_input_tokens_seen": 9779584,
      "step": 14955
    },
    {
      "epoch": 7.840670859538784,
      "grad_norm": 0.4238959550857544,
      "learning_rate": 3.80991022672021e-05,
      "loss": 0.4333,
      "num_input_tokens_seen": 9782784,
      "step": 14960
    },
    {
      "epoch": 7.84329140461216,
      "grad_norm": 0.5838655233383179,
      "learning_rate": 3.8089361855547744e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 9784960,
      "step": 14965
    },
    {
      "epoch": 7.845911949685535,
      "grad_norm": 0.6880454421043396,
      "learning_rate": 3.8079618705748214e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 9787904,
      "step": 14970
    },
    {
      "epoch": 7.84853249475891,
      "grad_norm": 0.9022308588027954,
      "learning_rate": 3.806987281984167e-05,
      "loss": 0.8092,
      "num_input_tokens_seen": 9790752,
      "step": 14975
    },
    {
      "epoch": 7.851153039832285,
      "grad_norm": 0.3448931574821472,
      "learning_rate": 3.806012419986684e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 9794624,
      "step": 14980
    },
    {
      "epoch": 7.85377358490566,
      "grad_norm": 0.31834205985069275,
      "learning_rate": 3.805037284786303e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 9797920,
      "step": 14985
    },
    {
      "epoch": 7.856394129979035,
      "grad_norm": 0.49662500619888306,
      "learning_rate": 3.804061876587009e-05,
      "loss": 0.5949,
      "num_input_tokens_seen": 9800896,
      "step": 14990
    },
    {
      "epoch": 7.859014675052411,
      "grad_norm": 0.5202686190605164,
      "learning_rate": 3.80308619559285e-05,
      "loss": 0.3144,
      "num_input_tokens_seen": 9803872,
      "step": 14995
    },
    {
      "epoch": 7.861635220125786,
      "grad_norm": 0.41924259066581726,
      "learning_rate": 3.802110242007925e-05,
      "loss": 0.4693,
      "num_input_tokens_seen": 9807136,
      "step": 15000
    },
    {
      "epoch": 7.864255765199162,
      "grad_norm": 0.3814500868320465,
      "learning_rate": 3.801134016036392e-05,
      "loss": 0.5151,
      "num_input_tokens_seen": 9810880,
      "step": 15005
    },
    {
      "epoch": 7.866876310272537,
      "grad_norm": 0.4153808355331421,
      "learning_rate": 3.8001575178824665e-05,
      "loss": 0.3283,
      "num_input_tokens_seen": 9814752,
      "step": 15010
    },
    {
      "epoch": 7.869496855345912,
      "grad_norm": 0.5094525218009949,
      "learning_rate": 3.799180747750422e-05,
      "loss": 0.4995,
      "num_input_tokens_seen": 9818592,
      "step": 15015
    },
    {
      "epoch": 7.872117400419287,
      "grad_norm": 0.42319056391716003,
      "learning_rate": 3.798203705844586e-05,
      "loss": 0.5491,
      "num_input_tokens_seen": 9821152,
      "step": 15020
    },
    {
      "epoch": 7.8747379454926625,
      "grad_norm": 0.3097524046897888,
      "learning_rate": 3.7972263923693476e-05,
      "loss": 0.6121,
      "num_input_tokens_seen": 9824896,
      "step": 15025
    },
    {
      "epoch": 7.877358490566038,
      "grad_norm": 0.5049062967300415,
      "learning_rate": 3.796248807529145e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 9828000,
      "step": 15030
    },
    {
      "epoch": 7.879979035639413,
      "grad_norm": 0.7189100384712219,
      "learning_rate": 3.795270951528482e-05,
      "loss": 0.544,
      "num_input_tokens_seen": 9830496,
      "step": 15035
    },
    {
      "epoch": 7.882599580712788,
      "grad_norm": 0.4059467613697052,
      "learning_rate": 3.7942928245719124e-05,
      "loss": 0.4792,
      "num_input_tokens_seen": 9833856,
      "step": 15040
    },
    {
      "epoch": 7.885220125786163,
      "grad_norm": 0.7589930295944214,
      "learning_rate": 3.7933144268640495e-05,
      "loss": 0.5343,
      "num_input_tokens_seen": 9836544,
      "step": 15045
    },
    {
      "epoch": 7.887840670859539,
      "grad_norm": 0.5447065830230713,
      "learning_rate": 3.7923357586095645e-05,
      "loss": 0.4711,
      "num_input_tokens_seen": 9839264,
      "step": 15050
    },
    {
      "epoch": 7.890461215932914,
      "grad_norm": 0.2969032824039459,
      "learning_rate": 3.7913568200131823e-05,
      "loss": 0.4513,
      "num_input_tokens_seen": 9845216,
      "step": 15055
    },
    {
      "epoch": 7.8930817610062896,
      "grad_norm": 0.3699047863483429,
      "learning_rate": 3.790377611279685e-05,
      "loss": 0.6641,
      "num_input_tokens_seen": 9848256,
      "step": 15060
    },
    {
      "epoch": 7.895702306079665,
      "grad_norm": 0.3108087182044983,
      "learning_rate": 3.789398132613914e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 9851008,
      "step": 15065
    },
    {
      "epoch": 7.89832285115304,
      "grad_norm": 0.5659593343734741,
      "learning_rate": 3.788418384220763e-05,
      "loss": 0.495,
      "num_input_tokens_seen": 9853920,
      "step": 15070
    },
    {
      "epoch": 7.900943396226415,
      "grad_norm": 0.3509724736213684,
      "learning_rate": 3.7874383663051846e-05,
      "loss": 0.4886,
      "num_input_tokens_seen": 9856736,
      "step": 15075
    },
    {
      "epoch": 7.90356394129979,
      "grad_norm": 0.5938308835029602,
      "learning_rate": 3.7864580790721894e-05,
      "loss": 0.3951,
      "num_input_tokens_seen": 9860896,
      "step": 15080
    },
    {
      "epoch": 7.906184486373165,
      "grad_norm": 0.41613513231277466,
      "learning_rate": 3.785477522726839e-05,
      "loss": 0.4403,
      "num_input_tokens_seen": 9867296,
      "step": 15085
    },
    {
      "epoch": 7.908805031446541,
      "grad_norm": 0.5054610967636108,
      "learning_rate": 3.784496697474258e-05,
      "loss": 0.6458,
      "num_input_tokens_seen": 9870816,
      "step": 15090
    },
    {
      "epoch": 7.911425576519916,
      "grad_norm": 0.32375603914260864,
      "learning_rate": 3.783515603519622e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 9873792,
      "step": 15095
    },
    {
      "epoch": 7.914046121593291,
      "grad_norm": 0.42286768555641174,
      "learning_rate": 3.782534241068164e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 9876544,
      "step": 15100
    },
    {
      "epoch": 7.916666666666667,
      "grad_norm": 0.40415143966674805,
      "learning_rate": 3.781552610325175e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 9880032,
      "step": 15105
    },
    {
      "epoch": 7.919287211740042,
      "grad_norm": 0.7411772608757019,
      "learning_rate": 3.780570711496001e-05,
      "loss": 0.4746,
      "num_input_tokens_seen": 9883296,
      "step": 15110
    },
    {
      "epoch": 7.921907756813417,
      "grad_norm": 0.3113524615764618,
      "learning_rate": 3.779588544786044e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 9885824,
      "step": 15115
    },
    {
      "epoch": 7.9245283018867925,
      "grad_norm": 0.44323182106018066,
      "learning_rate": 3.778606110400761e-05,
      "loss": 0.5096,
      "num_input_tokens_seen": 9888896,
      "step": 15120
    },
    {
      "epoch": 7.927148846960168,
      "grad_norm": 0.4182837903499603,
      "learning_rate": 3.7776234085456665e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 9891552,
      "step": 15125
    },
    {
      "epoch": 7.929769392033543,
      "grad_norm": 0.8844102621078491,
      "learning_rate": 3.776640439426332e-05,
      "loss": 0.4596,
      "num_input_tokens_seen": 9894400,
      "step": 15130
    },
    {
      "epoch": 7.932389937106918,
      "grad_norm": 0.38876399397850037,
      "learning_rate": 3.775657203248381e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 9897728,
      "step": 15135
    },
    {
      "epoch": 7.935010482180293,
      "grad_norm": 0.4399588704109192,
      "learning_rate": 3.774673700217497e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 9901248,
      "step": 15140
    },
    {
      "epoch": 7.937631027253669,
      "grad_norm": 0.39605703949928284,
      "learning_rate": 3.773689930539418e-05,
      "loss": 0.758,
      "num_input_tokens_seen": 9904384,
      "step": 15145
    },
    {
      "epoch": 7.940251572327044,
      "grad_norm": 0.5096534490585327,
      "learning_rate": 3.7727058944199354e-05,
      "loss": 0.4097,
      "num_input_tokens_seen": 9907872,
      "step": 15150
    },
    {
      "epoch": 7.9428721174004195,
      "grad_norm": 0.3102307915687561,
      "learning_rate": 3.7717215920648994e-05,
      "loss": 0.3246,
      "num_input_tokens_seen": 9911040,
      "step": 15155
    },
    {
      "epoch": 7.945492662473795,
      "grad_norm": 0.4284020662307739,
      "learning_rate": 3.770737023680215e-05,
      "loss": 0.4388,
      "num_input_tokens_seen": 9914208,
      "step": 15160
    },
    {
      "epoch": 7.94811320754717,
      "grad_norm": 0.4255194365978241,
      "learning_rate": 3.7697521894718436e-05,
      "loss": 0.544,
      "num_input_tokens_seen": 9917856,
      "step": 15165
    },
    {
      "epoch": 7.950733752620545,
      "grad_norm": 0.6484808921813965,
      "learning_rate": 3.7687670896458e-05,
      "loss": 0.493,
      "num_input_tokens_seen": 9921280,
      "step": 15170
    },
    {
      "epoch": 7.95335429769392,
      "grad_norm": 0.4381152391433716,
      "learning_rate": 3.7677817244081556e-05,
      "loss": 0.421,
      "num_input_tokens_seen": 9925376,
      "step": 15175
    },
    {
      "epoch": 7.955974842767295,
      "grad_norm": 0.310894250869751,
      "learning_rate": 3.766796093965038e-05,
      "loss": 0.5143,
      "num_input_tokens_seen": 9928256,
      "step": 15180
    },
    {
      "epoch": 7.9585953878406706,
      "grad_norm": 0.6259674429893494,
      "learning_rate": 3.76581019852263e-05,
      "loss": 0.5591,
      "num_input_tokens_seen": 9931232,
      "step": 15185
    },
    {
      "epoch": 7.961215932914046,
      "grad_norm": 0.5306234955787659,
      "learning_rate": 3.7648240382871705e-05,
      "loss": 0.4252,
      "num_input_tokens_seen": 9934368,
      "step": 15190
    },
    {
      "epoch": 7.963836477987421,
      "grad_norm": 0.3953019082546234,
      "learning_rate": 3.763837613464951e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 9938432,
      "step": 15195
    },
    {
      "epoch": 7.966457023060797,
      "grad_norm": 0.5728835463523865,
      "learning_rate": 3.7628509242623215e-05,
      "loss": 0.3237,
      "num_input_tokens_seen": 9941024,
      "step": 15200
    },
    {
      "epoch": 7.969077568134172,
      "grad_norm": 0.5000131130218506,
      "learning_rate": 3.761863970885686e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 9944480,
      "step": 15205
    },
    {
      "epoch": 7.971698113207547,
      "grad_norm": 0.4453648030757904,
      "learning_rate": 3.760876753541504e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 9948192,
      "step": 15210
    },
    {
      "epoch": 7.9743186582809225,
      "grad_norm": 0.5264340043067932,
      "learning_rate": 3.7598892724362907e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 9951168,
      "step": 15215
    },
    {
      "epoch": 7.976939203354298,
      "grad_norm": 0.5393633842468262,
      "learning_rate": 3.758901527776615e-05,
      "loss": 0.5428,
      "num_input_tokens_seen": 9955136,
      "step": 15220
    },
    {
      "epoch": 7.979559748427673,
      "grad_norm": 0.38639047741889954,
      "learning_rate": 3.7579135197691015e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 9958464,
      "step": 15225
    },
    {
      "epoch": 7.982180293501048,
      "grad_norm": 0.47394540905952454,
      "learning_rate": 3.7569252486204306e-05,
      "loss": 0.515,
      "num_input_tokens_seen": 9961312,
      "step": 15230
    },
    {
      "epoch": 7.984800838574423,
      "grad_norm": 0.264628142118454,
      "learning_rate": 3.755936714537338e-05,
      "loss": 0.4374,
      "num_input_tokens_seen": 9965280,
      "step": 15235
    },
    {
      "epoch": 7.987421383647799,
      "grad_norm": 0.4335402846336365,
      "learning_rate": 3.754947917726612e-05,
      "loss": 0.6199,
      "num_input_tokens_seen": 9969504,
      "step": 15240
    },
    {
      "epoch": 7.990041928721174,
      "grad_norm": 0.36534759402275085,
      "learning_rate": 3.7539588583951e-05,
      "loss": 0.3629,
      "num_input_tokens_seen": 9972960,
      "step": 15245
    },
    {
      "epoch": 7.9926624737945495,
      "grad_norm": 0.527406632900238,
      "learning_rate": 3.752969536749701e-05,
      "loss": 0.5229,
      "num_input_tokens_seen": 9975712,
      "step": 15250
    },
    {
      "epoch": 7.995283018867925,
      "grad_norm": 0.32895973324775696,
      "learning_rate": 3.7519799529973685e-05,
      "loss": 0.5186,
      "num_input_tokens_seen": 9978752,
      "step": 15255
    },
    {
      "epoch": 7.9979035639413,
      "grad_norm": 0.4596031606197357,
      "learning_rate": 3.7509901073451134e-05,
      "loss": 0.6294,
      "num_input_tokens_seen": 9981472,
      "step": 15260
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.45058733224868774,
      "eval_runtime": 9.6296,
      "eval_samples_per_second": 88.062,
      "eval_steps_per_second": 22.016,
      "num_input_tokens_seen": 9983720,
      "step": 15264
    },
    {
      "epoch": 8.000524109014675,
      "grad_norm": 0.4102574288845062,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.3894,
      "num_input_tokens_seen": 9984424,
      "step": 15265
    },
    {
      "epoch": 8.00314465408805,
      "grad_norm": 0.6044232845306396,
      "learning_rate": 3.749009631169147e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 9987528,
      "step": 15270
    },
    {
      "epoch": 8.005765199161425,
      "grad_norm": 0.3456215560436249,
      "learning_rate": 3.74801900105973e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 9990536,
      "step": 15275
    },
    {
      "epoch": 8.0083857442348,
      "grad_norm": 0.4148038327693939,
      "learning_rate": 3.747028109878975e-05,
      "loss": 0.3645,
      "num_input_tokens_seen": 9994504,
      "step": 15280
    },
    {
      "epoch": 8.011006289308176,
      "grad_norm": 0.484381765127182,
      "learning_rate": 3.746036957834167e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 9997480,
      "step": 15285
    },
    {
      "epoch": 8.01362683438155,
      "grad_norm": 0.4147130846977234,
      "learning_rate": 3.7450455451326415e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 10000840,
      "step": 15290
    },
    {
      "epoch": 8.016247379454926,
      "grad_norm": 0.5358365178108215,
      "learning_rate": 3.7440538719817935e-05,
      "loss": 0.459,
      "num_input_tokens_seen": 10003304,
      "step": 15295
    },
    {
      "epoch": 8.018867924528301,
      "grad_norm": 0.4914815127849579,
      "learning_rate": 3.743061938589068e-05,
      "loss": 0.5139,
      "num_input_tokens_seen": 10006280,
      "step": 15300
    },
    {
      "epoch": 8.021488469601676,
      "grad_norm": 0.6807121634483337,
      "learning_rate": 3.742069745161966e-05,
      "loss": 0.4858,
      "num_input_tokens_seen": 10009800,
      "step": 15305
    },
    {
      "epoch": 8.024109014675052,
      "grad_norm": 0.49686160683631897,
      "learning_rate": 3.741077291908045e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 10012840,
      "step": 15310
    },
    {
      "epoch": 8.026729559748428,
      "grad_norm": 0.35047096014022827,
      "learning_rate": 3.740084579034913e-05,
      "loss": 0.3883,
      "num_input_tokens_seen": 10015656,
      "step": 15315
    },
    {
      "epoch": 8.029350104821804,
      "grad_norm": 0.5698813796043396,
      "learning_rate": 3.739091606750234e-05,
      "loss": 0.4424,
      "num_input_tokens_seen": 10018888,
      "step": 15320
    },
    {
      "epoch": 8.031970649895179,
      "grad_norm": 0.3141813576221466,
      "learning_rate": 3.7380983752617275e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 10022280,
      "step": 15325
    },
    {
      "epoch": 8.034591194968554,
      "grad_norm": 0.7417407035827637,
      "learning_rate": 3.737104884777166e-05,
      "loss": 0.3896,
      "num_input_tokens_seen": 10025416,
      "step": 15330
    },
    {
      "epoch": 8.03721174004193,
      "grad_norm": 0.8704700469970703,
      "learning_rate": 3.736111135504376e-05,
      "loss": 0.5414,
      "num_input_tokens_seen": 10027944,
      "step": 15335
    },
    {
      "epoch": 8.039832285115304,
      "grad_norm": 0.25301894545555115,
      "learning_rate": 3.735117127651239e-05,
      "loss": 0.4281,
      "num_input_tokens_seen": 10030824,
      "step": 15340
    },
    {
      "epoch": 8.04245283018868,
      "grad_norm": 0.5671461224555969,
      "learning_rate": 3.734122861425689e-05,
      "loss": 0.3697,
      "num_input_tokens_seen": 10033128,
      "step": 15345
    },
    {
      "epoch": 8.045073375262055,
      "grad_norm": 0.4581946134567261,
      "learning_rate": 3.7331283370357173e-05,
      "loss": 0.4132,
      "num_input_tokens_seen": 10036488,
      "step": 15350
    },
    {
      "epoch": 8.04769392033543,
      "grad_norm": 0.6058677434921265,
      "learning_rate": 3.732133554689365e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 10039368,
      "step": 15355
    },
    {
      "epoch": 8.050314465408805,
      "grad_norm": 0.5478532314300537,
      "learning_rate": 3.73113851459473e-05,
      "loss": 0.4744,
      "num_input_tokens_seen": 10042472,
      "step": 15360
    },
    {
      "epoch": 8.05293501048218,
      "grad_norm": 0.4914112687110901,
      "learning_rate": 3.730143216959964e-05,
      "loss": 0.386,
      "num_input_tokens_seen": 10045640,
      "step": 15365
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 0.35733768343925476,
      "learning_rate": 3.729147661993271e-05,
      "loss": 0.4463,
      "num_input_tokens_seen": 10049864,
      "step": 15370
    },
    {
      "epoch": 8.05817610062893,
      "grad_norm": 0.46794724464416504,
      "learning_rate": 3.728151849902909e-05,
      "loss": 0.5804,
      "num_input_tokens_seen": 10053480,
      "step": 15375
    },
    {
      "epoch": 8.060796645702306,
      "grad_norm": 0.4041869640350342,
      "learning_rate": 3.727155780897193e-05,
      "loss": 0.5015,
      "num_input_tokens_seen": 10056872,
      "step": 15380
    },
    {
      "epoch": 8.06341719077568,
      "grad_norm": 0.3050481081008911,
      "learning_rate": 3.7261594551844865e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 10060264,
      "step": 15385
    },
    {
      "epoch": 8.066037735849056,
      "grad_norm": 0.5430972576141357,
      "learning_rate": 3.7251628729732116e-05,
      "loss": 0.3174,
      "num_input_tokens_seen": 10063240,
      "step": 15390
    },
    {
      "epoch": 8.068658280922431,
      "grad_norm": 0.5317696332931519,
      "learning_rate": 3.724166034471841e-05,
      "loss": 0.4121,
      "num_input_tokens_seen": 10065768,
      "step": 15395
    },
    {
      "epoch": 8.071278825995806,
      "grad_norm": 0.5074896216392517,
      "learning_rate": 3.723168939888901e-05,
      "loss": 0.4751,
      "num_input_tokens_seen": 10068872,
      "step": 15400
    },
    {
      "epoch": 8.073899371069182,
      "grad_norm": 0.430887907743454,
      "learning_rate": 3.7221715894329747e-05,
      "loss": 0.5176,
      "num_input_tokens_seen": 10071496,
      "step": 15405
    },
    {
      "epoch": 8.076519916142558,
      "grad_norm": 0.372949481010437,
      "learning_rate": 3.7211739833126945e-05,
      "loss": 0.6309,
      "num_input_tokens_seen": 10074632,
      "step": 15410
    },
    {
      "epoch": 8.079140461215934,
      "grad_norm": 0.8977055549621582,
      "learning_rate": 3.720176121736748e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 10077064,
      "step": 15415
    },
    {
      "epoch": 8.081761006289309,
      "grad_norm": 0.4378196895122528,
      "learning_rate": 3.719178004913879e-05,
      "loss": 0.4353,
      "num_input_tokens_seen": 10081288,
      "step": 15420
    },
    {
      "epoch": 8.084381551362684,
      "grad_norm": 0.6002017259597778,
      "learning_rate": 3.7181796330528796e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 10084392,
      "step": 15425
    },
    {
      "epoch": 8.08700209643606,
      "grad_norm": 0.4610254466533661,
      "learning_rate": 3.717181006362599e-05,
      "loss": 0.3083,
      "num_input_tokens_seen": 10088936,
      "step": 15430
    },
    {
      "epoch": 8.089622641509434,
      "grad_norm": 0.43188294768333435,
      "learning_rate": 3.7161821250519384e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 10091432,
      "step": 15435
    },
    {
      "epoch": 8.09224318658281,
      "grad_norm": 0.49490219354629517,
      "learning_rate": 3.7151829893298515e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 10093704,
      "step": 15440
    },
    {
      "epoch": 8.094863731656185,
      "grad_norm": 0.38820087909698486,
      "learning_rate": 3.714183599405347e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 10098120,
      "step": 15445
    },
    {
      "epoch": 8.09748427672956,
      "grad_norm": 0.4489736557006836,
      "learning_rate": 3.713183955487487e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 10100872,
      "step": 15450
    },
    {
      "epoch": 8.100104821802935,
      "grad_norm": 0.3896796703338623,
      "learning_rate": 3.7121840577853815e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 10104456,
      "step": 15455
    },
    {
      "epoch": 8.10272536687631,
      "grad_norm": 0.7736926674842834,
      "learning_rate": 3.7111839065082027e-05,
      "loss": 0.4448,
      "num_input_tokens_seen": 10107368,
      "step": 15460
    },
    {
      "epoch": 8.105345911949685,
      "grad_norm": 0.6361538171768188,
      "learning_rate": 3.710183501865168e-05,
      "loss": 0.4656,
      "num_input_tokens_seen": 10110344,
      "step": 15465
    },
    {
      "epoch": 8.10796645702306,
      "grad_norm": 0.42484840750694275,
      "learning_rate": 3.7091828440655514e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 10113224,
      "step": 15470
    },
    {
      "epoch": 8.110587002096436,
      "grad_norm": 0.6591414213180542,
      "learning_rate": 3.708181933318678e-05,
      "loss": 0.6395,
      "num_input_tokens_seen": 10117160,
      "step": 15475
    },
    {
      "epoch": 8.11320754716981,
      "grad_norm": 0.5240538120269775,
      "learning_rate": 3.707180769833929e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 10121064,
      "step": 15480
    },
    {
      "epoch": 8.115828092243186,
      "grad_norm": 0.4717348515987396,
      "learning_rate": 3.706179353820734e-05,
      "loss": 0.4856,
      "num_input_tokens_seen": 10124360,
      "step": 15485
    },
    {
      "epoch": 8.118448637316561,
      "grad_norm": 0.45081061124801636,
      "learning_rate": 3.7051776854885794e-05,
      "loss": 0.5393,
      "num_input_tokens_seen": 10127400,
      "step": 15490
    },
    {
      "epoch": 8.121069182389936,
      "grad_norm": 0.5917088389396667,
      "learning_rate": 3.7041757650470025e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 10129960,
      "step": 15495
    },
    {
      "epoch": 8.123689727463312,
      "grad_norm": 0.41502776741981506,
      "learning_rate": 3.703173592705594e-05,
      "loss": 0.4299,
      "num_input_tokens_seen": 10132936,
      "step": 15500
    },
    {
      "epoch": 8.126310272536688,
      "grad_norm": 0.22779403626918793,
      "learning_rate": 3.702171168673996e-05,
      "loss": 0.4259,
      "num_input_tokens_seen": 10136328,
      "step": 15505
    },
    {
      "epoch": 8.128930817610064,
      "grad_norm": 0.3430510461330414,
      "learning_rate": 3.701168493161905e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 10139688,
      "step": 15510
    },
    {
      "epoch": 8.131551362683439,
      "grad_norm": 0.3391459584236145,
      "learning_rate": 3.7001655663790676e-05,
      "loss": 0.4062,
      "num_input_tokens_seen": 10143016,
      "step": 15515
    },
    {
      "epoch": 8.134171907756814,
      "grad_norm": 0.27972903847694397,
      "learning_rate": 3.699162388535288e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 10147336,
      "step": 15520
    },
    {
      "epoch": 8.13679245283019,
      "grad_norm": 0.4732416272163391,
      "learning_rate": 3.698158959840416e-05,
      "loss": 0.3734,
      "num_input_tokens_seen": 10150504,
      "step": 15525
    },
    {
      "epoch": 8.139412997903564,
      "grad_norm": 0.5614420175552368,
      "learning_rate": 3.6971552805043585e-05,
      "loss": 0.5895,
      "num_input_tokens_seen": 10153000,
      "step": 15530
    },
    {
      "epoch": 8.14203354297694,
      "grad_norm": 0.994845986366272,
      "learning_rate": 3.6961513507370754e-05,
      "loss": 0.4816,
      "num_input_tokens_seen": 10155784,
      "step": 15535
    },
    {
      "epoch": 8.144654088050315,
      "grad_norm": 0.30322644114494324,
      "learning_rate": 3.695147170748576e-05,
      "loss": 0.928,
      "num_input_tokens_seen": 10159592,
      "step": 15540
    },
    {
      "epoch": 8.14727463312369,
      "grad_norm": 0.48544812202453613,
      "learning_rate": 3.694142740748923e-05,
      "loss": 0.3993,
      "num_input_tokens_seen": 10162600,
      "step": 15545
    },
    {
      "epoch": 8.149895178197065,
      "grad_norm": 0.8403131365776062,
      "learning_rate": 3.693138060948232e-05,
      "loss": 0.6202,
      "num_input_tokens_seen": 10164968,
      "step": 15550
    },
    {
      "epoch": 8.15251572327044,
      "grad_norm": 0.5230485796928406,
      "learning_rate": 3.692133131556671e-05,
      "loss": 0.643,
      "num_input_tokens_seen": 10171912,
      "step": 15555
    },
    {
      "epoch": 8.155136268343815,
      "grad_norm": 0.6404860019683838,
      "learning_rate": 3.6911279527844596e-05,
      "loss": 0.56,
      "num_input_tokens_seen": 10175112,
      "step": 15560
    },
    {
      "epoch": 8.15775681341719,
      "grad_norm": 0.5605084896087646,
      "learning_rate": 3.69012252484187e-05,
      "loss": 0.5529,
      "num_input_tokens_seen": 10178280,
      "step": 15565
    },
    {
      "epoch": 8.160377358490566,
      "grad_norm": 0.30007389187812805,
      "learning_rate": 3.6891168479392245e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 10181352,
      "step": 15570
    },
    {
      "epoch": 8.16299790356394,
      "grad_norm": 0.40250882506370544,
      "learning_rate": 3.688110922286901e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 10185256,
      "step": 15575
    },
    {
      "epoch": 8.165618448637316,
      "grad_norm": 0.2958693504333496,
      "learning_rate": 3.687104748095328e-05,
      "loss": 0.568,
      "num_input_tokens_seen": 10188872,
      "step": 15580
    },
    {
      "epoch": 8.168238993710691,
      "grad_norm": 0.35106298327445984,
      "learning_rate": 3.686098325574984e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 10192008,
      "step": 15585
    },
    {
      "epoch": 8.170859538784066,
      "grad_norm": 0.3334946930408478,
      "learning_rate": 3.685091654936403e-05,
      "loss": 0.3636,
      "num_input_tokens_seen": 10196808,
      "step": 15590
    },
    {
      "epoch": 8.173480083857442,
      "grad_norm": 0.41473910212516785,
      "learning_rate": 3.6840847363901675e-05,
      "loss": 0.5891,
      "num_input_tokens_seen": 10199848,
      "step": 15595
    },
    {
      "epoch": 8.176100628930818,
      "grad_norm": 0.33791494369506836,
      "learning_rate": 3.683077570146914e-05,
      "loss": 0.3089,
      "num_input_tokens_seen": 10203048,
      "step": 15600
    },
    {
      "epoch": 8.178721174004194,
      "grad_norm": 0.45799174904823303,
      "learning_rate": 3.68207015641733e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 10207144,
      "step": 15605
    },
    {
      "epoch": 8.181341719077569,
      "grad_norm": 0.729883074760437,
      "learning_rate": 3.681062495412155e-05,
      "loss": 0.4948,
      "num_input_tokens_seen": 10209672,
      "step": 15610
    },
    {
      "epoch": 8.183962264150944,
      "grad_norm": 0.545222818851471,
      "learning_rate": 3.68005458734218e-05,
      "loss": 0.4142,
      "num_input_tokens_seen": 10212520,
      "step": 15615
    },
    {
      "epoch": 8.18658280922432,
      "grad_norm": 0.3766210675239563,
      "learning_rate": 3.679046432418248e-05,
      "loss": 0.4185,
      "num_input_tokens_seen": 10215048,
      "step": 15620
    },
    {
      "epoch": 8.189203354297694,
      "grad_norm": 0.6982791423797607,
      "learning_rate": 3.6780380308512543e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 10217352,
      "step": 15625
    },
    {
      "epoch": 8.19182389937107,
      "grad_norm": 0.4727267026901245,
      "learning_rate": 3.677029382852144e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 10221096,
      "step": 15630
    },
    {
      "epoch": 8.194444444444445,
      "grad_norm": 0.2964255213737488,
      "learning_rate": 3.676020488631915e-05,
      "loss": 0.3962,
      "num_input_tokens_seen": 10229608,
      "step": 15635
    },
    {
      "epoch": 8.19706498951782,
      "grad_norm": 0.6113890409469604,
      "learning_rate": 3.675011348401617e-05,
      "loss": 0.3897,
      "num_input_tokens_seen": 10233000,
      "step": 15640
    },
    {
      "epoch": 8.199685534591195,
      "grad_norm": 0.38338181376457214,
      "learning_rate": 3.6740019623723494e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 10236328,
      "step": 15645
    },
    {
      "epoch": 8.20230607966457,
      "grad_norm": 0.40718090534210205,
      "learning_rate": 3.672992330755265e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 10239624,
      "step": 15650
    },
    {
      "epoch": 8.204926624737945,
      "grad_norm": 0.8130720853805542,
      "learning_rate": 3.671982453761568e-05,
      "loss": 0.5614,
      "num_input_tokens_seen": 10242440,
      "step": 15655
    },
    {
      "epoch": 8.20754716981132,
      "grad_norm": 0.29858481884002686,
      "learning_rate": 3.6709723316025114e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 10245608,
      "step": 15660
    },
    {
      "epoch": 8.210167714884696,
      "grad_norm": 0.5368508100509644,
      "learning_rate": 3.669961964489403e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 10247848,
      "step": 15665
    },
    {
      "epoch": 8.21278825995807,
      "grad_norm": 0.3550315797328949,
      "learning_rate": 3.668951352633599e-05,
      "loss": 0.4917,
      "num_input_tokens_seen": 10252616,
      "step": 15670
    },
    {
      "epoch": 8.215408805031446,
      "grad_norm": 1.032500982284546,
      "learning_rate": 3.6679404962465084e-05,
      "loss": 0.6216,
      "num_input_tokens_seen": 10255880,
      "step": 15675
    },
    {
      "epoch": 8.218029350104821,
      "grad_norm": 0.38319122791290283,
      "learning_rate": 3.66692939553959e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 10258632,
      "step": 15680
    },
    {
      "epoch": 8.220649895178196,
      "grad_norm": 0.4557751417160034,
      "learning_rate": 3.6659180507243554e-05,
      "loss": 0.4304,
      "num_input_tokens_seen": 10262504,
      "step": 15685
    },
    {
      "epoch": 8.223270440251572,
      "grad_norm": 0.31537115573883057,
      "learning_rate": 3.664906462012366e-05,
      "loss": 0.421,
      "num_input_tokens_seen": 10265480,
      "step": 15690
    },
    {
      "epoch": 8.225890985324948,
      "grad_norm": 0.5666506886482239,
      "learning_rate": 3.663894629615234e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 10269000,
      "step": 15695
    },
    {
      "epoch": 8.228511530398324,
      "grad_norm": 0.32405605912208557,
      "learning_rate": 3.6628825537446246e-05,
      "loss": 0.3597,
      "num_input_tokens_seen": 10272072,
      "step": 15700
    },
    {
      "epoch": 8.231132075471699,
      "grad_norm": 0.43619465827941895,
      "learning_rate": 3.6618702346122525e-05,
      "loss": 0.5605,
      "num_input_tokens_seen": 10276008,
      "step": 15705
    },
    {
      "epoch": 8.233752620545074,
      "grad_norm": 0.4537862539291382,
      "learning_rate": 3.6608576724298817e-05,
      "loss": 0.4386,
      "num_input_tokens_seen": 10279304,
      "step": 15710
    },
    {
      "epoch": 8.23637316561845,
      "grad_norm": 0.4014291763305664,
      "learning_rate": 3.65984486740933e-05,
      "loss": 0.564,
      "num_input_tokens_seen": 10282856,
      "step": 15715
    },
    {
      "epoch": 8.238993710691824,
      "grad_norm": 0.34907323122024536,
      "learning_rate": 3.6588318197624644e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 10286376,
      "step": 15720
    },
    {
      "epoch": 8.2416142557652,
      "grad_norm": 0.45926082134246826,
      "learning_rate": 3.6578185297012015e-05,
      "loss": 0.527,
      "num_input_tokens_seen": 10288904,
      "step": 15725
    },
    {
      "epoch": 8.244234800838575,
      "grad_norm": 0.6436542868614197,
      "learning_rate": 3.656804997437512e-05,
      "loss": 0.5195,
      "num_input_tokens_seen": 10294824,
      "step": 15730
    },
    {
      "epoch": 8.24685534591195,
      "grad_norm": 0.46487218141555786,
      "learning_rate": 3.6557912231834155e-05,
      "loss": 0.4459,
      "num_input_tokens_seen": 10299400,
      "step": 15735
    },
    {
      "epoch": 8.249475890985325,
      "grad_norm": 0.3232268691062927,
      "learning_rate": 3.65477720715098e-05,
      "loss": 0.3122,
      "num_input_tokens_seen": 10302312,
      "step": 15740
    },
    {
      "epoch": 8.2520964360587,
      "grad_norm": 0.8826063871383667,
      "learning_rate": 3.653762949552328e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 10305000,
      "step": 15745
    },
    {
      "epoch": 8.254716981132075,
      "grad_norm": 0.4746079742908478,
      "learning_rate": 3.6527484505996295e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 10308296,
      "step": 15750
    },
    {
      "epoch": 8.25733752620545,
      "grad_norm": 0.47632306814193726,
      "learning_rate": 3.651733710505106e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 10311464,
      "step": 15755
    },
    {
      "epoch": 8.259958071278826,
      "grad_norm": 0.3856002688407898,
      "learning_rate": 3.6507187294810305e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 10314920,
      "step": 15760
    },
    {
      "epoch": 8.2625786163522,
      "grad_norm": 0.4701390266418457,
      "learning_rate": 3.649703507739724e-05,
      "loss": 0.3793,
      "num_input_tokens_seen": 10317896,
      "step": 15765
    },
    {
      "epoch": 8.265199161425576,
      "grad_norm": 0.40128621459007263,
      "learning_rate": 3.6486880454935606e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 10320136,
      "step": 15770
    },
    {
      "epoch": 8.267819706498951,
      "grad_norm": 0.32002124190330505,
      "learning_rate": 3.647672342954963e-05,
      "loss": 0.3393,
      "num_input_tokens_seen": 10323496,
      "step": 15775
    },
    {
      "epoch": 8.270440251572326,
      "grad_norm": 0.39634817838668823,
      "learning_rate": 3.646656400336405e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 10327784,
      "step": 15780
    },
    {
      "epoch": 8.273060796645701,
      "grad_norm": 0.3738378584384918,
      "learning_rate": 3.6456402178504094e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 10330952,
      "step": 15785
    },
    {
      "epoch": 8.275681341719078,
      "grad_norm": 0.38138583302497864,
      "learning_rate": 3.6446237957095505e-05,
      "loss": 0.3935,
      "num_input_tokens_seen": 10334568,
      "step": 15790
    },
    {
      "epoch": 8.278301886792454,
      "grad_norm": 0.4585021734237671,
      "learning_rate": 3.643607134126452e-05,
      "loss": 0.3739,
      "num_input_tokens_seen": 10338184,
      "step": 15795
    },
    {
      "epoch": 8.280922431865829,
      "grad_norm": 0.46272021532058716,
      "learning_rate": 3.642590233313789e-05,
      "loss": 0.5873,
      "num_input_tokens_seen": 10341352,
      "step": 15800
    },
    {
      "epoch": 8.283542976939204,
      "grad_norm": 0.930478572845459,
      "learning_rate": 3.6415730934842827e-05,
      "loss": 0.546,
      "num_input_tokens_seen": 10344136,
      "step": 15805
    },
    {
      "epoch": 8.286163522012579,
      "grad_norm": 0.8687489628791809,
      "learning_rate": 3.64055571485071e-05,
      "loss": 0.4865,
      "num_input_tokens_seen": 10346472,
      "step": 15810
    },
    {
      "epoch": 8.288784067085954,
      "grad_norm": 0.8152505159378052,
      "learning_rate": 3.639538097625894e-05,
      "loss": 0.4381,
      "num_input_tokens_seen": 10349928,
      "step": 15815
    },
    {
      "epoch": 8.29140461215933,
      "grad_norm": 0.267179399728775,
      "learning_rate": 3.638520242022708e-05,
      "loss": 0.4937,
      "num_input_tokens_seen": 10352936,
      "step": 15820
    },
    {
      "epoch": 8.294025157232705,
      "grad_norm": 0.5191807150840759,
      "learning_rate": 3.637502148254077e-05,
      "loss": 0.3908,
      "num_input_tokens_seen": 10356040,
      "step": 15825
    },
    {
      "epoch": 8.29664570230608,
      "grad_norm": 0.5913297533988953,
      "learning_rate": 3.636483816532974e-05,
      "loss": 0.3828,
      "num_input_tokens_seen": 10359048,
      "step": 15830
    },
    {
      "epoch": 8.299266247379455,
      "grad_norm": 0.3967724144458771,
      "learning_rate": 3.6354652470724217e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 10361832,
      "step": 15835
    },
    {
      "epoch": 8.30188679245283,
      "grad_norm": 0.5413452386856079,
      "learning_rate": 3.634446440085494e-05,
      "loss": 0.5572,
      "num_input_tokens_seen": 10364904,
      "step": 15840
    },
    {
      "epoch": 8.304507337526205,
      "grad_norm": 0.8278183341026306,
      "learning_rate": 3.633427395785313e-05,
      "loss": 0.6095,
      "num_input_tokens_seen": 10367816,
      "step": 15845
    },
    {
      "epoch": 8.30712788259958,
      "grad_norm": 0.5289857983589172,
      "learning_rate": 3.6324081143850516e-05,
      "loss": 0.4733,
      "num_input_tokens_seen": 10371080,
      "step": 15850
    },
    {
      "epoch": 8.309748427672956,
      "grad_norm": 0.4104997515678406,
      "learning_rate": 3.6313885960979326e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 10373928,
      "step": 15855
    },
    {
      "epoch": 8.31236897274633,
      "grad_norm": 1.1074355840682983,
      "learning_rate": 3.630368841137226e-05,
      "loss": 0.5225,
      "num_input_tokens_seen": 10377384,
      "step": 15860
    },
    {
      "epoch": 8.314989517819706,
      "grad_norm": 0.9665669798851013,
      "learning_rate": 3.6293488497162554e-05,
      "loss": 0.6455,
      "num_input_tokens_seen": 10380392,
      "step": 15865
    },
    {
      "epoch": 8.317610062893081,
      "grad_norm": 0.5043712854385376,
      "learning_rate": 3.628328622048388e-05,
      "loss": 0.5481,
      "num_input_tokens_seen": 10382824,
      "step": 15870
    },
    {
      "epoch": 8.320230607966456,
      "grad_norm": 0.6211522221565247,
      "learning_rate": 3.627308158347046e-05,
      "loss": 0.432,
      "num_input_tokens_seen": 10385192,
      "step": 15875
    },
    {
      "epoch": 8.322851153039831,
      "grad_norm": 0.40923309326171875,
      "learning_rate": 3.626287458825699e-05,
      "loss": 0.4315,
      "num_input_tokens_seen": 10389256,
      "step": 15880
    },
    {
      "epoch": 8.325471698113208,
      "grad_norm": 0.5980260372161865,
      "learning_rate": 3.625266523697864e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 10392168,
      "step": 15885
    },
    {
      "epoch": 8.328092243186584,
      "grad_norm": 0.35660097002983093,
      "learning_rate": 3.62424535317711e-05,
      "loss": 0.3537,
      "num_input_tokens_seen": 10395240,
      "step": 15890
    },
    {
      "epoch": 8.330712788259959,
      "grad_norm": 0.3870498836040497,
      "learning_rate": 3.623223947477055e-05,
      "loss": 0.3426,
      "num_input_tokens_seen": 10398056,
      "step": 15895
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 0.4037272036075592,
      "learning_rate": 3.6222023068113635e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 10401608,
      "step": 15900
    },
    {
      "epoch": 8.335953878406709,
      "grad_norm": 0.44908368587493896,
      "learning_rate": 3.621180431393754e-05,
      "loss": 0.3731,
      "num_input_tokens_seen": 10404424,
      "step": 15905
    },
    {
      "epoch": 8.338574423480084,
      "grad_norm": 0.45442596077919006,
      "learning_rate": 3.6201583214379874e-05,
      "loss": 0.3468,
      "num_input_tokens_seen": 10407016,
      "step": 15910
    },
    {
      "epoch": 8.34119496855346,
      "grad_norm": 0.5314989686012268,
      "learning_rate": 3.61913597715788e-05,
      "loss": 0.5337,
      "num_input_tokens_seen": 10410280,
      "step": 15915
    },
    {
      "epoch": 8.343815513626835,
      "grad_norm": 0.44416263699531555,
      "learning_rate": 3.6181133987672944e-05,
      "loss": 0.5388,
      "num_input_tokens_seen": 10413000,
      "step": 15920
    },
    {
      "epoch": 8.34643605870021,
      "grad_norm": 0.8701305985450745,
      "learning_rate": 3.617090586480142e-05,
      "loss": 0.4377,
      "num_input_tokens_seen": 10415336,
      "step": 15925
    },
    {
      "epoch": 8.349056603773585,
      "grad_norm": 0.4818797707557678,
      "learning_rate": 3.6160675405103826e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 10418536,
      "step": 15930
    },
    {
      "epoch": 8.35167714884696,
      "grad_norm": 0.7727368474006653,
      "learning_rate": 3.615044261072027e-05,
      "loss": 0.3986,
      "num_input_tokens_seen": 10421320,
      "step": 15935
    },
    {
      "epoch": 8.354297693920335,
      "grad_norm": 0.39815860986709595,
      "learning_rate": 3.614020748379133e-05,
      "loss": 0.4677,
      "num_input_tokens_seen": 10425224,
      "step": 15940
    },
    {
      "epoch": 8.35691823899371,
      "grad_norm": 0.4344692826271057,
      "learning_rate": 3.6129970026458076e-05,
      "loss": 0.382,
      "num_input_tokens_seen": 10427912,
      "step": 15945
    },
    {
      "epoch": 8.359538784067086,
      "grad_norm": 0.5210294127464294,
      "learning_rate": 3.611973024086207e-05,
      "loss": 0.3654,
      "num_input_tokens_seen": 10430600,
      "step": 15950
    },
    {
      "epoch": 8.36215932914046,
      "grad_norm": 0.8253464102745056,
      "learning_rate": 3.610948812914536e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 10433992,
      "step": 15955
    },
    {
      "epoch": 8.364779874213836,
      "grad_norm": 0.30583998560905457,
      "learning_rate": 3.609924369345048e-05,
      "loss": 0.345,
      "num_input_tokens_seen": 10437832,
      "step": 15960
    },
    {
      "epoch": 8.367400419287211,
      "grad_norm": 0.6067686676979065,
      "learning_rate": 3.6088996935920444e-05,
      "loss": 0.403,
      "num_input_tokens_seen": 10440840,
      "step": 15965
    },
    {
      "epoch": 8.370020964360586,
      "grad_norm": 0.8107262849807739,
      "learning_rate": 3.607874785869876e-05,
      "loss": 0.5427,
      "num_input_tokens_seen": 10443496,
      "step": 15970
    },
    {
      "epoch": 8.372641509433961,
      "grad_norm": 0.4015785753726959,
      "learning_rate": 3.6068496463929416e-05,
      "loss": 0.5146,
      "num_input_tokens_seen": 10446184,
      "step": 15975
    },
    {
      "epoch": 8.375262054507338,
      "grad_norm": 0.5059352517127991,
      "learning_rate": 3.605824275375689e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 10449064,
      "step": 15980
    },
    {
      "epoch": 8.377882599580714,
      "grad_norm": 0.43131446838378906,
      "learning_rate": 3.604798673032613e-05,
      "loss": 0.4229,
      "num_input_tokens_seen": 10452808,
      "step": 15985
    },
    {
      "epoch": 8.380503144654089,
      "grad_norm": 0.43446171283721924,
      "learning_rate": 3.60377283957826e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 10456808,
      "step": 15990
    },
    {
      "epoch": 8.383123689727464,
      "grad_norm": 0.4503459334373474,
      "learning_rate": 3.6027467752272203e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 10459880,
      "step": 15995
    },
    {
      "epoch": 8.385744234800839,
      "grad_norm": 0.31576523184776306,
      "learning_rate": 3.601720480194137e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 10463208,
      "step": 16000
    },
    {
      "epoch": 8.388364779874214,
      "grad_norm": 0.4036617875099182,
      "learning_rate": 3.600693954693698e-05,
      "loss": 0.4735,
      "num_input_tokens_seen": 10467368,
      "step": 16005
    },
    {
      "epoch": 8.39098532494759,
      "grad_norm": 0.7416204810142517,
      "learning_rate": 3.5996671989406414e-05,
      "loss": 0.425,
      "num_input_tokens_seen": 10470344,
      "step": 16010
    },
    {
      "epoch": 8.393605870020965,
      "grad_norm": 0.44607460498809814,
      "learning_rate": 3.598640213149752e-05,
      "loss": 0.445,
      "num_input_tokens_seen": 10474312,
      "step": 16015
    },
    {
      "epoch": 8.39622641509434,
      "grad_norm": 0.4993949234485626,
      "learning_rate": 3.597612997535864e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 10477224,
      "step": 16020
    },
    {
      "epoch": 8.398846960167715,
      "grad_norm": 0.48996007442474365,
      "learning_rate": 3.596585552313859e-05,
      "loss": 0.4299,
      "num_input_tokens_seen": 10481800,
      "step": 16025
    },
    {
      "epoch": 8.40146750524109,
      "grad_norm": 0.37170225381851196,
      "learning_rate": 3.595557877698666e-05,
      "loss": 0.3882,
      "num_input_tokens_seen": 10484584,
      "step": 16030
    },
    {
      "epoch": 8.404088050314465,
      "grad_norm": 0.6280762553215027,
      "learning_rate": 3.594529973905266e-05,
      "loss": 0.392,
      "num_input_tokens_seen": 10487272,
      "step": 16035
    },
    {
      "epoch": 8.40670859538784,
      "grad_norm": 0.3268515467643738,
      "learning_rate": 3.593501841148681e-05,
      "loss": 0.3446,
      "num_input_tokens_seen": 10490856,
      "step": 16040
    },
    {
      "epoch": 8.409329140461216,
      "grad_norm": 0.6395310759544373,
      "learning_rate": 3.5924734796439864e-05,
      "loss": 0.5724,
      "num_input_tokens_seen": 10493800,
      "step": 16045
    },
    {
      "epoch": 8.41194968553459,
      "grad_norm": 0.5671816468238831,
      "learning_rate": 3.5914448896063044e-05,
      "loss": 0.5118,
      "num_input_tokens_seen": 10497512,
      "step": 16050
    },
    {
      "epoch": 8.414570230607966,
      "grad_norm": 0.22284935414791107,
      "learning_rate": 3.590416071250802e-05,
      "loss": 0.4444,
      "num_input_tokens_seen": 10501096,
      "step": 16055
    },
    {
      "epoch": 8.417190775681341,
      "grad_norm": 0.36503660678863525,
      "learning_rate": 3.589387024792699e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 10504648,
      "step": 16060
    },
    {
      "epoch": 8.419811320754716,
      "grad_norm": 0.4038386940956116,
      "learning_rate": 3.5883577504472576e-05,
      "loss": 0.2957,
      "num_input_tokens_seen": 10507016,
      "step": 16065
    },
    {
      "epoch": 8.422431865828091,
      "grad_norm": 0.1985575556755066,
      "learning_rate": 3.587328248429792e-05,
      "loss": 0.3779,
      "num_input_tokens_seen": 10510792,
      "step": 16070
    },
    {
      "epoch": 8.425052410901468,
      "grad_norm": 0.33574971556663513,
      "learning_rate": 3.586298518955661e-05,
      "loss": 0.5198,
      "num_input_tokens_seen": 10514696,
      "step": 16075
    },
    {
      "epoch": 8.427672955974844,
      "grad_norm": 0.2527278661727905,
      "learning_rate": 3.5852685622402746e-05,
      "loss": 0.3878,
      "num_input_tokens_seen": 10518152,
      "step": 16080
    },
    {
      "epoch": 8.430293501048219,
      "grad_norm": 0.3454640209674835,
      "learning_rate": 3.584238378499085e-05,
      "loss": 0.5371,
      "num_input_tokens_seen": 10521128,
      "step": 16085
    },
    {
      "epoch": 8.432914046121594,
      "grad_norm": 0.7762149572372437,
      "learning_rate": 3.5832079679475974e-05,
      "loss": 0.6962,
      "num_input_tokens_seen": 10524456,
      "step": 16090
    },
    {
      "epoch": 8.435534591194969,
      "grad_norm": 0.393350750207901,
      "learning_rate": 3.582177330801359e-05,
      "loss": 0.4261,
      "num_input_tokens_seen": 10526920,
      "step": 16095
    },
    {
      "epoch": 8.438155136268344,
      "grad_norm": 0.3422168791294098,
      "learning_rate": 3.58114646727597e-05,
      "loss": 0.4169,
      "num_input_tokens_seen": 10529640,
      "step": 16100
    },
    {
      "epoch": 8.44077568134172,
      "grad_norm": 0.6212261319160461,
      "learning_rate": 3.580115377587075e-05,
      "loss": 0.4187,
      "num_input_tokens_seen": 10532968,
      "step": 16105
    },
    {
      "epoch": 8.443396226415095,
      "grad_norm": 0.3616514205932617,
      "learning_rate": 3.5790840619503654e-05,
      "loss": 0.4828,
      "num_input_tokens_seen": 10535944,
      "step": 16110
    },
    {
      "epoch": 8.44601677148847,
      "grad_norm": 0.4303516149520874,
      "learning_rate": 3.57805252058158e-05,
      "loss": 0.3785,
      "num_input_tokens_seen": 10539304,
      "step": 16115
    },
    {
      "epoch": 8.448637316561845,
      "grad_norm": 0.44338756799697876,
      "learning_rate": 3.577020753696507e-05,
      "loss": 0.3737,
      "num_input_tokens_seen": 10542120,
      "step": 16120
    },
    {
      "epoch": 8.45125786163522,
      "grad_norm": 0.31841015815734863,
      "learning_rate": 3.57598876151098e-05,
      "loss": 0.4364,
      "num_input_tokens_seen": 10544392,
      "step": 16125
    },
    {
      "epoch": 8.453878406708595,
      "grad_norm": 0.5657811760902405,
      "learning_rate": 3.5749565442408796e-05,
      "loss": 0.497,
      "num_input_tokens_seen": 10547656,
      "step": 16130
    },
    {
      "epoch": 8.45649895178197,
      "grad_norm": 0.7439377307891846,
      "learning_rate": 3.573924102102133e-05,
      "loss": 0.4965,
      "num_input_tokens_seen": 10550984,
      "step": 16135
    },
    {
      "epoch": 8.459119496855346,
      "grad_norm": 0.4329819083213806,
      "learning_rate": 3.572891435310717e-05,
      "loss": 0.5907,
      "num_input_tokens_seen": 10554952,
      "step": 16140
    },
    {
      "epoch": 8.46174004192872,
      "grad_norm": 0.5751445889472961,
      "learning_rate": 3.5718585440826527e-05,
      "loss": 0.5285,
      "num_input_tokens_seen": 10557608,
      "step": 16145
    },
    {
      "epoch": 8.464360587002096,
      "grad_norm": 0.5001906752586365,
      "learning_rate": 3.570825428634009e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 10561352,
      "step": 16150
    },
    {
      "epoch": 8.466981132075471,
      "grad_norm": 0.5682220458984375,
      "learning_rate": 3.569792089180903e-05,
      "loss": 0.5084,
      "num_input_tokens_seen": 10565064,
      "step": 16155
    },
    {
      "epoch": 8.469601677148846,
      "grad_norm": 0.3831090033054352,
      "learning_rate": 3.568758525939497e-05,
      "loss": 0.3766,
      "num_input_tokens_seen": 10567976,
      "step": 16160
    },
    {
      "epoch": 8.472222222222221,
      "grad_norm": 0.40562328696250916,
      "learning_rate": 3.5677247391259995e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 10571528,
      "step": 16165
    },
    {
      "epoch": 8.474842767295598,
      "grad_norm": 0.49712520837783813,
      "learning_rate": 3.566690728956668e-05,
      "loss": 0.493,
      "num_input_tokens_seen": 10574024,
      "step": 16170
    },
    {
      "epoch": 8.477463312368974,
      "grad_norm": 0.4978523254394531,
      "learning_rate": 3.5656564956478056e-05,
      "loss": 0.523,
      "num_input_tokens_seen": 10578024,
      "step": 16175
    },
    {
      "epoch": 8.480083857442349,
      "grad_norm": 0.2833017408847809,
      "learning_rate": 3.5646220394157626e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 10581448,
      "step": 16180
    },
    {
      "epoch": 8.482704402515724,
      "grad_norm": 0.9808169007301331,
      "learning_rate": 3.563587360476933e-05,
      "loss": 0.4199,
      "num_input_tokens_seen": 10584328,
      "step": 16185
    },
    {
      "epoch": 8.485324947589099,
      "grad_norm": 0.5193940997123718,
      "learning_rate": 3.562552459047763e-05,
      "loss": 0.6024,
      "num_input_tokens_seen": 10587400,
      "step": 16190
    },
    {
      "epoch": 8.487945492662474,
      "grad_norm": 0.4594464600086212,
      "learning_rate": 3.561517335344741e-05,
      "loss": 0.5039,
      "num_input_tokens_seen": 10591368,
      "step": 16195
    },
    {
      "epoch": 8.49056603773585,
      "grad_norm": 0.3988182544708252,
      "learning_rate": 3.560481989584402e-05,
      "loss": 0.5256,
      "num_input_tokens_seen": 10594856,
      "step": 16200
    },
    {
      "epoch": 8.493186582809225,
      "grad_norm": 0.7401639819145203,
      "learning_rate": 3.55944642198333e-05,
      "loss": 0.5416,
      "num_input_tokens_seen": 10597928,
      "step": 16205
    },
    {
      "epoch": 8.4958071278826,
      "grad_norm": 0.27240243554115295,
      "learning_rate": 3.5584106327581536e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 10601448,
      "step": 16210
    },
    {
      "epoch": 8.498427672955975,
      "grad_norm": 0.5946344137191772,
      "learning_rate": 3.557374622125547e-05,
      "loss": 0.4475,
      "num_input_tokens_seen": 10604520,
      "step": 16215
    },
    {
      "epoch": 8.50104821802935,
      "grad_norm": 0.344599187374115,
      "learning_rate": 3.556338390302234e-05,
      "loss": 0.5218,
      "num_input_tokens_seen": 10607944,
      "step": 16220
    },
    {
      "epoch": 8.503668763102725,
      "grad_norm": 0.45586463809013367,
      "learning_rate": 3.55530193750498e-05,
      "loss": 0.4937,
      "num_input_tokens_seen": 10611208,
      "step": 16225
    },
    {
      "epoch": 8.5062893081761,
      "grad_norm": 0.5375574827194214,
      "learning_rate": 3.5542652639505996e-05,
      "loss": 0.5233,
      "num_input_tokens_seen": 10613896,
      "step": 16230
    },
    {
      "epoch": 8.508909853249476,
      "grad_norm": 0.6975520849227905,
      "learning_rate": 3.5532283698559556e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 10616392,
      "step": 16235
    },
    {
      "epoch": 8.51153039832285,
      "grad_norm": 0.3411480784416199,
      "learning_rate": 3.55219125543795e-05,
      "loss": 0.5507,
      "num_input_tokens_seen": 10620232,
      "step": 16240
    },
    {
      "epoch": 8.514150943396226,
      "grad_norm": 0.4980601370334625,
      "learning_rate": 3.5511539209135396e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 10627272,
      "step": 16245
    },
    {
      "epoch": 8.516771488469601,
      "grad_norm": 0.425078809261322,
      "learning_rate": 3.550116366499721e-05,
      "loss": 0.5102,
      "num_input_tokens_seen": 10630344,
      "step": 16250
    },
    {
      "epoch": 8.519392033542976,
      "grad_norm": 0.5866686105728149,
      "learning_rate": 3.549078592413538e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 10633480,
      "step": 16255
    },
    {
      "epoch": 8.522012578616351,
      "grad_norm": 0.7272480130195618,
      "learning_rate": 3.548040598872083e-05,
      "loss": 0.5095,
      "num_input_tokens_seen": 10636104,
      "step": 16260
    },
    {
      "epoch": 8.524633123689728,
      "grad_norm": 0.4706535339355469,
      "learning_rate": 3.5470023860924905e-05,
      "loss": 0.4044,
      "num_input_tokens_seen": 10638696,
      "step": 16265
    },
    {
      "epoch": 8.527253668763104,
      "grad_norm": 0.3483477532863617,
      "learning_rate": 3.5459639542919444e-05,
      "loss": 0.4706,
      "num_input_tokens_seen": 10641320,
      "step": 16270
    },
    {
      "epoch": 8.529874213836479,
      "grad_norm": 0.28997746109962463,
      "learning_rate": 3.5449253036876715e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 10644360,
      "step": 16275
    },
    {
      "epoch": 8.532494758909854,
      "grad_norm": 0.5766061544418335,
      "learning_rate": 3.543886434496947e-05,
      "loss": 0.553,
      "num_input_tokens_seen": 10647464,
      "step": 16280
    },
    {
      "epoch": 8.535115303983229,
      "grad_norm": 0.6198086142539978,
      "learning_rate": 3.542847346937089e-05,
      "loss": 0.4419,
      "num_input_tokens_seen": 10650920,
      "step": 16285
    },
    {
      "epoch": 8.537735849056604,
      "grad_norm": 0.6053262948989868,
      "learning_rate": 3.541808041225464e-05,
      "loss": 0.4687,
      "num_input_tokens_seen": 10654984,
      "step": 16290
    },
    {
      "epoch": 8.54035639412998,
      "grad_norm": 0.5459084510803223,
      "learning_rate": 3.540768517579481e-05,
      "loss": 0.3795,
      "num_input_tokens_seen": 10658216,
      "step": 16295
    },
    {
      "epoch": 8.542976939203355,
      "grad_norm": 0.43174299597740173,
      "learning_rate": 3.539728776216599e-05,
      "loss": 0.546,
      "num_input_tokens_seen": 10660904,
      "step": 16300
    },
    {
      "epoch": 8.54559748427673,
      "grad_norm": 0.4538331627845764,
      "learning_rate": 3.5386888173543185e-05,
      "loss": 0.4629,
      "num_input_tokens_seen": 10664040,
      "step": 16305
    },
    {
      "epoch": 8.548218029350105,
      "grad_norm": 0.427482545375824,
      "learning_rate": 3.5376486412101875e-05,
      "loss": 0.5574,
      "num_input_tokens_seen": 10668008,
      "step": 16310
    },
    {
      "epoch": 8.55083857442348,
      "grad_norm": 0.29505953192710876,
      "learning_rate": 3.5366082480017994e-05,
      "loss": 0.4201,
      "num_input_tokens_seen": 10673032,
      "step": 16315
    },
    {
      "epoch": 8.553459119496855,
      "grad_norm": 0.8385735154151917,
      "learning_rate": 3.535567637946792e-05,
      "loss": 0.6352,
      "num_input_tokens_seen": 10676200,
      "step": 16320
    },
    {
      "epoch": 8.55607966457023,
      "grad_norm": 0.29937320947647095,
      "learning_rate": 3.534526811262848e-05,
      "loss": 0.4724,
      "num_input_tokens_seen": 10680168,
      "step": 16325
    },
    {
      "epoch": 8.558700209643606,
      "grad_norm": 0.45602938532829285,
      "learning_rate": 3.533485768167698e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 10683304,
      "step": 16330
    },
    {
      "epoch": 8.56132075471698,
      "grad_norm": 0.2926417589187622,
      "learning_rate": 3.532444508879117e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 10686856,
      "step": 16335
    },
    {
      "epoch": 8.563941299790356,
      "grad_norm": 0.5207971930503845,
      "learning_rate": 3.531403033614922e-05,
      "loss": 0.5046,
      "num_input_tokens_seen": 10691048,
      "step": 16340
    },
    {
      "epoch": 8.566561844863731,
      "grad_norm": 0.4768727421760559,
      "learning_rate": 3.530361342592981e-05,
      "loss": 0.4937,
      "num_input_tokens_seen": 10694504,
      "step": 16345
    },
    {
      "epoch": 8.569182389937106,
      "grad_norm": 0.4386007785797119,
      "learning_rate": 3.5293194360312006e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 10697576,
      "step": 16350
    },
    {
      "epoch": 8.571802935010481,
      "grad_norm": 0.6651846170425415,
      "learning_rate": 3.528277314147538e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 10700552,
      "step": 16355
    },
    {
      "epoch": 8.574423480083858,
      "grad_norm": 0.763926088809967,
      "learning_rate": 3.5272349771599925e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 10703240,
      "step": 16360
    },
    {
      "epoch": 8.577044025157234,
      "grad_norm": 1.832004427909851,
      "learning_rate": 3.526192425286609e-05,
      "loss": 0.474,
      "num_input_tokens_seen": 10705960,
      "step": 16365
    },
    {
      "epoch": 8.579664570230609,
      "grad_norm": 0.486131489276886,
      "learning_rate": 3.525149658745476e-05,
      "loss": 0.4104,
      "num_input_tokens_seen": 10709672,
      "step": 16370
    },
    {
      "epoch": 8.582285115303984,
      "grad_norm": 0.4425395727157593,
      "learning_rate": 3.5241066777547315e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 10713096,
      "step": 16375
    },
    {
      "epoch": 8.584905660377359,
      "grad_norm": 0.4358510971069336,
      "learning_rate": 3.523063482532554e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 10716552,
      "step": 16380
    },
    {
      "epoch": 8.587526205450734,
      "grad_norm": 0.6785373687744141,
      "learning_rate": 3.5220200732971665e-05,
      "loss": 0.4078,
      "num_input_tokens_seen": 10719336,
      "step": 16385
    },
    {
      "epoch": 8.59014675052411,
      "grad_norm": 0.378128319978714,
      "learning_rate": 3.52097645026684e-05,
      "loss": 0.3358,
      "num_input_tokens_seen": 10723528,
      "step": 16390
    },
    {
      "epoch": 8.592767295597485,
      "grad_norm": 0.5446103811264038,
      "learning_rate": 3.519932613659889e-05,
      "loss": 0.3986,
      "num_input_tokens_seen": 10726248,
      "step": 16395
    },
    {
      "epoch": 8.59538784067086,
      "grad_norm": 0.4680326282978058,
      "learning_rate": 3.5188885636946696e-05,
      "loss": 0.4232,
      "num_input_tokens_seen": 10729000,
      "step": 16400
    },
    {
      "epoch": 8.598008385744235,
      "grad_norm": 0.21805702149868011,
      "learning_rate": 3.5178443005895884e-05,
      "loss": 0.35,
      "num_input_tokens_seen": 10732328,
      "step": 16405
    },
    {
      "epoch": 8.60062893081761,
      "grad_norm": 0.5072724223136902,
      "learning_rate": 3.51679982456309e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 10735112,
      "step": 16410
    },
    {
      "epoch": 8.603249475890985,
      "grad_norm": 0.3341515362262726,
      "learning_rate": 3.51575513583367e-05,
      "loss": 0.58,
      "num_input_tokens_seen": 10739016,
      "step": 16415
    },
    {
      "epoch": 8.60587002096436,
      "grad_norm": 0.6408006548881531,
      "learning_rate": 3.5147102346198635e-05,
      "loss": 0.4002,
      "num_input_tokens_seen": 10741992,
      "step": 16420
    },
    {
      "epoch": 8.608490566037736,
      "grad_norm": 0.3032618761062622,
      "learning_rate": 3.5136651211402515e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 10746120,
      "step": 16425
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 0.6771662831306458,
      "learning_rate": 3.5126197956134614e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 10748584,
      "step": 16430
    },
    {
      "epoch": 8.613731656184486,
      "grad_norm": 0.40291520953178406,
      "learning_rate": 3.5115742582581626e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 10751400,
      "step": 16435
    },
    {
      "epoch": 8.616352201257861,
      "grad_norm": 0.5005500912666321,
      "learning_rate": 3.51052850929307e-05,
      "loss": 0.409,
      "num_input_tokens_seen": 10754024,
      "step": 16440
    },
    {
      "epoch": 8.618972746331236,
      "grad_norm": 0.2935566008090973,
      "learning_rate": 3.509482548936941e-05,
      "loss": 0.3935,
      "num_input_tokens_seen": 10757608,
      "step": 16445
    },
    {
      "epoch": 8.621593291404611,
      "grad_norm": 0.579084038734436,
      "learning_rate": 3.5084363774085804e-05,
      "loss": 0.5594,
      "num_input_tokens_seen": 10760232,
      "step": 16450
    },
    {
      "epoch": 8.624213836477988,
      "grad_norm": 0.5674508810043335,
      "learning_rate": 3.507389994926834e-05,
      "loss": 0.4911,
      "num_input_tokens_seen": 10763656,
      "step": 16455
    },
    {
      "epoch": 8.626834381551364,
      "grad_norm": 0.31142890453338623,
      "learning_rate": 3.5063434017105946e-05,
      "loss": 0.3838,
      "num_input_tokens_seen": 10768328,
      "step": 16460
    },
    {
      "epoch": 8.629454926624739,
      "grad_norm": 0.5036408305168152,
      "learning_rate": 3.505296597978795e-05,
      "loss": 0.562,
      "num_input_tokens_seen": 10771112,
      "step": 16465
    },
    {
      "epoch": 8.632075471698114,
      "grad_norm": 0.32237333059310913,
      "learning_rate": 3.504249583950417e-05,
      "loss": 0.5439,
      "num_input_tokens_seen": 10774312,
      "step": 16470
    },
    {
      "epoch": 8.634696016771489,
      "grad_norm": 0.28484517335891724,
      "learning_rate": 3.503202359844483e-05,
      "loss": 0.3711,
      "num_input_tokens_seen": 10777160,
      "step": 16475
    },
    {
      "epoch": 8.637316561844864,
      "grad_norm": 0.40469977259635925,
      "learning_rate": 3.50215492588006e-05,
      "loss": 0.3824,
      "num_input_tokens_seen": 10781096,
      "step": 16480
    },
    {
      "epoch": 8.63993710691824,
      "grad_norm": 0.319193571805954,
      "learning_rate": 3.5011072822762605e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 10784264,
      "step": 16485
    },
    {
      "epoch": 8.642557651991615,
      "grad_norm": 0.6231058239936829,
      "learning_rate": 3.500059429252238e-05,
      "loss": 0.4144,
      "num_input_tokens_seen": 10786856,
      "step": 16490
    },
    {
      "epoch": 8.64517819706499,
      "grad_norm": 0.35953545570373535,
      "learning_rate": 3.499011367027193e-05,
      "loss": 0.4691,
      "num_input_tokens_seen": 10789992,
      "step": 16495
    },
    {
      "epoch": 8.647798742138365,
      "grad_norm": 0.5970838665962219,
      "learning_rate": 3.4979630958203665e-05,
      "loss": 0.4436,
      "num_input_tokens_seen": 10793064,
      "step": 16500
    },
    {
      "epoch": 8.65041928721174,
      "grad_norm": 0.4086698889732361,
      "learning_rate": 3.496914615851046e-05,
      "loss": 0.5049,
      "num_input_tokens_seen": 10795784,
      "step": 16505
    },
    {
      "epoch": 8.653039832285115,
      "grad_norm": 0.8325532078742981,
      "learning_rate": 3.495865927338562e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 10798824,
      "step": 16510
    },
    {
      "epoch": 8.65566037735849,
      "grad_norm": 0.3219333291053772,
      "learning_rate": 3.494817030502286e-05,
      "loss": 0.5245,
      "num_input_tokens_seen": 10801960,
      "step": 16515
    },
    {
      "epoch": 8.658280922431866,
      "grad_norm": 0.400520384311676,
      "learning_rate": 3.493767925561638e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 10805320,
      "step": 16520
    },
    {
      "epoch": 8.66090146750524,
      "grad_norm": 0.27882876992225647,
      "learning_rate": 3.492718612736077e-05,
      "loss": 0.3434,
      "num_input_tokens_seen": 10808520,
      "step": 16525
    },
    {
      "epoch": 8.663522012578616,
      "grad_norm": 0.4947868585586548,
      "learning_rate": 3.491669092245108e-05,
      "loss": 0.5061,
      "num_input_tokens_seen": 10811176,
      "step": 16530
    },
    {
      "epoch": 8.666142557651991,
      "grad_norm": 0.7526494860649109,
      "learning_rate": 3.4906193643082784e-05,
      "loss": 0.4556,
      "num_input_tokens_seen": 10815240,
      "step": 16535
    },
    {
      "epoch": 8.668763102725366,
      "grad_norm": 0.5293556451797485,
      "learning_rate": 3.4895694291451795e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 10818152,
      "step": 16540
    },
    {
      "epoch": 8.671383647798741,
      "grad_norm": 0.6238390207290649,
      "learning_rate": 3.4885192869754444e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 10821128,
      "step": 16545
    },
    {
      "epoch": 8.674004192872118,
      "grad_norm": 0.43930408358573914,
      "learning_rate": 3.4874689380187537e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 10824040,
      "step": 16550
    },
    {
      "epoch": 8.676624737945493,
      "grad_norm": 0.25866442918777466,
      "learning_rate": 3.486418382494826e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 10827304,
      "step": 16555
    },
    {
      "epoch": 8.679245283018869,
      "grad_norm": 0.9876655340194702,
      "learning_rate": 3.485367620623426e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 10831144,
      "step": 16560
    },
    {
      "epoch": 8.681865828092244,
      "grad_norm": 0.6080846190452576,
      "learning_rate": 3.484316652624362e-05,
      "loss": 0.4281,
      "num_input_tokens_seen": 10834696,
      "step": 16565
    },
    {
      "epoch": 8.684486373165619,
      "grad_norm": 0.4735814034938812,
      "learning_rate": 3.483265478717484e-05,
      "loss": 0.3834,
      "num_input_tokens_seen": 10837320,
      "step": 16570
    },
    {
      "epoch": 8.687106918238994,
      "grad_norm": 0.6102723479270935,
      "learning_rate": 3.482214099122686e-05,
      "loss": 0.736,
      "num_input_tokens_seen": 10840008,
      "step": 16575
    },
    {
      "epoch": 8.68972746331237,
      "grad_norm": 0.7016969323158264,
      "learning_rate": 3.481162514059903e-05,
      "loss": 0.4052,
      "num_input_tokens_seen": 10844328,
      "step": 16580
    },
    {
      "epoch": 8.692348008385745,
      "grad_norm": 0.8035849928855896,
      "learning_rate": 3.480110723749116e-05,
      "loss": 0.5972,
      "num_input_tokens_seen": 10847400,
      "step": 16585
    },
    {
      "epoch": 8.69496855345912,
      "grad_norm": 0.32396841049194336,
      "learning_rate": 3.479058728410348e-05,
      "loss": 0.4031,
      "num_input_tokens_seen": 10850664,
      "step": 16590
    },
    {
      "epoch": 8.697589098532495,
      "grad_norm": 0.4104573130607605,
      "learning_rate": 3.478006528263663e-05,
      "loss": 0.5307,
      "num_input_tokens_seen": 10853864,
      "step": 16595
    },
    {
      "epoch": 8.70020964360587,
      "grad_norm": 0.6533710360527039,
      "learning_rate": 3.4769541235291704e-05,
      "loss": 0.4755,
      "num_input_tokens_seen": 10856936,
      "step": 16600
    },
    {
      "epoch": 8.702830188679245,
      "grad_norm": 0.3054073452949524,
      "learning_rate": 3.475901514427021e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 10860488,
      "step": 16605
    },
    {
      "epoch": 8.70545073375262,
      "grad_norm": 0.592666506767273,
      "learning_rate": 3.4748487011774075e-05,
      "loss": 0.3696,
      "num_input_tokens_seen": 10863592,
      "step": 16610
    },
    {
      "epoch": 8.708071278825996,
      "grad_norm": 0.2690662145614624,
      "learning_rate": 3.473795684000569e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 10866856,
      "step": 16615
    },
    {
      "epoch": 8.71069182389937,
      "grad_norm": 0.38409435749053955,
      "learning_rate": 3.4727424631167815e-05,
      "loss": 0.4091,
      "num_input_tokens_seen": 10869928,
      "step": 16620
    },
    {
      "epoch": 8.713312368972746,
      "grad_norm": 1.1269161701202393,
      "learning_rate": 3.471689038746369e-05,
      "loss": 0.759,
      "num_input_tokens_seen": 10872456,
      "step": 16625
    },
    {
      "epoch": 8.715932914046121,
      "grad_norm": 0.4230235815048218,
      "learning_rate": 3.470635411109695e-05,
      "loss": 0.386,
      "num_input_tokens_seen": 10876744,
      "step": 16630
    },
    {
      "epoch": 8.718553459119496,
      "grad_norm": 0.5411796569824219,
      "learning_rate": 3.469581580427167e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 10880072,
      "step": 16635
    },
    {
      "epoch": 8.721174004192871,
      "grad_norm": 0.5817884802818298,
      "learning_rate": 3.4685275469192335e-05,
      "loss": 0.5296,
      "num_input_tokens_seen": 10883112,
      "step": 16640
    },
    {
      "epoch": 8.723794549266248,
      "grad_norm": 0.4518725574016571,
      "learning_rate": 3.467473310806387e-05,
      "loss": 0.4031,
      "num_input_tokens_seen": 10886280,
      "step": 16645
    },
    {
      "epoch": 8.726415094339622,
      "grad_norm": 0.48757562041282654,
      "learning_rate": 3.466418872309161e-05,
      "loss": 0.4978,
      "num_input_tokens_seen": 10890280,
      "step": 16650
    },
    {
      "epoch": 8.729035639412999,
      "grad_norm": 0.5909173488616943,
      "learning_rate": 3.465364231648133e-05,
      "loss": 0.4005,
      "num_input_tokens_seen": 10893608,
      "step": 16655
    },
    {
      "epoch": 8.731656184486374,
      "grad_norm": 0.545526921749115,
      "learning_rate": 3.464309389043921e-05,
      "loss": 0.4707,
      "num_input_tokens_seen": 10896904,
      "step": 16660
    },
    {
      "epoch": 8.734276729559749,
      "grad_norm": 0.3804890513420105,
      "learning_rate": 3.4632543447171856e-05,
      "loss": 0.4281,
      "num_input_tokens_seen": 10899752,
      "step": 16665
    },
    {
      "epoch": 8.736897274633124,
      "grad_norm": 0.4868405759334564,
      "learning_rate": 3.4621990988886315e-05,
      "loss": 0.5045,
      "num_input_tokens_seen": 10903592,
      "step": 16670
    },
    {
      "epoch": 8.7395178197065,
      "grad_norm": 0.460302472114563,
      "learning_rate": 3.461143651779003e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 10907752,
      "step": 16675
    },
    {
      "epoch": 8.742138364779874,
      "grad_norm": 0.7736353278160095,
      "learning_rate": 3.4600880036090876e-05,
      "loss": 0.3781,
      "num_input_tokens_seen": 10910600,
      "step": 16680
    },
    {
      "epoch": 8.74475890985325,
      "grad_norm": 0.5150860548019409,
      "learning_rate": 3.459032154599715e-05,
      "loss": 0.3714,
      "num_input_tokens_seen": 10913800,
      "step": 16685
    },
    {
      "epoch": 8.747379454926625,
      "grad_norm": 0.44401493668556213,
      "learning_rate": 3.457976104971757e-05,
      "loss": 0.4764,
      "num_input_tokens_seen": 10916968,
      "step": 16690
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.3928193151950836,
      "learning_rate": 3.4569198549461276e-05,
      "loss": 0.3584,
      "num_input_tokens_seen": 10919944,
      "step": 16695
    },
    {
      "epoch": 8.752620545073375,
      "grad_norm": 0.541535496711731,
      "learning_rate": 3.455863404743781e-05,
      "loss": 0.3566,
      "num_input_tokens_seen": 10923144,
      "step": 16700
    },
    {
      "epoch": 8.75524109014675,
      "grad_norm": 1.2659432888031006,
      "learning_rate": 3.454806754585716e-05,
      "loss": 0.6855,
      "num_input_tokens_seen": 10926088,
      "step": 16705
    },
    {
      "epoch": 8.757861635220126,
      "grad_norm": 0.39808112382888794,
      "learning_rate": 3.453749904692971e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 10929416,
      "step": 16710
    },
    {
      "epoch": 8.7604821802935,
      "grad_norm": 0.2854675054550171,
      "learning_rate": 3.452692855286626e-05,
      "loss": 0.4852,
      "num_input_tokens_seen": 10934088,
      "step": 16715
    },
    {
      "epoch": 8.763102725366876,
      "grad_norm": 0.329956978559494,
      "learning_rate": 3.4516356065878055e-05,
      "loss": 0.5338,
      "num_input_tokens_seen": 10937256,
      "step": 16720
    },
    {
      "epoch": 8.765723270440251,
      "grad_norm": 0.7342736124992371,
      "learning_rate": 3.450578158817671e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 10940072,
      "step": 16725
    },
    {
      "epoch": 8.768343815513626,
      "grad_norm": 0.3221096396446228,
      "learning_rate": 3.4495205121974325e-05,
      "loss": 0.5441,
      "num_input_tokens_seen": 10943176,
      "step": 16730
    },
    {
      "epoch": 8.770964360587001,
      "grad_norm": 0.581536591053009,
      "learning_rate": 3.448462666948335e-05,
      "loss": 0.5369,
      "num_input_tokens_seen": 10946312,
      "step": 16735
    },
    {
      "epoch": 8.773584905660378,
      "grad_norm": 0.574982225894928,
      "learning_rate": 3.4474046232916674e-05,
      "loss": 0.3216,
      "num_input_tokens_seen": 10950792,
      "step": 16740
    },
    {
      "epoch": 8.776205450733752,
      "grad_norm": 0.35063081979751587,
      "learning_rate": 3.4463463814487616e-05,
      "loss": 0.4553,
      "num_input_tokens_seen": 10952936,
      "step": 16745
    },
    {
      "epoch": 8.778825995807129,
      "grad_norm": 0.47461968660354614,
      "learning_rate": 3.44528794164099e-05,
      "loss": 0.5306,
      "num_input_tokens_seen": 10957512,
      "step": 16750
    },
    {
      "epoch": 8.781446540880504,
      "grad_norm": 1.2603530883789062,
      "learning_rate": 3.4442293040897646e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 10960328,
      "step": 16755
    },
    {
      "epoch": 8.784067085953879,
      "grad_norm": 0.50763338804245,
      "learning_rate": 3.443170469016542e-05,
      "loss": 0.4079,
      "num_input_tokens_seen": 10963944,
      "step": 16760
    },
    {
      "epoch": 8.786687631027254,
      "grad_norm": 0.448595255613327,
      "learning_rate": 3.442111436642817e-05,
      "loss": 0.3738,
      "num_input_tokens_seen": 10967272,
      "step": 16765
    },
    {
      "epoch": 8.78930817610063,
      "grad_norm": 0.9222992658615112,
      "learning_rate": 3.441052207190127e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 10970440,
      "step": 16770
    },
    {
      "epoch": 8.791928721174004,
      "grad_norm": 0.26820626854896545,
      "learning_rate": 3.4399927808800526e-05,
      "loss": 0.378,
      "num_input_tokens_seen": 10974568,
      "step": 16775
    },
    {
      "epoch": 8.79454926624738,
      "grad_norm": 0.598086416721344,
      "learning_rate": 3.438933157934211e-05,
      "loss": 0.5379,
      "num_input_tokens_seen": 10977096,
      "step": 16780
    },
    {
      "epoch": 8.797169811320755,
      "grad_norm": 0.34317272901535034,
      "learning_rate": 3.4378733385742653e-05,
      "loss": 0.2967,
      "num_input_tokens_seen": 10980200,
      "step": 16785
    },
    {
      "epoch": 8.79979035639413,
      "grad_norm": 0.457347571849823,
      "learning_rate": 3.436813323021918e-05,
      "loss": 0.4593,
      "num_input_tokens_seen": 10983272,
      "step": 16790
    },
    {
      "epoch": 8.802410901467505,
      "grad_norm": 0.4549928307533264,
      "learning_rate": 3.43575311149891e-05,
      "loss": 0.6521,
      "num_input_tokens_seen": 10986952,
      "step": 16795
    },
    {
      "epoch": 8.80503144654088,
      "grad_norm": 0.2166144847869873,
      "learning_rate": 3.434692704227028e-05,
      "loss": 0.3783,
      "num_input_tokens_seen": 10990728,
      "step": 16800
    },
    {
      "epoch": 8.807651991614255,
      "grad_norm": 0.3631468415260315,
      "learning_rate": 3.4336321014280944e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 10994472,
      "step": 16805
    },
    {
      "epoch": 8.81027253668763,
      "grad_norm": 0.6620362401008606,
      "learning_rate": 3.432571303323977e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 11001224,
      "step": 16810
    },
    {
      "epoch": 8.812893081761006,
      "grad_norm": 0.4289611577987671,
      "learning_rate": 3.431510310136583e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 11005032,
      "step": 16815
    },
    {
      "epoch": 8.815513626834381,
      "grad_norm": 0.5104908347129822,
      "learning_rate": 3.4304491220878585e-05,
      "loss": 0.4391,
      "num_input_tokens_seen": 11007624,
      "step": 16820
    },
    {
      "epoch": 8.818134171907756,
      "grad_norm": 0.49442198872566223,
      "learning_rate": 3.429387739399793e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 11009704,
      "step": 16825
    },
    {
      "epoch": 8.820754716981131,
      "grad_norm": 0.5974138975143433,
      "learning_rate": 3.428326162294416e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 11013128,
      "step": 16830
    },
    {
      "epoch": 8.823375262054507,
      "grad_norm": 0.2698252201080322,
      "learning_rate": 3.427264390993795e-05,
      "loss": 0.3927,
      "num_input_tokens_seen": 11016328,
      "step": 16835
    },
    {
      "epoch": 8.825995807127882,
      "grad_norm": 0.3771113455295563,
      "learning_rate": 3.426202425720044e-05,
      "loss": 0.6259,
      "num_input_tokens_seen": 11019368,
      "step": 16840
    },
    {
      "epoch": 8.828616352201259,
      "grad_norm": 0.49800458550453186,
      "learning_rate": 3.425140266695312e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 11022696,
      "step": 16845
    },
    {
      "epoch": 8.831236897274634,
      "grad_norm": 0.516868531703949,
      "learning_rate": 3.4240779141417905e-05,
      "loss": 0.418,
      "num_input_tokens_seen": 11025992,
      "step": 16850
    },
    {
      "epoch": 8.833857442348009,
      "grad_norm": 0.30528321862220764,
      "learning_rate": 3.423015368281711e-05,
      "loss": 0.4126,
      "num_input_tokens_seen": 11028904,
      "step": 16855
    },
    {
      "epoch": 8.836477987421384,
      "grad_norm": 0.3453669846057892,
      "learning_rate": 3.421952629337348e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 11032712,
      "step": 16860
    },
    {
      "epoch": 8.83909853249476,
      "grad_norm": 0.6101237535476685,
      "learning_rate": 3.420889697531013e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 11036392,
      "step": 16865
    },
    {
      "epoch": 8.841719077568134,
      "grad_norm": 0.558711051940918,
      "learning_rate": 3.419826573085058e-05,
      "loss": 0.482,
      "num_input_tokens_seen": 11039368,
      "step": 16870
    },
    {
      "epoch": 8.84433962264151,
      "grad_norm": 0.47311824560165405,
      "learning_rate": 3.418763256221879e-05,
      "loss": 0.3562,
      "num_input_tokens_seen": 11042088,
      "step": 16875
    },
    {
      "epoch": 8.846960167714885,
      "grad_norm": 0.5150076746940613,
      "learning_rate": 3.4176997471639085e-05,
      "loss": 0.563,
      "num_input_tokens_seen": 11045416,
      "step": 16880
    },
    {
      "epoch": 8.84958071278826,
      "grad_norm": 0.6739036440849304,
      "learning_rate": 3.4166360461336206e-05,
      "loss": 0.4779,
      "num_input_tokens_seen": 11048552,
      "step": 16885
    },
    {
      "epoch": 8.852201257861635,
      "grad_norm": 0.4537280797958374,
      "learning_rate": 3.41557215335353e-05,
      "loss": 0.3454,
      "num_input_tokens_seen": 11052072,
      "step": 16890
    },
    {
      "epoch": 8.85482180293501,
      "grad_norm": 0.594774603843689,
      "learning_rate": 3.4145080690461887e-05,
      "loss": 0.5501,
      "num_input_tokens_seen": 11055016,
      "step": 16895
    },
    {
      "epoch": 8.857442348008385,
      "grad_norm": 0.28030669689178467,
      "learning_rate": 3.4134437934341936e-05,
      "loss": 0.3625,
      "num_input_tokens_seen": 11058376,
      "step": 16900
    },
    {
      "epoch": 8.86006289308176,
      "grad_norm": 0.3981115221977234,
      "learning_rate": 3.4123793267401784e-05,
      "loss": 0.442,
      "num_input_tokens_seen": 11061384,
      "step": 16905
    },
    {
      "epoch": 8.862683438155136,
      "grad_norm": 0.46841001510620117,
      "learning_rate": 3.411314669186816e-05,
      "loss": 0.537,
      "num_input_tokens_seen": 11064232,
      "step": 16910
    },
    {
      "epoch": 8.865303983228511,
      "grad_norm": 0.39767974615097046,
      "learning_rate": 3.410249820996823e-05,
      "loss": 0.3808,
      "num_input_tokens_seen": 11067368,
      "step": 16915
    },
    {
      "epoch": 8.867924528301886,
      "grad_norm": 0.4866785705089569,
      "learning_rate": 3.4091847823929513e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 11070056,
      "step": 16920
    },
    {
      "epoch": 8.870545073375261,
      "grad_norm": 0.49234458804130554,
      "learning_rate": 3.4081195535979946e-05,
      "loss": 0.4334,
      "num_input_tokens_seen": 11073768,
      "step": 16925
    },
    {
      "epoch": 8.873165618448636,
      "grad_norm": 0.5233428478240967,
      "learning_rate": 3.40705413483479e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 11076744,
      "step": 16930
    },
    {
      "epoch": 8.875786163522012,
      "grad_norm": 0.5984055399894714,
      "learning_rate": 3.405988526326207e-05,
      "loss": 0.4377,
      "num_input_tokens_seen": 11079848,
      "step": 16935
    },
    {
      "epoch": 8.878406708595389,
      "grad_norm": 0.49848541617393494,
      "learning_rate": 3.40492272829516e-05,
      "loss": 0.4951,
      "num_input_tokens_seen": 11082600,
      "step": 16940
    },
    {
      "epoch": 8.881027253668764,
      "grad_norm": 0.5562469363212585,
      "learning_rate": 3.403856740964604e-05,
      "loss": 0.4923,
      "num_input_tokens_seen": 11086280,
      "step": 16945
    },
    {
      "epoch": 8.883647798742139,
      "grad_norm": 0.4148798882961273,
      "learning_rate": 3.4027905645575275e-05,
      "loss": 0.5328,
      "num_input_tokens_seen": 11089352,
      "step": 16950
    },
    {
      "epoch": 8.886268343815514,
      "grad_norm": 0.7050207853317261,
      "learning_rate": 3.401724199296966e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 11092520,
      "step": 16955
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 0.5092171430587769,
      "learning_rate": 3.400657645405989e-05,
      "loss": 0.4906,
      "num_input_tokens_seen": 11096136,
      "step": 16960
    },
    {
      "epoch": 8.891509433962264,
      "grad_norm": 0.8411828875541687,
      "learning_rate": 3.3995909031077065e-05,
      "loss": 0.528,
      "num_input_tokens_seen": 11099656,
      "step": 16965
    },
    {
      "epoch": 8.89412997903564,
      "grad_norm": 1.0594422817230225,
      "learning_rate": 3.398523972625272e-05,
      "loss": 0.5481,
      "num_input_tokens_seen": 11102056,
      "step": 16970
    },
    {
      "epoch": 8.896750524109015,
      "grad_norm": 0.4323408901691437,
      "learning_rate": 3.3974568541818723e-05,
      "loss": 0.3264,
      "num_input_tokens_seen": 11104712,
      "step": 16975
    },
    {
      "epoch": 8.89937106918239,
      "grad_norm": 0.277832567691803,
      "learning_rate": 3.396389548000738e-05,
      "loss": 0.2981,
      "num_input_tokens_seen": 11108616,
      "step": 16980
    },
    {
      "epoch": 8.901991614255765,
      "grad_norm": 0.8680922389030457,
      "learning_rate": 3.395322054305137e-05,
      "loss": 0.4948,
      "num_input_tokens_seen": 11111688,
      "step": 16985
    },
    {
      "epoch": 8.90461215932914,
      "grad_norm": 0.6752920746803284,
      "learning_rate": 3.394254373318376e-05,
      "loss": 0.3933,
      "num_input_tokens_seen": 11114632,
      "step": 16990
    },
    {
      "epoch": 8.907232704402515,
      "grad_norm": 0.444749653339386,
      "learning_rate": 3.3931865052638026e-05,
      "loss": 0.3858,
      "num_input_tokens_seen": 11117960,
      "step": 16995
    },
    {
      "epoch": 8.90985324947589,
      "grad_norm": 0.41120463609695435,
      "learning_rate": 3.392118450364802e-05,
      "loss": 0.5698,
      "num_input_tokens_seen": 11121384,
      "step": 17000
    },
    {
      "epoch": 8.912473794549266,
      "grad_norm": 0.6276853084564209,
      "learning_rate": 3.391050208844799e-05,
      "loss": 0.4472,
      "num_input_tokens_seen": 11124488,
      "step": 17005
    },
    {
      "epoch": 8.915094339622641,
      "grad_norm": 0.4035504162311554,
      "learning_rate": 3.389981780927259e-05,
      "loss": 0.4038,
      "num_input_tokens_seen": 11127656,
      "step": 17010
    },
    {
      "epoch": 8.917714884696016,
      "grad_norm": 0.37104368209838867,
      "learning_rate": 3.3889131668356834e-05,
      "loss": 0.4174,
      "num_input_tokens_seen": 11130024,
      "step": 17015
    },
    {
      "epoch": 8.920335429769391,
      "grad_norm": 0.5487117767333984,
      "learning_rate": 3.387844366793614e-05,
      "loss": 0.5666,
      "num_input_tokens_seen": 11132872,
      "step": 17020
    },
    {
      "epoch": 8.922955974842766,
      "grad_norm": 0.3495144546031952,
      "learning_rate": 3.3867753810246326e-05,
      "loss": 0.4403,
      "num_input_tokens_seen": 11136488,
      "step": 17025
    },
    {
      "epoch": 8.925576519916142,
      "grad_norm": 0.37089037895202637,
      "learning_rate": 3.3857062097523586e-05,
      "loss": 0.3008,
      "num_input_tokens_seen": 11140584,
      "step": 17030
    },
    {
      "epoch": 8.928197064989519,
      "grad_norm": 0.4112498164176941,
      "learning_rate": 3.38463685320045e-05,
      "loss": 0.462,
      "num_input_tokens_seen": 11143304,
      "step": 17035
    },
    {
      "epoch": 8.930817610062894,
      "grad_norm": 0.3400362730026245,
      "learning_rate": 3.3835673115926045e-05,
      "loss": 0.3937,
      "num_input_tokens_seen": 11146792,
      "step": 17040
    },
    {
      "epoch": 8.933438155136269,
      "grad_norm": 0.48928406834602356,
      "learning_rate": 3.382497585152557e-05,
      "loss": 0.3299,
      "num_input_tokens_seen": 11149832,
      "step": 17045
    },
    {
      "epoch": 8.936058700209644,
      "grad_norm": 0.31599146127700806,
      "learning_rate": 3.3814276741040835e-05,
      "loss": 0.376,
      "num_input_tokens_seen": 11153192,
      "step": 17050
    },
    {
      "epoch": 8.93867924528302,
      "grad_norm": 0.36744752526283264,
      "learning_rate": 3.380357578670996e-05,
      "loss": 0.4713,
      "num_input_tokens_seen": 11155912,
      "step": 17055
    },
    {
      "epoch": 8.941299790356394,
      "grad_norm": 0.5376017093658447,
      "learning_rate": 3.379287299077146e-05,
      "loss": 0.4028,
      "num_input_tokens_seen": 11158472,
      "step": 17060
    },
    {
      "epoch": 8.94392033542977,
      "grad_norm": 0.5411480069160461,
      "learning_rate": 3.3782168355464263e-05,
      "loss": 0.5488,
      "num_input_tokens_seen": 11161192,
      "step": 17065
    },
    {
      "epoch": 8.946540880503145,
      "grad_norm": 0.4865163564682007,
      "learning_rate": 3.377146188302762e-05,
      "loss": 0.3737,
      "num_input_tokens_seen": 11163688,
      "step": 17070
    },
    {
      "epoch": 8.94916142557652,
      "grad_norm": 0.6923845410346985,
      "learning_rate": 3.376075357570124e-05,
      "loss": 0.5037,
      "num_input_tokens_seen": 11166408,
      "step": 17075
    },
    {
      "epoch": 8.951781970649895,
      "grad_norm": 0.7203994989395142,
      "learning_rate": 3.3750043435725144e-05,
      "loss": 0.5281,
      "num_input_tokens_seen": 11169640,
      "step": 17080
    },
    {
      "epoch": 8.95440251572327,
      "grad_norm": 0.7381361722946167,
      "learning_rate": 3.373933146533979e-05,
      "loss": 0.4133,
      "num_input_tokens_seen": 11172616,
      "step": 17085
    },
    {
      "epoch": 8.957023060796645,
      "grad_norm": 0.4083919823169708,
      "learning_rate": 3.3728617666786e-05,
      "loss": 0.7064,
      "num_input_tokens_seen": 11175880,
      "step": 17090
    },
    {
      "epoch": 8.95964360587002,
      "grad_norm": 0.34420108795166016,
      "learning_rate": 3.371790204230497e-05,
      "loss": 0.577,
      "num_input_tokens_seen": 11178696,
      "step": 17095
    },
    {
      "epoch": 8.962264150943396,
      "grad_norm": 0.353825181722641,
      "learning_rate": 3.3707184594138295e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 11182152,
      "step": 17100
    },
    {
      "epoch": 8.964884696016771,
      "grad_norm": 0.5948107242584229,
      "learning_rate": 3.369646532452795e-05,
      "loss": 0.5581,
      "num_input_tokens_seen": 11185416,
      "step": 17105
    },
    {
      "epoch": 8.967505241090146,
      "grad_norm": 0.6811187863349915,
      "learning_rate": 3.3685744235716254e-05,
      "loss": 0.5229,
      "num_input_tokens_seen": 11188392,
      "step": 17110
    },
    {
      "epoch": 8.970125786163521,
      "grad_norm": 0.3676571249961853,
      "learning_rate": 3.367502132994598e-05,
      "loss": 0.4686,
      "num_input_tokens_seen": 11191528,
      "step": 17115
    },
    {
      "epoch": 8.972746331236896,
      "grad_norm": 0.3213559687137604,
      "learning_rate": 3.366429660946019e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 11194792,
      "step": 17120
    },
    {
      "epoch": 8.975366876310272,
      "grad_norm": 0.28003978729248047,
      "learning_rate": 3.365357007650241e-05,
      "loss": 0.459,
      "num_input_tokens_seen": 11198600,
      "step": 17125
    },
    {
      "epoch": 8.977987421383649,
      "grad_norm": 0.5378799438476562,
      "learning_rate": 3.364284173331649e-05,
      "loss": 0.3192,
      "num_input_tokens_seen": 11201448,
      "step": 17130
    },
    {
      "epoch": 8.980607966457024,
      "grad_norm": 0.4751434624195099,
      "learning_rate": 3.3632111582146686e-05,
      "loss": 0.3704,
      "num_input_tokens_seen": 11205160,
      "step": 17135
    },
    {
      "epoch": 8.983228511530399,
      "grad_norm": 0.34951189160346985,
      "learning_rate": 3.362137962523761e-05,
      "loss": 0.3055,
      "num_input_tokens_seen": 11208328,
      "step": 17140
    },
    {
      "epoch": 8.985849056603774,
      "grad_norm": 0.3443482220172882,
      "learning_rate": 3.361064586483429e-05,
      "loss": 0.4447,
      "num_input_tokens_seen": 11211848,
      "step": 17145
    },
    {
      "epoch": 8.98846960167715,
      "grad_norm": 0.3369937837123871,
      "learning_rate": 3.359991030318208e-05,
      "loss": 0.5502,
      "num_input_tokens_seen": 11214376,
      "step": 17150
    },
    {
      "epoch": 8.991090146750524,
      "grad_norm": 0.4960320293903351,
      "learning_rate": 3.358917294252675e-05,
      "loss": 0.3774,
      "num_input_tokens_seen": 11218376,
      "step": 17155
    },
    {
      "epoch": 8.9937106918239,
      "grad_norm": 0.3452340364456177,
      "learning_rate": 3.357843378511443e-05,
      "loss": 0.4417,
      "num_input_tokens_seen": 11221768,
      "step": 17160
    },
    {
      "epoch": 8.996331236897275,
      "grad_norm": 0.5198101997375488,
      "learning_rate": 3.3567692833191635e-05,
      "loss": 0.4484,
      "num_input_tokens_seen": 11224712,
      "step": 17165
    },
    {
      "epoch": 8.99895178197065,
      "grad_norm": 0.42118123173713684,
      "learning_rate": 3.355695008900525e-05,
      "loss": 0.3541,
      "num_input_tokens_seen": 11229096,
      "step": 17170
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.4490000903606415,
      "eval_runtime": 9.6461,
      "eval_samples_per_second": 87.911,
      "eval_steps_per_second": 21.978,
      "num_input_tokens_seen": 11229792,
      "step": 17172
    },
    {
      "epoch": 9.001572327044025,
      "grad_norm": 0.43935590982437134,
      "learning_rate": 3.354620555480251e-05,
      "loss": 0.2568,
      "num_input_tokens_seen": 11232064,
      "step": 17175
    },
    {
      "epoch": 9.0041928721174,
      "grad_norm": 0.38569456338882446,
      "learning_rate": 3.353545923283108e-05,
      "loss": 0.6333,
      "num_input_tokens_seen": 11236960,
      "step": 17180
    },
    {
      "epoch": 9.006813417190775,
      "grad_norm": 0.44529077410697937,
      "learning_rate": 3.352471112533895e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 11239840,
      "step": 17185
    },
    {
      "epoch": 9.00943396226415,
      "grad_norm": 0.5747714042663574,
      "learning_rate": 3.351396123457451e-05,
      "loss": 0.4498,
      "num_input_tokens_seen": 11242432,
      "step": 17190
    },
    {
      "epoch": 9.012054507337526,
      "grad_norm": 0.4334632158279419,
      "learning_rate": 3.350320956278652e-05,
      "loss": 0.3111,
      "num_input_tokens_seen": 11245312,
      "step": 17195
    },
    {
      "epoch": 9.014675052410901,
      "grad_norm": 0.8173248767852783,
      "learning_rate": 3.349245611222409e-05,
      "loss": 0.4241,
      "num_input_tokens_seen": 11249152,
      "step": 17200
    },
    {
      "epoch": 9.017295597484276,
      "grad_norm": 0.7433425188064575,
      "learning_rate": 3.348170088513673e-05,
      "loss": 0.5935,
      "num_input_tokens_seen": 11251968,
      "step": 17205
    },
    {
      "epoch": 9.019916142557651,
      "grad_norm": 0.5117957592010498,
      "learning_rate": 3.347094388377431e-05,
      "loss": 0.5449,
      "num_input_tokens_seen": 11256224,
      "step": 17210
    },
    {
      "epoch": 9.022536687631026,
      "grad_norm": 0.5129490494728088,
      "learning_rate": 3.3460185110387076e-05,
      "loss": 0.447,
      "num_input_tokens_seen": 11258944,
      "step": 17215
    },
    {
      "epoch": 9.025157232704403,
      "grad_norm": 0.6249842047691345,
      "learning_rate": 3.344942456722563e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 11263488,
      "step": 17220
    },
    {
      "epoch": 9.027777777777779,
      "grad_norm": 0.4317395091056824,
      "learning_rate": 3.343866225654096e-05,
      "loss": 0.4069,
      "num_input_tokens_seen": 11266560,
      "step": 17225
    },
    {
      "epoch": 9.030398322851154,
      "grad_norm": 0.3386372923851013,
      "learning_rate": 3.342789818058443e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 11270336,
      "step": 17230
    },
    {
      "epoch": 9.033018867924529,
      "grad_norm": 0.49456658959388733,
      "learning_rate": 3.341713234160774e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 11273536,
      "step": 17235
    },
    {
      "epoch": 9.035639412997904,
      "grad_norm": 0.4060143232345581,
      "learning_rate": 3.3406364741862986e-05,
      "loss": 0.4138,
      "num_input_tokens_seen": 11277408,
      "step": 17240
    },
    {
      "epoch": 9.03825995807128,
      "grad_norm": 0.4070079028606415,
      "learning_rate": 3.3395595383602644e-05,
      "loss": 0.3971,
      "num_input_tokens_seen": 11280576,
      "step": 17245
    },
    {
      "epoch": 9.040880503144654,
      "grad_norm": 0.30726465582847595,
      "learning_rate": 3.338482426907953e-05,
      "loss": 0.4752,
      "num_input_tokens_seen": 11283904,
      "step": 17250
    },
    {
      "epoch": 9.04350104821803,
      "grad_norm": 0.34798911213874817,
      "learning_rate": 3.337405140054683e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 11286592,
      "step": 17255
    },
    {
      "epoch": 9.046121593291405,
      "grad_norm": 0.5284243226051331,
      "learning_rate": 3.336327678025813e-05,
      "loss": 0.3862,
      "num_input_tokens_seen": 11289376,
      "step": 17260
    },
    {
      "epoch": 9.04874213836478,
      "grad_norm": 0.9036619663238525,
      "learning_rate": 3.3352500410467324e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 11292320,
      "step": 17265
    },
    {
      "epoch": 9.051362683438155,
      "grad_norm": 0.4260571002960205,
      "learning_rate": 3.3341722293428726e-05,
      "loss": 0.6319,
      "num_input_tokens_seen": 11295392,
      "step": 17270
    },
    {
      "epoch": 9.05398322851153,
      "grad_norm": 0.5901222229003906,
      "learning_rate": 3.3330942431397e-05,
      "loss": 0.5394,
      "num_input_tokens_seen": 11298656,
      "step": 17275
    },
    {
      "epoch": 9.056603773584905,
      "grad_norm": 0.5573202967643738,
      "learning_rate": 3.332016082662715e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 11301824,
      "step": 17280
    },
    {
      "epoch": 9.05922431865828,
      "grad_norm": 0.30713945627212524,
      "learning_rate": 3.3309377481374584e-05,
      "loss": 0.4563,
      "num_input_tokens_seen": 11304352,
      "step": 17285
    },
    {
      "epoch": 9.061844863731656,
      "grad_norm": 0.5877187252044678,
      "learning_rate": 3.3298592397895046e-05,
      "loss": 0.4745,
      "num_input_tokens_seen": 11306976,
      "step": 17290
    },
    {
      "epoch": 9.064465408805031,
      "grad_norm": 0.6907803416252136,
      "learning_rate": 3.3287805578444645e-05,
      "loss": 0.4611,
      "num_input_tokens_seen": 11310048,
      "step": 17295
    },
    {
      "epoch": 9.067085953878406,
      "grad_norm": 0.35561123490333557,
      "learning_rate": 3.327701702527989e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 11313984,
      "step": 17300
    },
    {
      "epoch": 9.069706498951781,
      "grad_norm": 0.40400540828704834,
      "learning_rate": 3.3266226740657584e-05,
      "loss": 0.3194,
      "num_input_tokens_seen": 11317536,
      "step": 17305
    },
    {
      "epoch": 9.072327044025156,
      "grad_norm": 0.2929593622684479,
      "learning_rate": 3.3255434726834954e-05,
      "loss": 0.398,
      "num_input_tokens_seen": 11321152,
      "step": 17310
    },
    {
      "epoch": 9.074947589098532,
      "grad_norm": 0.766919732093811,
      "learning_rate": 3.324464098606957e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 11323968,
      "step": 17315
    },
    {
      "epoch": 9.077568134171909,
      "grad_norm": 0.4665292799472809,
      "learning_rate": 3.323384552061934e-05,
      "loss": 0.4142,
      "num_input_tokens_seen": 11326496,
      "step": 17320
    },
    {
      "epoch": 9.080188679245284,
      "grad_norm": 0.648837685585022,
      "learning_rate": 3.322304833274257e-05,
      "loss": 0.4229,
      "num_input_tokens_seen": 11329696,
      "step": 17325
    },
    {
      "epoch": 9.082809224318659,
      "grad_norm": 0.27910202741622925,
      "learning_rate": 3.32122494246979e-05,
      "loss": 0.375,
      "num_input_tokens_seen": 11333120,
      "step": 17330
    },
    {
      "epoch": 9.085429769392034,
      "grad_norm": 0.3053613305091858,
      "learning_rate": 3.320144879874434e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 11336352,
      "step": 17335
    },
    {
      "epoch": 9.08805031446541,
      "grad_norm": 0.3675723075866699,
      "learning_rate": 3.319064645714126e-05,
      "loss": 0.4176,
      "num_input_tokens_seen": 11339968,
      "step": 17340
    },
    {
      "epoch": 9.090670859538784,
      "grad_norm": 0.5359514355659485,
      "learning_rate": 3.3179842402148386e-05,
      "loss": 0.4137,
      "num_input_tokens_seen": 11343296,
      "step": 17345
    },
    {
      "epoch": 9.09329140461216,
      "grad_norm": 0.36263787746429443,
      "learning_rate": 3.316903663602581e-05,
      "loss": 0.5596,
      "num_input_tokens_seen": 11347808,
      "step": 17350
    },
    {
      "epoch": 9.095911949685535,
      "grad_norm": 0.5668581128120422,
      "learning_rate": 3.315822916103395e-05,
      "loss": 0.5709,
      "num_input_tokens_seen": 11350880,
      "step": 17355
    },
    {
      "epoch": 9.09853249475891,
      "grad_norm": 0.27220186591148376,
      "learning_rate": 3.3147419979433635e-05,
      "loss": 0.3962,
      "num_input_tokens_seen": 11354016,
      "step": 17360
    },
    {
      "epoch": 9.101153039832285,
      "grad_norm": 0.3445053696632385,
      "learning_rate": 3.313660909348601e-05,
      "loss": 0.3619,
      "num_input_tokens_seen": 11357504,
      "step": 17365
    },
    {
      "epoch": 9.10377358490566,
      "grad_norm": 0.36188191175460815,
      "learning_rate": 3.312579650545259e-05,
      "loss": 0.4739,
      "num_input_tokens_seen": 11360512,
      "step": 17370
    },
    {
      "epoch": 9.106394129979035,
      "grad_norm": 0.5511714816093445,
      "learning_rate": 3.311498221759524e-05,
      "loss": 0.5283,
      "num_input_tokens_seen": 11363392,
      "step": 17375
    },
    {
      "epoch": 9.10901467505241,
      "grad_norm": 0.37779203057289124,
      "learning_rate": 3.310416623217619e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 11366400,
      "step": 17380
    },
    {
      "epoch": 9.111635220125786,
      "grad_norm": 0.6839309334754944,
      "learning_rate": 3.309334855145803e-05,
      "loss": 0.4159,
      "num_input_tokens_seen": 11369824,
      "step": 17385
    },
    {
      "epoch": 9.114255765199161,
      "grad_norm": 0.4776456654071808,
      "learning_rate": 3.3082529177703684e-05,
      "loss": 0.5213,
      "num_input_tokens_seen": 11373376,
      "step": 17390
    },
    {
      "epoch": 9.116876310272536,
      "grad_norm": 0.43228062987327576,
      "learning_rate": 3.3071708113176435e-05,
      "loss": 0.4412,
      "num_input_tokens_seen": 11376672,
      "step": 17395
    },
    {
      "epoch": 9.119496855345911,
      "grad_norm": 0.3913607597351074,
      "learning_rate": 3.306088536013995e-05,
      "loss": 0.5034,
      "num_input_tokens_seen": 11379456,
      "step": 17400
    },
    {
      "epoch": 9.122117400419286,
      "grad_norm": 0.5373203754425049,
      "learning_rate": 3.3050060920858206e-05,
      "loss": 0.4823,
      "num_input_tokens_seen": 11382560,
      "step": 17405
    },
    {
      "epoch": 9.124737945492662,
      "grad_norm": 0.317547470331192,
      "learning_rate": 3.3039234797595544e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 11385792,
      "step": 17410
    },
    {
      "epoch": 9.127358490566039,
      "grad_norm": 0.4491869807243347,
      "learning_rate": 3.3028406992616695e-05,
      "loss": 0.4031,
      "num_input_tokens_seen": 11391648,
      "step": 17415
    },
    {
      "epoch": 9.129979035639414,
      "grad_norm": 0.36720767617225647,
      "learning_rate": 3.301757750818669e-05,
      "loss": 0.3287,
      "num_input_tokens_seen": 11395744,
      "step": 17420
    },
    {
      "epoch": 9.132599580712789,
      "grad_norm": 0.35237133502960205,
      "learning_rate": 3.300674634657094e-05,
      "loss": 0.4264,
      "num_input_tokens_seen": 11398496,
      "step": 17425
    },
    {
      "epoch": 9.135220125786164,
      "grad_norm": 0.5664875507354736,
      "learning_rate": 3.2995913510035206e-05,
      "loss": 0.3851,
      "num_input_tokens_seen": 11401408,
      "step": 17430
    },
    {
      "epoch": 9.13784067085954,
      "grad_norm": 0.5742624998092651,
      "learning_rate": 3.298507900084558e-05,
      "loss": 0.5313,
      "num_input_tokens_seen": 11404448,
      "step": 17435
    },
    {
      "epoch": 9.140461215932914,
      "grad_norm": 0.6094236373901367,
      "learning_rate": 3.2974242821268524e-05,
      "loss": 0.421,
      "num_input_tokens_seen": 11407680,
      "step": 17440
    },
    {
      "epoch": 9.14308176100629,
      "grad_norm": 0.35580629110336304,
      "learning_rate": 3.2963404973570855e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 11410752,
      "step": 17445
    },
    {
      "epoch": 9.145702306079665,
      "grad_norm": 0.447096049785614,
      "learning_rate": 3.2952565460019715e-05,
      "loss": 0.3913,
      "num_input_tokens_seen": 11414432,
      "step": 17450
    },
    {
      "epoch": 9.14832285115304,
      "grad_norm": 0.5220810770988464,
      "learning_rate": 3.29417242828826e-05,
      "loss": 0.548,
      "num_input_tokens_seen": 11417568,
      "step": 17455
    },
    {
      "epoch": 9.150943396226415,
      "grad_norm": 0.9776028394699097,
      "learning_rate": 3.2930881444427374e-05,
      "loss": 0.5327,
      "num_input_tokens_seen": 11421248,
      "step": 17460
    },
    {
      "epoch": 9.15356394129979,
      "grad_norm": 0.48964524269104004,
      "learning_rate": 3.292003694692223e-05,
      "loss": 0.3875,
      "num_input_tokens_seen": 11425472,
      "step": 17465
    },
    {
      "epoch": 9.156184486373165,
      "grad_norm": 0.2952135503292084,
      "learning_rate": 3.290919079263572e-05,
      "loss": 0.4802,
      "num_input_tokens_seen": 11428928,
      "step": 17470
    },
    {
      "epoch": 9.15880503144654,
      "grad_norm": 0.6448747515678406,
      "learning_rate": 3.2898342983836715e-05,
      "loss": 0.5159,
      "num_input_tokens_seen": 11433664,
      "step": 17475
    },
    {
      "epoch": 9.161425576519916,
      "grad_norm": 0.36912402510643005,
      "learning_rate": 3.288749352279447e-05,
      "loss": 0.6773,
      "num_input_tokens_seen": 11437024,
      "step": 17480
    },
    {
      "epoch": 9.164046121593291,
      "grad_norm": 0.404232382774353,
      "learning_rate": 3.2876642411778573e-05,
      "loss": 0.3535,
      "num_input_tokens_seen": 11442304,
      "step": 17485
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 0.42528143525123596,
      "learning_rate": 3.286578965305894e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 11445248,
      "step": 17490
    },
    {
      "epoch": 9.169287211740041,
      "grad_norm": 1.6021690368652344,
      "learning_rate": 3.285493524890585e-05,
      "loss": 0.6096,
      "num_input_tokens_seen": 11448064,
      "step": 17495
    },
    {
      "epoch": 9.171907756813416,
      "grad_norm": 0.3001665472984314,
      "learning_rate": 3.284407920158992e-05,
      "loss": 0.3648,
      "num_input_tokens_seen": 11451232,
      "step": 17500
    },
    {
      "epoch": 9.174528301886792,
      "grad_norm": 0.4304790794849396,
      "learning_rate": 3.283322151338211e-05,
      "loss": 0.411,
      "num_input_tokens_seen": 11455648,
      "step": 17505
    },
    {
      "epoch": 9.177148846960169,
      "grad_norm": 0.4743507504463196,
      "learning_rate": 3.282236218655373e-05,
      "loss": 0.5534,
      "num_input_tokens_seen": 11459296,
      "step": 17510
    },
    {
      "epoch": 9.179769392033544,
      "grad_norm": 0.4234893321990967,
      "learning_rate": 3.281150122337642e-05,
      "loss": 0.6243,
      "num_input_tokens_seen": 11462240,
      "step": 17515
    },
    {
      "epoch": 9.182389937106919,
      "grad_norm": 0.5024523138999939,
      "learning_rate": 3.280063862612216e-05,
      "loss": 0.5099,
      "num_input_tokens_seen": 11465600,
      "step": 17520
    },
    {
      "epoch": 9.185010482180294,
      "grad_norm": 0.40229207277297974,
      "learning_rate": 3.2789774397063314e-05,
      "loss": 0.447,
      "num_input_tokens_seen": 11468896,
      "step": 17525
    },
    {
      "epoch": 9.18763102725367,
      "grad_norm": 0.6377207636833191,
      "learning_rate": 3.277890853847253e-05,
      "loss": 0.5641,
      "num_input_tokens_seen": 11471808,
      "step": 17530
    },
    {
      "epoch": 9.190251572327044,
      "grad_norm": 0.6478783488273621,
      "learning_rate": 3.2768041052622824e-05,
      "loss": 0.3396,
      "num_input_tokens_seen": 11474752,
      "step": 17535
    },
    {
      "epoch": 9.19287211740042,
      "grad_norm": 0.47677138447761536,
      "learning_rate": 3.275717194178755e-05,
      "loss": 0.384,
      "num_input_tokens_seen": 11477760,
      "step": 17540
    },
    {
      "epoch": 9.195492662473795,
      "grad_norm": 0.43984562158584595,
      "learning_rate": 3.27463012082404e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 11480864,
      "step": 17545
    },
    {
      "epoch": 9.19811320754717,
      "grad_norm": 0.7572346329689026,
      "learning_rate": 3.2735428854255416e-05,
      "loss": 0.3634,
      "num_input_tokens_seen": 11483776,
      "step": 17550
    },
    {
      "epoch": 9.200733752620545,
      "grad_norm": 0.4577483832836151,
      "learning_rate": 3.272455488210697e-05,
      "loss": 0.4427,
      "num_input_tokens_seen": 11487680,
      "step": 17555
    },
    {
      "epoch": 9.20335429769392,
      "grad_norm": 0.34481316804885864,
      "learning_rate": 3.271367929406977e-05,
      "loss": 0.4735,
      "num_input_tokens_seen": 11491072,
      "step": 17560
    },
    {
      "epoch": 9.205974842767295,
      "grad_norm": 0.3710155785083771,
      "learning_rate": 3.270280209241886e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 11494432,
      "step": 17565
    },
    {
      "epoch": 9.20859538784067,
      "grad_norm": 0.5538485050201416,
      "learning_rate": 3.2691923279429624e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 11498144,
      "step": 17570
    },
    {
      "epoch": 9.211215932914046,
      "grad_norm": 0.26831573247909546,
      "learning_rate": 3.2681042857377804e-05,
      "loss": 0.2849,
      "num_input_tokens_seen": 11501728,
      "step": 17575
    },
    {
      "epoch": 9.213836477987421,
      "grad_norm": 0.4101082384586334,
      "learning_rate": 3.2670160828539435e-05,
      "loss": 0.477,
      "num_input_tokens_seen": 11505088,
      "step": 17580
    },
    {
      "epoch": 9.216457023060796,
      "grad_norm": 0.5259235501289368,
      "learning_rate": 3.2659277195190934e-05,
      "loss": 0.5628,
      "num_input_tokens_seen": 11507712,
      "step": 17585
    },
    {
      "epoch": 9.219077568134171,
      "grad_norm": 0.8490913510322571,
      "learning_rate": 3.2648391959609026e-05,
      "loss": 0.3999,
      "num_input_tokens_seen": 11510560,
      "step": 17590
    },
    {
      "epoch": 9.221698113207546,
      "grad_norm": 0.3660670816898346,
      "learning_rate": 3.263750512407078e-05,
      "loss": 0.5584,
      "num_input_tokens_seen": 11513760,
      "step": 17595
    },
    {
      "epoch": 9.224318658280922,
      "grad_norm": 0.38422271609306335,
      "learning_rate": 3.262661669085359e-05,
      "loss": 0.5175,
      "num_input_tokens_seen": 11516288,
      "step": 17600
    },
    {
      "epoch": 9.226939203354299,
      "grad_norm": 0.35639744997024536,
      "learning_rate": 3.2615726662235205e-05,
      "loss": 0.3669,
      "num_input_tokens_seen": 11522496,
      "step": 17605
    },
    {
      "epoch": 9.229559748427674,
      "grad_norm": 0.5644003748893738,
      "learning_rate": 3.2604835040493676e-05,
      "loss": 0.7075,
      "num_input_tokens_seen": 11525216,
      "step": 17610
    },
    {
      "epoch": 9.232180293501049,
      "grad_norm": 0.5417116284370422,
      "learning_rate": 3.259394182790742e-05,
      "loss": 0.4859,
      "num_input_tokens_seen": 11527552,
      "step": 17615
    },
    {
      "epoch": 9.234800838574424,
      "grad_norm": 0.29635176062583923,
      "learning_rate": 3.2583047026755174e-05,
      "loss": 0.4375,
      "num_input_tokens_seen": 11530880,
      "step": 17620
    },
    {
      "epoch": 9.2374213836478,
      "grad_norm": 0.4968317151069641,
      "learning_rate": 3.2572150639316e-05,
      "loss": 0.509,
      "num_input_tokens_seen": 11534208,
      "step": 17625
    },
    {
      "epoch": 9.240041928721174,
      "grad_norm": 0.3171270787715912,
      "learning_rate": 3.256125266786931e-05,
      "loss": 0.4197,
      "num_input_tokens_seen": 11537632,
      "step": 17630
    },
    {
      "epoch": 9.24266247379455,
      "grad_norm": 0.4076448976993561,
      "learning_rate": 3.255035311469481e-05,
      "loss": 0.4488,
      "num_input_tokens_seen": 11541024,
      "step": 17635
    },
    {
      "epoch": 9.245283018867925,
      "grad_norm": 0.5844096541404724,
      "learning_rate": 3.2539451982072586e-05,
      "loss": 0.5635,
      "num_input_tokens_seen": 11544192,
      "step": 17640
    },
    {
      "epoch": 9.2479035639413,
      "grad_norm": 0.717190682888031,
      "learning_rate": 3.2528549272283016e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 11547840,
      "step": 17645
    },
    {
      "epoch": 9.250524109014675,
      "grad_norm": 0.2770443260669708,
      "learning_rate": 3.251764498760683e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 11551552,
      "step": 17650
    },
    {
      "epoch": 9.25314465408805,
      "grad_norm": 0.5947600603103638,
      "learning_rate": 3.2506739130325084e-05,
      "loss": 0.604,
      "num_input_tokens_seen": 11554880,
      "step": 17655
    },
    {
      "epoch": 9.255765199161425,
      "grad_norm": 0.3210661709308624,
      "learning_rate": 3.2495831702719146e-05,
      "loss": 0.4004,
      "num_input_tokens_seen": 11557472,
      "step": 17660
    },
    {
      "epoch": 9.2583857442348,
      "grad_norm": 0.36663955450057983,
      "learning_rate": 3.248492270707074e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 11560992,
      "step": 17665
    },
    {
      "epoch": 9.261006289308176,
      "grad_norm": 0.35526975989341736,
      "learning_rate": 3.247401214566189e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 11564480,
      "step": 17670
    },
    {
      "epoch": 9.26362683438155,
      "grad_norm": 0.3509760797023773,
      "learning_rate": 3.2463100020774965e-05,
      "loss": 0.3903,
      "num_input_tokens_seen": 11567936,
      "step": 17675
    },
    {
      "epoch": 9.266247379454926,
      "grad_norm": 0.6588388085365295,
      "learning_rate": 3.245218633469267e-05,
      "loss": 0.4553,
      "num_input_tokens_seen": 11571360,
      "step": 17680
    },
    {
      "epoch": 9.268867924528301,
      "grad_norm": 0.5843781232833862,
      "learning_rate": 3.244127108969801e-05,
      "loss": 0.6006,
      "num_input_tokens_seen": 11574336,
      "step": 17685
    },
    {
      "epoch": 9.271488469601676,
      "grad_norm": 0.5423325300216675,
      "learning_rate": 3.2430354288074314e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 11576928,
      "step": 17690
    },
    {
      "epoch": 9.274109014675052,
      "grad_norm": 0.8590419888496399,
      "learning_rate": 3.241943593210529e-05,
      "loss": 0.4669,
      "num_input_tokens_seen": 11580096,
      "step": 17695
    },
    {
      "epoch": 9.276729559748428,
      "grad_norm": 0.5565742254257202,
      "learning_rate": 3.2408516024074916e-05,
      "loss": 0.3826,
      "num_input_tokens_seen": 11583872,
      "step": 17700
    },
    {
      "epoch": 9.279350104821804,
      "grad_norm": 0.4636266231536865,
      "learning_rate": 3.2397594566267505e-05,
      "loss": 0.516,
      "num_input_tokens_seen": 11586496,
      "step": 17705
    },
    {
      "epoch": 9.281970649895179,
      "grad_norm": 0.5642245411872864,
      "learning_rate": 3.23866715609677e-05,
      "loss": 0.5187,
      "num_input_tokens_seen": 11589280,
      "step": 17710
    },
    {
      "epoch": 9.284591194968554,
      "grad_norm": 0.4381336271762848,
      "learning_rate": 3.237574701046049e-05,
      "loss": 0.5083,
      "num_input_tokens_seen": 11592832,
      "step": 17715
    },
    {
      "epoch": 9.28721174004193,
      "grad_norm": 0.29813337326049805,
      "learning_rate": 3.236482091703115e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 11595904,
      "step": 17720
    },
    {
      "epoch": 9.289832285115304,
      "grad_norm": 0.5857152938842773,
      "learning_rate": 3.235389328296528e-05,
      "loss": 0.4343,
      "num_input_tokens_seen": 11600672,
      "step": 17725
    },
    {
      "epoch": 9.29245283018868,
      "grad_norm": 0.4570995569229126,
      "learning_rate": 3.2342964110548845e-05,
      "loss": 0.4097,
      "num_input_tokens_seen": 11603968,
      "step": 17730
    },
    {
      "epoch": 9.295073375262055,
      "grad_norm": 0.4990711212158203,
      "learning_rate": 3.2332033402068093e-05,
      "loss": 0.4155,
      "num_input_tokens_seen": 11606944,
      "step": 17735
    },
    {
      "epoch": 9.29769392033543,
      "grad_norm": 0.3914327025413513,
      "learning_rate": 3.2321101159809605e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 11610368,
      "step": 17740
    },
    {
      "epoch": 9.300314465408805,
      "grad_norm": 0.5289729237556458,
      "learning_rate": 3.231016738606027e-05,
      "loss": 0.4129,
      "num_input_tokens_seen": 11613120,
      "step": 17745
    },
    {
      "epoch": 9.30293501048218,
      "grad_norm": 0.6464347243309021,
      "learning_rate": 3.2299232083107313e-05,
      "loss": 0.3846,
      "num_input_tokens_seen": 11616352,
      "step": 17750
    },
    {
      "epoch": 9.305555555555555,
      "grad_norm": 0.40100693702697754,
      "learning_rate": 3.228829525323829e-05,
      "loss": 0.4702,
      "num_input_tokens_seen": 11621536,
      "step": 17755
    },
    {
      "epoch": 9.30817610062893,
      "grad_norm": 0.45317646861076355,
      "learning_rate": 3.227735689874104e-05,
      "loss": 0.3988,
      "num_input_tokens_seen": 11624480,
      "step": 17760
    },
    {
      "epoch": 9.310796645702306,
      "grad_norm": 0.43497687578201294,
      "learning_rate": 3.226641702190375e-05,
      "loss": 0.3715,
      "num_input_tokens_seen": 11627360,
      "step": 17765
    },
    {
      "epoch": 9.31341719077568,
      "grad_norm": 0.3601279556751251,
      "learning_rate": 3.225547562501493e-05,
      "loss": 0.3526,
      "num_input_tokens_seen": 11631360,
      "step": 17770
    },
    {
      "epoch": 9.316037735849056,
      "grad_norm": 0.34199467301368713,
      "learning_rate": 3.2244532710363375e-05,
      "loss": 0.4352,
      "num_input_tokens_seen": 11634720,
      "step": 17775
    },
    {
      "epoch": 9.318658280922431,
      "grad_norm": 0.2527521252632141,
      "learning_rate": 3.2233588280238224e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 11638144,
      "step": 17780
    },
    {
      "epoch": 9.321278825995806,
      "grad_norm": 0.5217648148536682,
      "learning_rate": 3.2222642336928944e-05,
      "loss": 0.4276,
      "num_input_tokens_seen": 11640960,
      "step": 17785
    },
    {
      "epoch": 9.323899371069182,
      "grad_norm": 0.6051300168037415,
      "learning_rate": 3.2211694882725275e-05,
      "loss": 0.3428,
      "num_input_tokens_seen": 11644320,
      "step": 17790
    },
    {
      "epoch": 9.326519916142558,
      "grad_norm": 0.35241636633872986,
      "learning_rate": 3.220074591991731e-05,
      "loss": 0.4036,
      "num_input_tokens_seen": 11647008,
      "step": 17795
    },
    {
      "epoch": 9.329140461215934,
      "grad_norm": 0.5731803178787231,
      "learning_rate": 3.218979545079546e-05,
      "loss": 0.4079,
      "num_input_tokens_seen": 11650880,
      "step": 17800
    },
    {
      "epoch": 9.331761006289309,
      "grad_norm": 0.3362880051136017,
      "learning_rate": 3.217884347765042e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 11654336,
      "step": 17805
    },
    {
      "epoch": 9.334381551362684,
      "grad_norm": 0.5783414840698242,
      "learning_rate": 3.216789000277323e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 11657120,
      "step": 17810
    },
    {
      "epoch": 9.33700209643606,
      "grad_norm": 0.3504593074321747,
      "learning_rate": 3.2156935028455224e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 11660480,
      "step": 17815
    },
    {
      "epoch": 9.339622641509434,
      "grad_norm": 0.4146130383014679,
      "learning_rate": 3.2145978556988056e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 11663264,
      "step": 17820
    },
    {
      "epoch": 9.34224318658281,
      "grad_norm": 0.4898262321949005,
      "learning_rate": 3.213502059066371e-05,
      "loss": 0.4094,
      "num_input_tokens_seen": 11665632,
      "step": 17825
    },
    {
      "epoch": 9.344863731656185,
      "grad_norm": 0.3767939507961273,
      "learning_rate": 3.212406113177445e-05,
      "loss": 0.5267,
      "num_input_tokens_seen": 11668544,
      "step": 17830
    },
    {
      "epoch": 9.34748427672956,
      "grad_norm": 0.48472991585731506,
      "learning_rate": 3.211310018261287e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 11672192,
      "step": 17835
    },
    {
      "epoch": 9.350104821802935,
      "grad_norm": 0.5463157296180725,
      "learning_rate": 3.21021377454719e-05,
      "loss": 0.5243,
      "num_input_tokens_seen": 11675104,
      "step": 17840
    },
    {
      "epoch": 9.35272536687631,
      "grad_norm": 0.39088109135627747,
      "learning_rate": 3.209117382264473e-05,
      "loss": 0.4268,
      "num_input_tokens_seen": 11678304,
      "step": 17845
    },
    {
      "epoch": 9.355345911949685,
      "grad_norm": 0.5205724239349365,
      "learning_rate": 3.208020841642489e-05,
      "loss": 0.4408,
      "num_input_tokens_seen": 11681760,
      "step": 17850
    },
    {
      "epoch": 9.35796645702306,
      "grad_norm": 0.9766994714736938,
      "learning_rate": 3.206924152910623e-05,
      "loss": 0.3702,
      "num_input_tokens_seen": 11684480,
      "step": 17855
    },
    {
      "epoch": 9.360587002096436,
      "grad_norm": 0.7556982040405273,
      "learning_rate": 3.205827316298289e-05,
      "loss": 0.4598,
      "num_input_tokens_seen": 11687488,
      "step": 17860
    },
    {
      "epoch": 9.36320754716981,
      "grad_norm": 0.525449275970459,
      "learning_rate": 3.204730332034934e-05,
      "loss": 0.4201,
      "num_input_tokens_seen": 11691456,
      "step": 17865
    },
    {
      "epoch": 9.365828092243186,
      "grad_norm": 0.4095878601074219,
      "learning_rate": 3.2036332003500324e-05,
      "loss": 0.6316,
      "num_input_tokens_seen": 11694016,
      "step": 17870
    },
    {
      "epoch": 9.368448637316561,
      "grad_norm": 0.599380612373352,
      "learning_rate": 3.2025359214730924e-05,
      "loss": 0.4898,
      "num_input_tokens_seen": 11696960,
      "step": 17875
    },
    {
      "epoch": 9.371069182389936,
      "grad_norm": 0.7148800492286682,
      "learning_rate": 3.201438495633654e-05,
      "loss": 0.408,
      "num_input_tokens_seen": 11700128,
      "step": 17880
    },
    {
      "epoch": 9.373689727463312,
      "grad_norm": 0.2500390410423279,
      "learning_rate": 3.200340923061284e-05,
      "loss": 0.392,
      "num_input_tokens_seen": 11704192,
      "step": 17885
    },
    {
      "epoch": 9.376310272536688,
      "grad_norm": 0.5698798894882202,
      "learning_rate": 3.199243203985583e-05,
      "loss": 0.4436,
      "num_input_tokens_seen": 11707552,
      "step": 17890
    },
    {
      "epoch": 9.378930817610064,
      "grad_norm": 0.7167071104049683,
      "learning_rate": 3.1981453386361804e-05,
      "loss": 0.4122,
      "num_input_tokens_seen": 11714304,
      "step": 17895
    },
    {
      "epoch": 9.381551362683439,
      "grad_norm": 0.602136492729187,
      "learning_rate": 3.1970473272427386e-05,
      "loss": 0.4099,
      "num_input_tokens_seen": 11716896,
      "step": 17900
    },
    {
      "epoch": 9.384171907756814,
      "grad_norm": 0.5167325735092163,
      "learning_rate": 3.195949170034948e-05,
      "loss": 0.4243,
      "num_input_tokens_seen": 11719680,
      "step": 17905
    },
    {
      "epoch": 9.38679245283019,
      "grad_norm": 0.34845080971717834,
      "learning_rate": 3.19485086724253e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 11723008,
      "step": 17910
    },
    {
      "epoch": 9.389412997903564,
      "grad_norm": 0.3921818733215332,
      "learning_rate": 3.193752419095239e-05,
      "loss": 0.3543,
      "num_input_tokens_seen": 11726272,
      "step": 17915
    },
    {
      "epoch": 9.39203354297694,
      "grad_norm": 0.3487931489944458,
      "learning_rate": 3.192653825822857e-05,
      "loss": 0.4674,
      "num_input_tokens_seen": 11729664,
      "step": 17920
    },
    {
      "epoch": 9.394654088050315,
      "grad_norm": 0.5336987972259521,
      "learning_rate": 3.1915550876551964e-05,
      "loss": 0.4371,
      "num_input_tokens_seen": 11733056,
      "step": 17925
    },
    {
      "epoch": 9.39727463312369,
      "grad_norm": 1.2550901174545288,
      "learning_rate": 3.190456204822101e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 11736128,
      "step": 17930
    },
    {
      "epoch": 9.399895178197065,
      "grad_norm": 0.6435515880584717,
      "learning_rate": 3.1893571775534445e-05,
      "loss": 0.6499,
      "num_input_tokens_seen": 11739456,
      "step": 17935
    },
    {
      "epoch": 9.40251572327044,
      "grad_norm": 0.4244025945663452,
      "learning_rate": 3.1882580060791315e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 11744256,
      "step": 17940
    },
    {
      "epoch": 9.405136268343815,
      "grad_norm": 0.4583694338798523,
      "learning_rate": 3.187158690629095e-05,
      "loss": 0.4203,
      "num_input_tokens_seen": 11746976,
      "step": 17945
    },
    {
      "epoch": 9.40775681341719,
      "grad_norm": 0.25464096665382385,
      "learning_rate": 3.1860592314333e-05,
      "loss": 0.5205,
      "num_input_tokens_seen": 11749920,
      "step": 17950
    },
    {
      "epoch": 9.410377358490566,
      "grad_norm": 0.5436143279075623,
      "learning_rate": 3.184959628721742e-05,
      "loss": 0.482,
      "num_input_tokens_seen": 11752896,
      "step": 17955
    },
    {
      "epoch": 9.41299790356394,
      "grad_norm": 0.5933201313018799,
      "learning_rate": 3.183859882724443e-05,
      "loss": 0.5034,
      "num_input_tokens_seen": 11755424,
      "step": 17960
    },
    {
      "epoch": 9.415618448637316,
      "grad_norm": 0.5078055262565613,
      "learning_rate": 3.1827599936714584e-05,
      "loss": 0.4415,
      "num_input_tokens_seen": 11758976,
      "step": 17965
    },
    {
      "epoch": 9.418238993710691,
      "grad_norm": 0.8981307744979858,
      "learning_rate": 3.1816599617928736e-05,
      "loss": 0.3751,
      "num_input_tokens_seen": 11761248,
      "step": 17970
    },
    {
      "epoch": 9.420859538784066,
      "grad_norm": 0.5362889170646667,
      "learning_rate": 3.1805597873188e-05,
      "loss": 0.4053,
      "num_input_tokens_seen": 11764288,
      "step": 17975
    },
    {
      "epoch": 9.423480083857442,
      "grad_norm": 0.4585341215133667,
      "learning_rate": 3.179459470479384e-05,
      "loss": 0.6655,
      "num_input_tokens_seen": 11767328,
      "step": 17980
    },
    {
      "epoch": 9.426100628930818,
      "grad_norm": 0.3837202489376068,
      "learning_rate": 3.1783590115047983e-05,
      "loss": 0.4101,
      "num_input_tokens_seen": 11770656,
      "step": 17985
    },
    {
      "epoch": 9.428721174004194,
      "grad_norm": 0.404186874628067,
      "learning_rate": 3.177258410625246e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 11773760,
      "step": 17990
    },
    {
      "epoch": 9.431341719077569,
      "grad_norm": 0.32984134554862976,
      "learning_rate": 3.176157668070962e-05,
      "loss": 0.4275,
      "num_input_tokens_seen": 11776480,
      "step": 17995
    },
    {
      "epoch": 9.433962264150944,
      "grad_norm": 0.48023608326911926,
      "learning_rate": 3.175056784072207e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 11780160,
      "step": 18000
    },
    {
      "epoch": 9.43658280922432,
      "grad_norm": 0.3795560896396637,
      "learning_rate": 3.173955758859274e-05,
      "loss": 0.3802,
      "num_input_tokens_seen": 11783200,
      "step": 18005
    },
    {
      "epoch": 9.439203354297694,
      "grad_norm": 0.43229979276657104,
      "learning_rate": 3.172854592662486e-05,
      "loss": 0.4203,
      "num_input_tokens_seen": 11786528,
      "step": 18010
    },
    {
      "epoch": 9.44182389937107,
      "grad_norm": 0.47908568382263184,
      "learning_rate": 3.1717532857121926e-05,
      "loss": 0.426,
      "num_input_tokens_seen": 11789280,
      "step": 18015
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 0.7150943279266357,
      "learning_rate": 3.170651838238776e-05,
      "loss": 0.5112,
      "num_input_tokens_seen": 11792608,
      "step": 18020
    },
    {
      "epoch": 9.44706498951782,
      "grad_norm": 0.5199384093284607,
      "learning_rate": 3.169550250472646e-05,
      "loss": 0.5371,
      "num_input_tokens_seen": 11794944,
      "step": 18025
    },
    {
      "epoch": 9.449685534591195,
      "grad_norm": 0.6714401245117188,
      "learning_rate": 3.168448522644243e-05,
      "loss": 0.3916,
      "num_input_tokens_seen": 11798208,
      "step": 18030
    },
    {
      "epoch": 9.45230607966457,
      "grad_norm": 0.26504606008529663,
      "learning_rate": 3.167346654984034e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 11801440,
      "step": 18035
    },
    {
      "epoch": 9.454926624737945,
      "grad_norm": 0.9011780619621277,
      "learning_rate": 3.166244647722519e-05,
      "loss": 0.475,
      "num_input_tokens_seen": 11804192,
      "step": 18040
    },
    {
      "epoch": 9.45754716981132,
      "grad_norm": 0.4717022776603699,
      "learning_rate": 3.1651425010902244e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 11807584,
      "step": 18045
    },
    {
      "epoch": 9.460167714884696,
      "grad_norm": 0.3545355498790741,
      "learning_rate": 3.164040215317706e-05,
      "loss": 0.4386,
      "num_input_tokens_seen": 11811392,
      "step": 18050
    },
    {
      "epoch": 9.46278825995807,
      "grad_norm": 0.3346727788448334,
      "learning_rate": 3.162937790635552e-05,
      "loss": 0.4807,
      "num_input_tokens_seen": 11815872,
      "step": 18055
    },
    {
      "epoch": 9.465408805031446,
      "grad_norm": 0.3426615297794342,
      "learning_rate": 3.1618352272743746e-05,
      "loss": 0.3295,
      "num_input_tokens_seen": 11818784,
      "step": 18060
    },
    {
      "epoch": 9.468029350104821,
      "grad_norm": 1.775668978691101,
      "learning_rate": 3.160732525464818e-05,
      "loss": 0.5597,
      "num_input_tokens_seen": 11822624,
      "step": 18065
    },
    {
      "epoch": 9.470649895178196,
      "grad_norm": 0.4222603142261505,
      "learning_rate": 3.1596296854375554e-05,
      "loss": 0.4868,
      "num_input_tokens_seen": 11825312,
      "step": 18070
    },
    {
      "epoch": 9.473270440251572,
      "grad_norm": 0.8157133460044861,
      "learning_rate": 3.1585267074232884e-05,
      "loss": 0.4947,
      "num_input_tokens_seen": 11828224,
      "step": 18075
    },
    {
      "epoch": 9.475890985324948,
      "grad_norm": 0.4632703363895416,
      "learning_rate": 3.157423591652746e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 11831072,
      "step": 18080
    },
    {
      "epoch": 9.478511530398324,
      "grad_norm": 0.7117603421211243,
      "learning_rate": 3.1563203383566894e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 11835168,
      "step": 18085
    },
    {
      "epoch": 9.481132075471699,
      "grad_norm": 0.6109200119972229,
      "learning_rate": 3.155216947765906e-05,
      "loss": 0.5049,
      "num_input_tokens_seen": 11838048,
      "step": 18090
    },
    {
      "epoch": 9.483752620545074,
      "grad_norm": 0.6325680017471313,
      "learning_rate": 3.154113420111211e-05,
      "loss": 0.5946,
      "num_input_tokens_seen": 11842048,
      "step": 18095
    },
    {
      "epoch": 9.48637316561845,
      "grad_norm": 0.35772964358329773,
      "learning_rate": 3.1530097556234526e-05,
      "loss": 0.4851,
      "num_input_tokens_seen": 11845856,
      "step": 18100
    },
    {
      "epoch": 9.488993710691824,
      "grad_norm": 0.3368600904941559,
      "learning_rate": 3.151905954533502e-05,
      "loss": 0.4307,
      "num_input_tokens_seen": 11849152,
      "step": 18105
    },
    {
      "epoch": 9.4916142557652,
      "grad_norm": 0.3751490116119385,
      "learning_rate": 3.150802017072263e-05,
      "loss": 0.4267,
      "num_input_tokens_seen": 11852096,
      "step": 18110
    },
    {
      "epoch": 9.494234800838575,
      "grad_norm": 0.38126879930496216,
      "learning_rate": 3.149697943470668e-05,
      "loss": 0.387,
      "num_input_tokens_seen": 11854528,
      "step": 18115
    },
    {
      "epoch": 9.49685534591195,
      "grad_norm": 0.3921354413032532,
      "learning_rate": 3.1485937339596736e-05,
      "loss": 0.5693,
      "num_input_tokens_seen": 11857440,
      "step": 18120
    },
    {
      "epoch": 9.499475890985325,
      "grad_norm": 0.6685483455657959,
      "learning_rate": 3.147489388770271e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 11860192,
      "step": 18125
    },
    {
      "epoch": 9.5020964360587,
      "grad_norm": 0.3702546954154968,
      "learning_rate": 3.146384908133475e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 11863104,
      "step": 18130
    },
    {
      "epoch": 9.504716981132075,
      "grad_norm": 0.6269749402999878,
      "learning_rate": 3.14528029228033e-05,
      "loss": 0.5117,
      "num_input_tokens_seen": 11865792,
      "step": 18135
    },
    {
      "epoch": 9.50733752620545,
      "grad_norm": 0.5899245142936707,
      "learning_rate": 3.144175541441911e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 11868384,
      "step": 18140
    },
    {
      "epoch": 9.509958071278826,
      "grad_norm": 0.2522147595882416,
      "learning_rate": 3.1430706558493166e-05,
      "loss": 0.3968,
      "num_input_tokens_seen": 11871808,
      "step": 18145
    },
    {
      "epoch": 9.5125786163522,
      "grad_norm": 0.6582757234573364,
      "learning_rate": 3.141965635733677e-05,
      "loss": 0.5768,
      "num_input_tokens_seen": 11874368,
      "step": 18150
    },
    {
      "epoch": 9.515199161425576,
      "grad_norm": 0.45481711626052856,
      "learning_rate": 3.140860481326152e-05,
      "loss": 0.3929,
      "num_input_tokens_seen": 11877024,
      "step": 18155
    },
    {
      "epoch": 9.517819706498951,
      "grad_norm": 0.45359334349632263,
      "learning_rate": 3.139755192857925e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 11880224,
      "step": 18160
    },
    {
      "epoch": 9.520440251572326,
      "grad_norm": 0.5017049908638,
      "learning_rate": 3.1386497705602094e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 11883712,
      "step": 18165
    },
    {
      "epoch": 9.523060796645701,
      "grad_norm": 0.31332674622535706,
      "learning_rate": 3.137544214664249e-05,
      "loss": 0.3502,
      "num_input_tokens_seen": 11888384,
      "step": 18170
    },
    {
      "epoch": 9.525681341719078,
      "grad_norm": 0.37491583824157715,
      "learning_rate": 3.136438525401312e-05,
      "loss": 0.4458,
      "num_input_tokens_seen": 11891680,
      "step": 18175
    },
    {
      "epoch": 9.528301886792454,
      "grad_norm": 0.580970287322998,
      "learning_rate": 3.135332703002696e-05,
      "loss": 0.3958,
      "num_input_tokens_seen": 11894592,
      "step": 18180
    },
    {
      "epoch": 9.530922431865829,
      "grad_norm": 0.36352744698524475,
      "learning_rate": 3.1342267476997276e-05,
      "loss": 0.4497,
      "num_input_tokens_seen": 11897536,
      "step": 18185
    },
    {
      "epoch": 9.533542976939204,
      "grad_norm": 0.31742823123931885,
      "learning_rate": 3.1331206597237584e-05,
      "loss": 0.4362,
      "num_input_tokens_seen": 11901440,
      "step": 18190
    },
    {
      "epoch": 9.536163522012579,
      "grad_norm": 0.4281802475452423,
      "learning_rate": 3.1320144393061715e-05,
      "loss": 0.5445,
      "num_input_tokens_seen": 11904736,
      "step": 18195
    },
    {
      "epoch": 9.538784067085954,
      "grad_norm": 0.3421385586261749,
      "learning_rate": 3.130908086678373e-05,
      "loss": 0.3938,
      "num_input_tokens_seen": 11907936,
      "step": 18200
    },
    {
      "epoch": 9.54140461215933,
      "grad_norm": 0.5097522139549255,
      "learning_rate": 3.1298016020718014e-05,
      "loss": 0.4825,
      "num_input_tokens_seen": 11911072,
      "step": 18205
    },
    {
      "epoch": 9.544025157232705,
      "grad_norm": 0.4139309525489807,
      "learning_rate": 3.128694985717919e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 11914432,
      "step": 18210
    },
    {
      "epoch": 9.54664570230608,
      "grad_norm": 1.1372463703155518,
      "learning_rate": 3.1275882378482185e-05,
      "loss": 0.5216,
      "num_input_tokens_seen": 11917344,
      "step": 18215
    },
    {
      "epoch": 9.549266247379455,
      "grad_norm": 0.2943410277366638,
      "learning_rate": 3.1264813586942185e-05,
      "loss": 0.3413,
      "num_input_tokens_seen": 11919968,
      "step": 18220
    },
    {
      "epoch": 9.55188679245283,
      "grad_norm": 0.38445761799812317,
      "learning_rate": 3.125374348487466e-05,
      "loss": 0.3831,
      "num_input_tokens_seen": 11923520,
      "step": 18225
    },
    {
      "epoch": 9.554507337526205,
      "grad_norm": 0.8029131293296814,
      "learning_rate": 3.1242672074595334e-05,
      "loss": 0.4808,
      "num_input_tokens_seen": 11926304,
      "step": 18230
    },
    {
      "epoch": 9.55712788259958,
      "grad_norm": 0.3255026042461395,
      "learning_rate": 3.123159935842024e-05,
      "loss": 0.5662,
      "num_input_tokens_seen": 11929248,
      "step": 18235
    },
    {
      "epoch": 9.559748427672956,
      "grad_norm": 0.5072985887527466,
      "learning_rate": 3.122052533866565e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 11934464,
      "step": 18240
    },
    {
      "epoch": 9.56236897274633,
      "grad_norm": 0.40655434131622314,
      "learning_rate": 3.120945001764813e-05,
      "loss": 0.4467,
      "num_input_tokens_seen": 11938560,
      "step": 18245
    },
    {
      "epoch": 9.564989517819706,
      "grad_norm": 0.5063691735267639,
      "learning_rate": 3.11983733976845e-05,
      "loss": 0.355,
      "num_input_tokens_seen": 11941472,
      "step": 18250
    },
    {
      "epoch": 9.567610062893081,
      "grad_norm": 0.3542573153972626,
      "learning_rate": 3.1187295481091876e-05,
      "loss": 0.5255,
      "num_input_tokens_seen": 11944320,
      "step": 18255
    },
    {
      "epoch": 9.570230607966456,
      "grad_norm": 0.6983438730239868,
      "learning_rate": 3.1176216270187625e-05,
      "loss": 0.3826,
      "num_input_tokens_seen": 11947808,
      "step": 18260
    },
    {
      "epoch": 9.572851153039831,
      "grad_norm": 0.5562120079994202,
      "learning_rate": 3.116513576728939e-05,
      "loss": 0.5537,
      "num_input_tokens_seen": 11950944,
      "step": 18265
    },
    {
      "epoch": 9.575471698113208,
      "grad_norm": 0.6386443376541138,
      "learning_rate": 3.1154053974715085e-05,
      "loss": 0.545,
      "num_input_tokens_seen": 11954208,
      "step": 18270
    },
    {
      "epoch": 9.578092243186584,
      "grad_norm": 0.2532491683959961,
      "learning_rate": 3.1142970894782906e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 11957728,
      "step": 18275
    },
    {
      "epoch": 9.580712788259959,
      "grad_norm": 0.43135127425193787,
      "learning_rate": 3.1131886529811296e-05,
      "loss": 0.6569,
      "num_input_tokens_seen": 11960544,
      "step": 18280
    },
    {
      "epoch": 9.583333333333334,
      "grad_norm": 0.4179301857948303,
      "learning_rate": 3.1120800882118985e-05,
      "loss": 0.4696,
      "num_input_tokens_seen": 11963904,
      "step": 18285
    },
    {
      "epoch": 9.585953878406709,
      "grad_norm": 0.4840657711029053,
      "learning_rate": 3.1109713954024945e-05,
      "loss": 0.4747,
      "num_input_tokens_seen": 11966784,
      "step": 18290
    },
    {
      "epoch": 9.588574423480084,
      "grad_norm": 0.32268115878105164,
      "learning_rate": 3.109862574784846e-05,
      "loss": 0.3881,
      "num_input_tokens_seen": 11970240,
      "step": 18295
    },
    {
      "epoch": 9.59119496855346,
      "grad_norm": 0.3665376901626587,
      "learning_rate": 3.108753626590905e-05,
      "loss": 0.4108,
      "num_input_tokens_seen": 11973760,
      "step": 18300
    },
    {
      "epoch": 9.593815513626835,
      "grad_norm": 0.5030189752578735,
      "learning_rate": 3.1076445510526486e-05,
      "loss": 0.4299,
      "num_input_tokens_seen": 11976512,
      "step": 18305
    },
    {
      "epoch": 9.59643605870021,
      "grad_norm": 0.35178500413894653,
      "learning_rate": 3.106535348402086e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 11979904,
      "step": 18310
    },
    {
      "epoch": 9.599056603773585,
      "grad_norm": 0.3283143937587738,
      "learning_rate": 3.105426018871247e-05,
      "loss": 0.5219,
      "num_input_tokens_seen": 11983232,
      "step": 18315
    },
    {
      "epoch": 9.60167714884696,
      "grad_norm": 0.6276537179946899,
      "learning_rate": 3.104316562692193e-05,
      "loss": 0.3297,
      "num_input_tokens_seen": 11987392,
      "step": 18320
    },
    {
      "epoch": 9.604297693920335,
      "grad_norm": 0.2980533838272095,
      "learning_rate": 3.103206980097009e-05,
      "loss": 0.3402,
      "num_input_tokens_seen": 11991712,
      "step": 18325
    },
    {
      "epoch": 9.60691823899371,
      "grad_norm": 0.7079888582229614,
      "learning_rate": 3.102097271317805e-05,
      "loss": 0.4348,
      "num_input_tokens_seen": 11995040,
      "step": 18330
    },
    {
      "epoch": 9.609538784067086,
      "grad_norm": 0.488052099943161,
      "learning_rate": 3.100987436586721e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 11998048,
      "step": 18335
    },
    {
      "epoch": 9.61215932914046,
      "grad_norm": 0.11747413128614426,
      "learning_rate": 3.099877476135923e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 12003200,
      "step": 18340
    },
    {
      "epoch": 9.614779874213836,
      "grad_norm": 0.615714967250824,
      "learning_rate": 3.0987673901976e-05,
      "loss": 0.4286,
      "num_input_tokens_seen": 12005728,
      "step": 18345
    },
    {
      "epoch": 9.617400419287211,
      "grad_norm": 0.34503141045570374,
      "learning_rate": 3.0976571790039697e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 12009280,
      "step": 18350
    },
    {
      "epoch": 9.620020964360586,
      "grad_norm": 0.4723772704601288,
      "learning_rate": 3.096546842787277e-05,
      "loss": 0.5087,
      "num_input_tokens_seen": 12013216,
      "step": 18355
    },
    {
      "epoch": 9.622641509433961,
      "grad_norm": 0.5924317836761475,
      "learning_rate": 3.095436381779789e-05,
      "loss": 0.4254,
      "num_input_tokens_seen": 12015712,
      "step": 18360
    },
    {
      "epoch": 9.625262054507338,
      "grad_norm": 0.4091378152370453,
      "learning_rate": 3.0943257962138045e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 12018688,
      "step": 18365
    },
    {
      "epoch": 9.627882599580714,
      "grad_norm": 0.4559256136417389,
      "learning_rate": 3.093215086321644e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 12022464,
      "step": 18370
    },
    {
      "epoch": 9.630503144654089,
      "grad_norm": 0.6663120985031128,
      "learning_rate": 3.092104252335654e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 12025120,
      "step": 18375
    },
    {
      "epoch": 9.633123689727464,
      "grad_norm": 0.47406384348869324,
      "learning_rate": 3.09099329448821e-05,
      "loss": 0.4821,
      "num_input_tokens_seen": 12028064,
      "step": 18380
    },
    {
      "epoch": 9.635744234800839,
      "grad_norm": 0.40018388628959656,
      "learning_rate": 3.089882213011711e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 12031776,
      "step": 18385
    },
    {
      "epoch": 9.638364779874214,
      "grad_norm": 0.43959254026412964,
      "learning_rate": 3.088771008138583e-05,
      "loss": 0.5232,
      "num_input_tokens_seen": 12035008,
      "step": 18390
    },
    {
      "epoch": 9.64098532494759,
      "grad_norm": 0.3610750734806061,
      "learning_rate": 3.087659680101278e-05,
      "loss": 0.4889,
      "num_input_tokens_seen": 12039648,
      "step": 18395
    },
    {
      "epoch": 9.643605870020965,
      "grad_norm": 0.44094377756118774,
      "learning_rate": 3.086548229132271e-05,
      "loss": 0.5608,
      "num_input_tokens_seen": 12042304,
      "step": 18400
    },
    {
      "epoch": 9.64622641509434,
      "grad_norm": 0.48359397053718567,
      "learning_rate": 3.085436655464067e-05,
      "loss": 0.5675,
      "num_input_tokens_seen": 12045312,
      "step": 18405
    },
    {
      "epoch": 9.648846960167715,
      "grad_norm": 0.5100971460342407,
      "learning_rate": 3.084324959329193e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 12047936,
      "step": 18410
    },
    {
      "epoch": 9.65146750524109,
      "grad_norm": 0.3514813184738159,
      "learning_rate": 3.083213140960204e-05,
      "loss": 0.3734,
      "num_input_tokens_seen": 12050592,
      "step": 18415
    },
    {
      "epoch": 9.654088050314465,
      "grad_norm": 1.150041103363037,
      "learning_rate": 3.08210120058968e-05,
      "loss": 0.6128,
      "num_input_tokens_seen": 12053312,
      "step": 18420
    },
    {
      "epoch": 9.65670859538784,
      "grad_norm": 0.5186863541603088,
      "learning_rate": 3.080989138450225e-05,
      "loss": 0.4233,
      "num_input_tokens_seen": 12055520,
      "step": 18425
    },
    {
      "epoch": 9.659329140461216,
      "grad_norm": 0.4781563878059387,
      "learning_rate": 3.0798769547744713e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 12058464,
      "step": 18430
    },
    {
      "epoch": 9.66194968553459,
      "grad_norm": 0.5560152530670166,
      "learning_rate": 3.0787646497950736e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 12061152,
      "step": 18435
    },
    {
      "epoch": 9.664570230607966,
      "grad_norm": 0.510555624961853,
      "learning_rate": 3.0776522237447146e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 12063968,
      "step": 18440
    },
    {
      "epoch": 9.667190775681341,
      "grad_norm": 0.5402754545211792,
      "learning_rate": 3.076539676856101e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 12066816,
      "step": 18445
    },
    {
      "epoch": 9.669811320754716,
      "grad_norm": 0.5382971167564392,
      "learning_rate": 3.0754270093619634e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 12069536,
      "step": 18450
    },
    {
      "epoch": 9.672431865828091,
      "grad_norm": 1.382445216178894,
      "learning_rate": 3.074314221495061e-05,
      "loss": 0.6904,
      "num_input_tokens_seen": 12072608,
      "step": 18455
    },
    {
      "epoch": 9.675052410901468,
      "grad_norm": 0.4658169448375702,
      "learning_rate": 3.0732013134881756e-05,
      "loss": 0.332,
      "num_input_tokens_seen": 12076576,
      "step": 18460
    },
    {
      "epoch": 9.677672955974844,
      "grad_norm": 0.7990968227386475,
      "learning_rate": 3.072088285574114e-05,
      "loss": 0.4953,
      "num_input_tokens_seen": 12079712,
      "step": 18465
    },
    {
      "epoch": 9.680293501048219,
      "grad_norm": 0.500979483127594,
      "learning_rate": 3.070975137985711e-05,
      "loss": 0.4306,
      "num_input_tokens_seen": 12082944,
      "step": 18470
    },
    {
      "epoch": 9.682914046121594,
      "grad_norm": 0.39447230100631714,
      "learning_rate": 3.0698618709558226e-05,
      "loss": 0.5486,
      "num_input_tokens_seen": 12086752,
      "step": 18475
    },
    {
      "epoch": 9.685534591194969,
      "grad_norm": 0.5341734886169434,
      "learning_rate": 3.068748484717332e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 12089344,
      "step": 18480
    },
    {
      "epoch": 9.688155136268344,
      "grad_norm": 0.20310825109481812,
      "learning_rate": 3.067634979503147e-05,
      "loss": 0.5138,
      "num_input_tokens_seen": 12093024,
      "step": 18485
    },
    {
      "epoch": 9.69077568134172,
      "grad_norm": 0.33693090081214905,
      "learning_rate": 3.0665213555462e-05,
      "loss": 0.6006,
      "num_input_tokens_seen": 12096800,
      "step": 18490
    },
    {
      "epoch": 9.693396226415095,
      "grad_norm": 0.6451184749603271,
      "learning_rate": 3.0654076130794494e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 12099712,
      "step": 18495
    },
    {
      "epoch": 9.69601677148847,
      "grad_norm": 0.6302061676979065,
      "learning_rate": 3.064293752335876e-05,
      "loss": 0.5123,
      "num_input_tokens_seen": 12103328,
      "step": 18500
    },
    {
      "epoch": 9.698637316561845,
      "grad_norm": 0.35221362113952637,
      "learning_rate": 3.063179773548487e-05,
      "loss": 0.3794,
      "num_input_tokens_seen": 12106368,
      "step": 18505
    },
    {
      "epoch": 9.70125786163522,
      "grad_norm": 0.23143291473388672,
      "learning_rate": 3.0620656769503164e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 12110528,
      "step": 18510
    },
    {
      "epoch": 9.703878406708595,
      "grad_norm": 0.4892314076423645,
      "learning_rate": 3.060951462774417e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 12113312,
      "step": 18515
    },
    {
      "epoch": 9.70649895178197,
      "grad_norm": 0.6644513010978699,
      "learning_rate": 3.0598371312538714e-05,
      "loss": 0.349,
      "num_input_tokens_seen": 12116736,
      "step": 18520
    },
    {
      "epoch": 9.709119496855346,
      "grad_norm": 0.5432233810424805,
      "learning_rate": 3.058722682621785e-05,
      "loss": 0.3649,
      "num_input_tokens_seen": 12120064,
      "step": 18525
    },
    {
      "epoch": 9.71174004192872,
      "grad_norm": 0.4965478181838989,
      "learning_rate": 3.0576081171112875e-05,
      "loss": 0.4669,
      "num_input_tokens_seen": 12123200,
      "step": 18530
    },
    {
      "epoch": 9.714360587002096,
      "grad_norm": 0.7132389545440674,
      "learning_rate": 3.056493434955534e-05,
      "loss": 0.5942,
      "num_input_tokens_seen": 12126784,
      "step": 18535
    },
    {
      "epoch": 9.716981132075471,
      "grad_norm": 1.0726045370101929,
      "learning_rate": 3.0553786363877027e-05,
      "loss": 0.4096,
      "num_input_tokens_seen": 12129280,
      "step": 18540
    },
    {
      "epoch": 9.719601677148846,
      "grad_norm": 0.36857298016548157,
      "learning_rate": 3.054263721640996e-05,
      "loss": 0.5215,
      "num_input_tokens_seen": 12132672,
      "step": 18545
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 0.9210325479507446,
      "learning_rate": 3.053148690948642e-05,
      "loss": 0.4635,
      "num_input_tokens_seen": 12135584,
      "step": 18550
    },
    {
      "epoch": 9.724842767295598,
      "grad_norm": 0.5179809331893921,
      "learning_rate": 3.0520335445438935e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 12137952,
      "step": 18555
    },
    {
      "epoch": 9.727463312368974,
      "grad_norm": 0.3536068797111511,
      "learning_rate": 3.0509182826600246e-05,
      "loss": 0.3478,
      "num_input_tokens_seen": 12141280,
      "step": 18560
    },
    {
      "epoch": 9.730083857442349,
      "grad_norm": 0.7213699221611023,
      "learning_rate": 3.049802905530337e-05,
      "loss": 0.4449,
      "num_input_tokens_seen": 12143872,
      "step": 18565
    },
    {
      "epoch": 9.732704402515724,
      "grad_norm": 0.40474623441696167,
      "learning_rate": 3.0486874133881533e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 12146912,
      "step": 18570
    },
    {
      "epoch": 9.735324947589099,
      "grad_norm": 0.5413623452186584,
      "learning_rate": 3.047571806466823e-05,
      "loss": 0.3612,
      "num_input_tokens_seen": 12150688,
      "step": 18575
    },
    {
      "epoch": 9.737945492662474,
      "grad_norm": 0.38358184695243835,
      "learning_rate": 3.0464560849997177e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 12154048,
      "step": 18580
    },
    {
      "epoch": 9.74056603773585,
      "grad_norm": 0.5148059129714966,
      "learning_rate": 3.0453402492202325e-05,
      "loss": 0.5907,
      "num_input_tokens_seen": 12157312,
      "step": 18585
    },
    {
      "epoch": 9.743186582809225,
      "grad_norm": 0.2469509094953537,
      "learning_rate": 3.0442242993617904e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 12161472,
      "step": 18590
    },
    {
      "epoch": 9.7458071278826,
      "grad_norm": 0.687949538230896,
      "learning_rate": 3.0431082356578334e-05,
      "loss": 0.5239,
      "num_input_tokens_seen": 12164384,
      "step": 18595
    },
    {
      "epoch": 9.748427672955975,
      "grad_norm": 0.4074745178222656,
      "learning_rate": 3.0419920583418294e-05,
      "loss": 0.4445,
      "num_input_tokens_seen": 12167872,
      "step": 18600
    },
    {
      "epoch": 9.75104821802935,
      "grad_norm": 0.2963913679122925,
      "learning_rate": 3.0408757676472694e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 12170752,
      "step": 18605
    },
    {
      "epoch": 9.753668763102725,
      "grad_norm": 0.7448773980140686,
      "learning_rate": 3.0397593638076705e-05,
      "loss": 0.4614,
      "num_input_tokens_seen": 12173728,
      "step": 18610
    },
    {
      "epoch": 9.7562893081761,
      "grad_norm": 0.47396552562713623,
      "learning_rate": 3.0386428470565702e-05,
      "loss": 0.458,
      "num_input_tokens_seen": 12177024,
      "step": 18615
    },
    {
      "epoch": 9.758909853249476,
      "grad_norm": 0.5168859958648682,
      "learning_rate": 3.0375262176275308e-05,
      "loss": 0.5406,
      "num_input_tokens_seen": 12180192,
      "step": 18620
    },
    {
      "epoch": 9.76153039832285,
      "grad_norm": 0.21091541647911072,
      "learning_rate": 3.03640947575414e-05,
      "loss": 0.4331,
      "num_input_tokens_seen": 12183488,
      "step": 18625
    },
    {
      "epoch": 9.764150943396226,
      "grad_norm": 0.39640313386917114,
      "learning_rate": 3.0352926216700067e-05,
      "loss": 0.546,
      "num_input_tokens_seen": 12186880,
      "step": 18630
    },
    {
      "epoch": 9.766771488469601,
      "grad_norm": 0.42172756791114807,
      "learning_rate": 3.034175655608763e-05,
      "loss": 0.359,
      "num_input_tokens_seen": 12189376,
      "step": 18635
    },
    {
      "epoch": 9.769392033542976,
      "grad_norm": 0.8866410255432129,
      "learning_rate": 3.0330585778040677e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 12191616,
      "step": 18640
    },
    {
      "epoch": 9.772012578616351,
      "grad_norm": 0.3067001700401306,
      "learning_rate": 3.031941388489598e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 12195744,
      "step": 18645
    },
    {
      "epoch": 9.774633123689728,
      "grad_norm": 0.4961371123790741,
      "learning_rate": 3.0308240878990584e-05,
      "loss": 0.5048,
      "num_input_tokens_seen": 12198784,
      "step": 18650
    },
    {
      "epoch": 9.777253668763104,
      "grad_norm": 0.5043480396270752,
      "learning_rate": 3.0297066762661773e-05,
      "loss": 0.5066,
      "num_input_tokens_seen": 12201408,
      "step": 18655
    },
    {
      "epoch": 9.779874213836479,
      "grad_norm": 0.38798561692237854,
      "learning_rate": 3.0285891538247015e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 12204448,
      "step": 18660
    },
    {
      "epoch": 9.782494758909854,
      "grad_norm": 0.399795800447464,
      "learning_rate": 3.0274715208084053e-05,
      "loss": 0.4479,
      "num_input_tokens_seen": 12207712,
      "step": 18665
    },
    {
      "epoch": 9.785115303983229,
      "grad_norm": 0.6800791025161743,
      "learning_rate": 3.0263537774510847e-05,
      "loss": 0.4539,
      "num_input_tokens_seen": 12210624,
      "step": 18670
    },
    {
      "epoch": 9.787735849056604,
      "grad_norm": 0.4347088634967804,
      "learning_rate": 3.0252359239865585e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 12214112,
      "step": 18675
    },
    {
      "epoch": 9.79035639412998,
      "grad_norm": 0.2901316285133362,
      "learning_rate": 3.0241179606486704e-05,
      "loss": 0.4631,
      "num_input_tokens_seen": 12218656,
      "step": 18680
    },
    {
      "epoch": 9.792976939203355,
      "grad_norm": 0.418818861246109,
      "learning_rate": 3.0229998876712827e-05,
      "loss": 0.4584,
      "num_input_tokens_seen": 12222112,
      "step": 18685
    },
    {
      "epoch": 9.79559748427673,
      "grad_norm": 0.5722870230674744,
      "learning_rate": 3.0218817052882858e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 12225216,
      "step": 18690
    },
    {
      "epoch": 9.798218029350105,
      "grad_norm": 0.5568426847457886,
      "learning_rate": 3.020763413733591e-05,
      "loss": 0.4547,
      "num_input_tokens_seen": 12228192,
      "step": 18695
    },
    {
      "epoch": 9.80083857442348,
      "grad_norm": 0.3561501204967499,
      "learning_rate": 3.0196450132411302e-05,
      "loss": 0.3515,
      "num_input_tokens_seen": 12232512,
      "step": 18700
    },
    {
      "epoch": 9.803459119496855,
      "grad_norm": 0.5730090141296387,
      "learning_rate": 3.0185265040448616e-05,
      "loss": 0.513,
      "num_input_tokens_seen": 12235584,
      "step": 18705
    },
    {
      "epoch": 9.80607966457023,
      "grad_norm": 1.1191457509994507,
      "learning_rate": 3.017407886378764e-05,
      "loss": 0.4023,
      "num_input_tokens_seen": 12237920,
      "step": 18710
    },
    {
      "epoch": 9.808700209643606,
      "grad_norm": 0.6267119646072388,
      "learning_rate": 3.016289160476839e-05,
      "loss": 0.4755,
      "num_input_tokens_seen": 12240736,
      "step": 18715
    },
    {
      "epoch": 9.81132075471698,
      "grad_norm": 0.44718608260154724,
      "learning_rate": 3.0151703265731125e-05,
      "loss": 0.3107,
      "num_input_tokens_seen": 12243744,
      "step": 18720
    },
    {
      "epoch": 9.813941299790356,
      "grad_norm": 0.47160080075263977,
      "learning_rate": 3.0140513849016305e-05,
      "loss": 0.3665,
      "num_input_tokens_seen": 12248672,
      "step": 18725
    },
    {
      "epoch": 9.816561844863731,
      "grad_norm": 0.3493053615093231,
      "learning_rate": 3.0129323356964627e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 12252224,
      "step": 18730
    },
    {
      "epoch": 9.819182389937106,
      "grad_norm": 0.46962770819664,
      "learning_rate": 3.0118131791917032e-05,
      "loss": 0.5028,
      "num_input_tokens_seen": 12254624,
      "step": 18735
    },
    {
      "epoch": 9.821802935010481,
      "grad_norm": 0.43611806631088257,
      "learning_rate": 3.010693915621466e-05,
      "loss": 0.3832,
      "num_input_tokens_seen": 12258272,
      "step": 18740
    },
    {
      "epoch": 9.824423480083858,
      "grad_norm": 0.8417625427246094,
      "learning_rate": 3.009574545219887e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 12261088,
      "step": 18745
    },
    {
      "epoch": 9.827044025157234,
      "grad_norm": 0.5426942706108093,
      "learning_rate": 3.0084550682211272e-05,
      "loss": 0.5116,
      "num_input_tokens_seen": 12264544,
      "step": 18750
    },
    {
      "epoch": 9.829664570230609,
      "grad_norm": 0.6042071580886841,
      "learning_rate": 3.0073354848593677e-05,
      "loss": 0.579,
      "num_input_tokens_seen": 12267360,
      "step": 18755
    },
    {
      "epoch": 9.832285115303984,
      "grad_norm": 0.5494675040245056,
      "learning_rate": 3.0062157953688125e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 12270720,
      "step": 18760
    },
    {
      "epoch": 9.834905660377359,
      "grad_norm": 0.4273531436920166,
      "learning_rate": 3.0050959999836887e-05,
      "loss": 0.4389,
      "num_input_tokens_seen": 12274112,
      "step": 18765
    },
    {
      "epoch": 9.837526205450734,
      "grad_norm": 0.8614859580993652,
      "learning_rate": 3.003976098938244e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 12277600,
      "step": 18770
    },
    {
      "epoch": 9.84014675052411,
      "grad_norm": 0.47787928581237793,
      "learning_rate": 3.0028560924667487e-05,
      "loss": 0.4981,
      "num_input_tokens_seen": 12280640,
      "step": 18775
    },
    {
      "epoch": 9.842767295597485,
      "grad_norm": 0.5298669934272766,
      "learning_rate": 3.001735980803496e-05,
      "loss": 0.623,
      "num_input_tokens_seen": 12283776,
      "step": 18780
    },
    {
      "epoch": 9.84538784067086,
      "grad_norm": 0.4157264232635498,
      "learning_rate": 3.0006157641828004e-05,
      "loss": 0.327,
      "num_input_tokens_seen": 12287392,
      "step": 18785
    },
    {
      "epoch": 9.848008385744235,
      "grad_norm": 0.3899454176425934,
      "learning_rate": 2.9994954428389977e-05,
      "loss": 0.6231,
      "num_input_tokens_seen": 12290208,
      "step": 18790
    },
    {
      "epoch": 9.85062893081761,
      "grad_norm": 0.3916787803173065,
      "learning_rate": 2.9983750170064472e-05,
      "loss": 0.5133,
      "num_input_tokens_seen": 12293120,
      "step": 18795
    },
    {
      "epoch": 9.853249475890985,
      "grad_norm": 0.3073813319206238,
      "learning_rate": 2.9972544869195292e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 12300096,
      "step": 18800
    },
    {
      "epoch": 9.85587002096436,
      "grad_norm": 0.3110916018486023,
      "learning_rate": 2.9961338528126447e-05,
      "loss": 0.379,
      "num_input_tokens_seen": 12304864,
      "step": 18805
    },
    {
      "epoch": 9.858490566037736,
      "grad_norm": 0.4130687415599823,
      "learning_rate": 2.9950131149202192e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 12308512,
      "step": 18810
    },
    {
      "epoch": 9.86111111111111,
      "grad_norm": 0.49735188484191895,
      "learning_rate": 2.9938922734766973e-05,
      "loss": 0.4523,
      "num_input_tokens_seen": 12311232,
      "step": 18815
    },
    {
      "epoch": 9.863731656184486,
      "grad_norm": 0.4036542773246765,
      "learning_rate": 2.992771328716546e-05,
      "loss": 0.281,
      "num_input_tokens_seen": 12314656,
      "step": 18820
    },
    {
      "epoch": 9.866352201257861,
      "grad_norm": 0.35421496629714966,
      "learning_rate": 2.991650280874255e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 12318336,
      "step": 18825
    },
    {
      "epoch": 9.868972746331236,
      "grad_norm": 0.6371414661407471,
      "learning_rate": 2.990529130184334e-05,
      "loss": 0.3227,
      "num_input_tokens_seen": 12321056,
      "step": 18830
    },
    {
      "epoch": 9.871593291404611,
      "grad_norm": 0.4778088927268982,
      "learning_rate": 2.989407876881316e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 12324640,
      "step": 18835
    },
    {
      "epoch": 9.874213836477988,
      "grad_norm": 0.6223800778388977,
      "learning_rate": 2.988286521199753e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 12327232,
      "step": 18840
    },
    {
      "epoch": 9.876834381551364,
      "grad_norm": 0.39313003420829773,
      "learning_rate": 2.9871650633742198e-05,
      "loss": 0.5521,
      "num_input_tokens_seen": 12330400,
      "step": 18845
    },
    {
      "epoch": 9.879454926624739,
      "grad_norm": 0.3512715697288513,
      "learning_rate": 2.9860435036393142e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 12333184,
      "step": 18850
    },
    {
      "epoch": 9.882075471698114,
      "grad_norm": 0.6053887605667114,
      "learning_rate": 2.9849218422296527e-05,
      "loss": 0.5203,
      "num_input_tokens_seen": 12335552,
      "step": 18855
    },
    {
      "epoch": 9.884696016771489,
      "grad_norm": 0.47505104541778564,
      "learning_rate": 2.983800079379873e-05,
      "loss": 0.4588,
      "num_input_tokens_seen": 12338848,
      "step": 18860
    },
    {
      "epoch": 9.887316561844864,
      "grad_norm": 0.49069181084632874,
      "learning_rate": 2.982678215324638e-05,
      "loss": 0.4158,
      "num_input_tokens_seen": 12341568,
      "step": 18865
    },
    {
      "epoch": 9.88993710691824,
      "grad_norm": 0.5087106227874756,
      "learning_rate": 2.9815562502986256e-05,
      "loss": 0.4386,
      "num_input_tokens_seen": 12344480,
      "step": 18870
    },
    {
      "epoch": 9.892557651991615,
      "grad_norm": 0.33999189734458923,
      "learning_rate": 2.9804341845365398e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 12347008,
      "step": 18875
    },
    {
      "epoch": 9.89517819706499,
      "grad_norm": 0.4005701541900635,
      "learning_rate": 2.9793120182731037e-05,
      "loss": 0.4146,
      "num_input_tokens_seen": 12350272,
      "step": 18880
    },
    {
      "epoch": 9.897798742138365,
      "grad_norm": 0.6015216112136841,
      "learning_rate": 2.9781897517430617e-05,
      "loss": 0.4187,
      "num_input_tokens_seen": 12352640,
      "step": 18885
    },
    {
      "epoch": 9.90041928721174,
      "grad_norm": 0.5223448872566223,
      "learning_rate": 2.977067385181179e-05,
      "loss": 0.5246,
      "num_input_tokens_seen": 12355488,
      "step": 18890
    },
    {
      "epoch": 9.903039832285115,
      "grad_norm": 0.4191029369831085,
      "learning_rate": 2.9759449188222426e-05,
      "loss": 0.3942,
      "num_input_tokens_seen": 12358336,
      "step": 18895
    },
    {
      "epoch": 9.90566037735849,
      "grad_norm": 0.40690121054649353,
      "learning_rate": 2.9748223529010582e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 12361952,
      "step": 18900
    },
    {
      "epoch": 9.908280922431866,
      "grad_norm": 0.8880859017372131,
      "learning_rate": 2.9736996876524546e-05,
      "loss": 0.4644,
      "num_input_tokens_seen": 12365472,
      "step": 18905
    },
    {
      "epoch": 9.91090146750524,
      "grad_norm": 0.6360986232757568,
      "learning_rate": 2.9725769233112816e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 12369792,
      "step": 18910
    },
    {
      "epoch": 9.913522012578616,
      "grad_norm": 0.5886117815971375,
      "learning_rate": 2.9714540601124074e-05,
      "loss": 0.3785,
      "num_input_tokens_seen": 12373248,
      "step": 18915
    },
    {
      "epoch": 9.916142557651991,
      "grad_norm": 0.37375274300575256,
      "learning_rate": 2.9703310982907218e-05,
      "loss": 0.3677,
      "num_input_tokens_seen": 12376224,
      "step": 18920
    },
    {
      "epoch": 9.918763102725366,
      "grad_norm": 0.6106703877449036,
      "learning_rate": 2.9692080380811365e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 12379424,
      "step": 18925
    },
    {
      "epoch": 9.921383647798741,
      "grad_norm": 0.4573727250099182,
      "learning_rate": 2.9680848797185828e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 12382784,
      "step": 18930
    },
    {
      "epoch": 9.924004192872118,
      "grad_norm": 0.3781870901584625,
      "learning_rate": 2.966961623438012e-05,
      "loss": 0.4014,
      "num_input_tokens_seen": 12386080,
      "step": 18935
    },
    {
      "epoch": 9.926624737945493,
      "grad_norm": 0.6186949610710144,
      "learning_rate": 2.9658382694743973e-05,
      "loss": 0.5398,
      "num_input_tokens_seen": 12389600,
      "step": 18940
    },
    {
      "epoch": 9.929245283018869,
      "grad_norm": 0.4579779803752899,
      "learning_rate": 2.9647148180627317e-05,
      "loss": 0.4022,
      "num_input_tokens_seen": 12392992,
      "step": 18945
    },
    {
      "epoch": 9.931865828092244,
      "grad_norm": 0.4602932929992676,
      "learning_rate": 2.9635912694380275e-05,
      "loss": 0.4551,
      "num_input_tokens_seen": 12396096,
      "step": 18950
    },
    {
      "epoch": 9.934486373165619,
      "grad_norm": 0.5320014953613281,
      "learning_rate": 2.962467623835319e-05,
      "loss": 0.4454,
      "num_input_tokens_seen": 12399488,
      "step": 18955
    },
    {
      "epoch": 9.937106918238994,
      "grad_norm": 0.5162785649299622,
      "learning_rate": 2.961343881489659e-05,
      "loss": 0.4097,
      "num_input_tokens_seen": 12403200,
      "step": 18960
    },
    {
      "epoch": 9.93972746331237,
      "grad_norm": 0.4726848602294922,
      "learning_rate": 2.9602200426361226e-05,
      "loss": 0.5349,
      "num_input_tokens_seen": 12406176,
      "step": 18965
    },
    {
      "epoch": 9.942348008385745,
      "grad_norm": 0.40432316064834595,
      "learning_rate": 2.9590961075098038e-05,
      "loss": 0.4498,
      "num_input_tokens_seen": 12408800,
      "step": 18970
    },
    {
      "epoch": 9.94496855345912,
      "grad_norm": 0.32874444127082825,
      "learning_rate": 2.957972076345817e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 12413216,
      "step": 18975
    },
    {
      "epoch": 9.947589098532495,
      "grad_norm": 0.5052968859672546,
      "learning_rate": 2.9568479493792967e-05,
      "loss": 0.3928,
      "num_input_tokens_seen": 12416352,
      "step": 18980
    },
    {
      "epoch": 9.95020964360587,
      "grad_norm": 0.42358872294425964,
      "learning_rate": 2.955723726845397e-05,
      "loss": 0.4794,
      "num_input_tokens_seen": 12419552,
      "step": 18985
    },
    {
      "epoch": 9.952830188679245,
      "grad_norm": 0.5177071690559387,
      "learning_rate": 2.9545994089792928e-05,
      "loss": 0.467,
      "num_input_tokens_seen": 12423008,
      "step": 18990
    },
    {
      "epoch": 9.95545073375262,
      "grad_norm": 0.34840846061706543,
      "learning_rate": 2.9534749960161795e-05,
      "loss": 0.5536,
      "num_input_tokens_seen": 12426208,
      "step": 18995
    },
    {
      "epoch": 9.958071278825996,
      "grad_norm": 0.43591228127479553,
      "learning_rate": 2.9523504881912694e-05,
      "loss": 0.3768,
      "num_input_tokens_seen": 12429664,
      "step": 19000
    },
    {
      "epoch": 9.96069182389937,
      "grad_norm": 0.28790614008903503,
      "learning_rate": 2.9512258857397978e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 12432512,
      "step": 19005
    },
    {
      "epoch": 9.963312368972746,
      "grad_norm": 0.6415286064147949,
      "learning_rate": 2.9501011888970203e-05,
      "loss": 0.4726,
      "num_input_tokens_seen": 12435392,
      "step": 19010
    },
    {
      "epoch": 9.965932914046121,
      "grad_norm": 0.7258475422859192,
      "learning_rate": 2.9489763978982075e-05,
      "loss": 0.5421,
      "num_input_tokens_seen": 12437568,
      "step": 19015
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 0.4790554642677307,
      "learning_rate": 2.9478515129786556e-05,
      "loss": 0.3792,
      "num_input_tokens_seen": 12440224,
      "step": 19020
    },
    {
      "epoch": 9.971174004192871,
      "grad_norm": 0.4803984761238098,
      "learning_rate": 2.946726534373676e-05,
      "loss": 0.5425,
      "num_input_tokens_seen": 12443136,
      "step": 19025
    },
    {
      "epoch": 9.973794549266248,
      "grad_norm": 0.9639845490455627,
      "learning_rate": 2.945601462318601e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 12445888,
      "step": 19030
    },
    {
      "epoch": 9.976415094339622,
      "grad_norm": 0.401348739862442,
      "learning_rate": 2.944476297048786e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 12448576,
      "step": 19035
    },
    {
      "epoch": 9.979035639412999,
      "grad_norm": 0.5369595885276794,
      "learning_rate": 2.9433510387995987e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 12451168,
      "step": 19040
    },
    {
      "epoch": 9.981656184486374,
      "grad_norm": 0.42926621437072754,
      "learning_rate": 2.9422256878064325e-05,
      "loss": 0.6318,
      "num_input_tokens_seen": 12456800,
      "step": 19045
    },
    {
      "epoch": 9.984276729559749,
      "grad_norm": 0.4932592511177063,
      "learning_rate": 2.9411002443046982e-05,
      "loss": 0.3875,
      "num_input_tokens_seen": 12460160,
      "step": 19050
    },
    {
      "epoch": 9.986897274633124,
      "grad_norm": 1.1340066194534302,
      "learning_rate": 2.9399747085298245e-05,
      "loss": 0.5171,
      "num_input_tokens_seen": 12462272,
      "step": 19055
    },
    {
      "epoch": 9.9895178197065,
      "grad_norm": 0.5397242307662964,
      "learning_rate": 2.9388490807172614e-05,
      "loss": 0.4365,
      "num_input_tokens_seen": 12464992,
      "step": 19060
    },
    {
      "epoch": 9.992138364779874,
      "grad_norm": 0.8005796074867249,
      "learning_rate": 2.9377233611024775e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 12467744,
      "step": 19065
    },
    {
      "epoch": 9.99475890985325,
      "grad_norm": 0.4993544816970825,
      "learning_rate": 2.9365975499209596e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 12471456,
      "step": 19070
    },
    {
      "epoch": 9.997379454926625,
      "grad_norm": 0.34723982214927673,
      "learning_rate": 2.9354716474082154e-05,
      "loss": 0.5129,
      "num_input_tokens_seen": 12474304,
      "step": 19075
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.04428231716156,
      "learning_rate": 2.9343456537997714e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 12476552,
      "step": 19080
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.4484306275844574,
      "eval_runtime": 9.723,
      "eval_samples_per_second": 87.216,
      "eval_steps_per_second": 21.804,
      "num_input_tokens_seen": 12476552,
      "step": 19080
    },
    {
      "epoch": 10.002620545073375,
      "grad_norm": 0.5557212829589844,
      "learning_rate": 2.9332195693311714e-05,
      "loss": 0.3829,
      "num_input_tokens_seen": 12479208,
      "step": 19085
    },
    {
      "epoch": 10.00524109014675,
      "grad_norm": 0.4647151827812195,
      "learning_rate": 2.9320933942379803e-05,
      "loss": 0.5538,
      "num_input_tokens_seen": 12483080,
      "step": 19090
    },
    {
      "epoch": 10.007861635220126,
      "grad_norm": 0.33662348985671997,
      "learning_rate": 2.9309671287557805e-05,
      "loss": 0.673,
      "num_input_tokens_seen": 12486696,
      "step": 19095
    },
    {
      "epoch": 10.0104821802935,
      "grad_norm": 0.3619711101055145,
      "learning_rate": 2.9298407731201737e-05,
      "loss": 0.4281,
      "num_input_tokens_seen": 12489224,
      "step": 19100
    },
    {
      "epoch": 10.013102725366876,
      "grad_norm": 0.39169079065322876,
      "learning_rate": 2.928714327566781e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 12492968,
      "step": 19105
    },
    {
      "epoch": 10.015723270440251,
      "grad_norm": 0.6210252046585083,
      "learning_rate": 2.9275877923312427e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 12495976,
      "step": 19110
    },
    {
      "epoch": 10.018343815513626,
      "grad_norm": 0.42404040694236755,
      "learning_rate": 2.9264611676492165e-05,
      "loss": 0.3385,
      "num_input_tokens_seen": 12498696,
      "step": 19115
    },
    {
      "epoch": 10.020964360587001,
      "grad_norm": 0.42879563570022583,
      "learning_rate": 2.9253344537563782e-05,
      "loss": 0.4516,
      "num_input_tokens_seen": 12501224,
      "step": 19120
    },
    {
      "epoch": 10.023584905660377,
      "grad_norm": 0.31688496470451355,
      "learning_rate": 2.924207650888426e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 12504264,
      "step": 19125
    },
    {
      "epoch": 10.026205450733753,
      "grad_norm": 0.4170700013637543,
      "learning_rate": 2.9230807592810723e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 12507240,
      "step": 19130
    },
    {
      "epoch": 10.028825995807129,
      "grad_norm": 0.293252170085907,
      "learning_rate": 2.92195377917005e-05,
      "loss": 0.3503,
      "num_input_tokens_seen": 12511304,
      "step": 19135
    },
    {
      "epoch": 10.031446540880504,
      "grad_norm": 0.3708172142505646,
      "learning_rate": 2.920826710791112e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 12513960,
      "step": 19140
    },
    {
      "epoch": 10.034067085953879,
      "grad_norm": 0.2832498848438263,
      "learning_rate": 2.919699554380026e-05,
      "loss": 0.3627,
      "num_input_tokens_seen": 12517704,
      "step": 19145
    },
    {
      "epoch": 10.036687631027254,
      "grad_norm": 0.39612722396850586,
      "learning_rate": 2.9185723101725825e-05,
      "loss": 0.335,
      "num_input_tokens_seen": 12521320,
      "step": 19150
    },
    {
      "epoch": 10.03930817610063,
      "grad_norm": 0.23091386258602142,
      "learning_rate": 2.9174449784045867e-05,
      "loss": 0.4451,
      "num_input_tokens_seen": 12525448,
      "step": 19155
    },
    {
      "epoch": 10.041928721174004,
      "grad_norm": 0.47693896293640137,
      "learning_rate": 2.916317559311863e-05,
      "loss": 0.5002,
      "num_input_tokens_seen": 12529000,
      "step": 19160
    },
    {
      "epoch": 10.04454926624738,
      "grad_norm": 0.38267573714256287,
      "learning_rate": 2.9151900531302566e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 12531784,
      "step": 19165
    },
    {
      "epoch": 10.047169811320755,
      "grad_norm": 0.348908931016922,
      "learning_rate": 2.9140624600956272e-05,
      "loss": 0.5101,
      "num_input_tokens_seen": 12535688,
      "step": 19170
    },
    {
      "epoch": 10.04979035639413,
      "grad_norm": 0.5007607340812683,
      "learning_rate": 2.912934780443855e-05,
      "loss": 0.4011,
      "num_input_tokens_seen": 12538472,
      "step": 19175
    },
    {
      "epoch": 10.052410901467505,
      "grad_norm": 0.7031987905502319,
      "learning_rate": 2.9118070144108388e-05,
      "loss": 0.6853,
      "num_input_tokens_seen": 12544936,
      "step": 19180
    },
    {
      "epoch": 10.05503144654088,
      "grad_norm": 0.44783324003219604,
      "learning_rate": 2.9106791622324924e-05,
      "loss": 0.3776,
      "num_input_tokens_seen": 12547816,
      "step": 19185
    },
    {
      "epoch": 10.057651991614255,
      "grad_norm": 0.5255595445632935,
      "learning_rate": 2.909551224144751e-05,
      "loss": 0.4801,
      "num_input_tokens_seen": 12551016,
      "step": 19190
    },
    {
      "epoch": 10.06027253668763,
      "grad_norm": 0.45350760221481323,
      "learning_rate": 2.9084232003835653e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 12554184,
      "step": 19195
    },
    {
      "epoch": 10.062893081761006,
      "grad_norm": 0.33992236852645874,
      "learning_rate": 2.9072950911849058e-05,
      "loss": 0.4648,
      "num_input_tokens_seen": 12557512,
      "step": 19200
    },
    {
      "epoch": 10.065513626834381,
      "grad_norm": 0.38850027322769165,
      "learning_rate": 2.90616689678476e-05,
      "loss": 0.5803,
      "num_input_tokens_seen": 12560904,
      "step": 19205
    },
    {
      "epoch": 10.068134171907756,
      "grad_norm": 0.5458477735519409,
      "learning_rate": 2.9050386174191337e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 12563368,
      "step": 19210
    },
    {
      "epoch": 10.070754716981131,
      "grad_norm": 0.3785324692726135,
      "learning_rate": 2.903910253324048e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 12566728,
      "step": 19215
    },
    {
      "epoch": 10.073375262054507,
      "grad_norm": 0.3621012568473816,
      "learning_rate": 2.9027818047355477e-05,
      "loss": 0.4788,
      "num_input_tokens_seen": 12570248,
      "step": 19220
    },
    {
      "epoch": 10.075995807127883,
      "grad_norm": 0.9416763186454773,
      "learning_rate": 2.9016532718896873e-05,
      "loss": 0.5277,
      "num_input_tokens_seen": 12572936,
      "step": 19225
    },
    {
      "epoch": 10.078616352201259,
      "grad_norm": 0.34713494777679443,
      "learning_rate": 2.900524655022545e-05,
      "loss": 0.5342,
      "num_input_tokens_seen": 12576328,
      "step": 19230
    },
    {
      "epoch": 10.081236897274634,
      "grad_norm": 0.6157730221748352,
      "learning_rate": 2.899395954370215e-05,
      "loss": 0.4037,
      "num_input_tokens_seen": 12579208,
      "step": 19235
    },
    {
      "epoch": 10.083857442348009,
      "grad_norm": 0.4798515737056732,
      "learning_rate": 2.898267170168807e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 12581896,
      "step": 19240
    },
    {
      "epoch": 10.086477987421384,
      "grad_norm": 0.37058645486831665,
      "learning_rate": 2.8971383026544507e-05,
      "loss": 0.6067,
      "num_input_tokens_seen": 12584840,
      "step": 19245
    },
    {
      "epoch": 10.08909853249476,
      "grad_norm": 0.39122724533081055,
      "learning_rate": 2.896009352063293e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 12588840,
      "step": 19250
    },
    {
      "epoch": 10.091719077568134,
      "grad_norm": 0.6706674098968506,
      "learning_rate": 2.894880318631496e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 12591752,
      "step": 19255
    },
    {
      "epoch": 10.09433962264151,
      "grad_norm": 0.9067119359970093,
      "learning_rate": 2.8937512025952417e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 12594504,
      "step": 19260
    },
    {
      "epoch": 10.096960167714885,
      "grad_norm": 0.2654620110988617,
      "learning_rate": 2.8926220041907283e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 12597544,
      "step": 19265
    },
    {
      "epoch": 10.09958071278826,
      "grad_norm": 0.33433210849761963,
      "learning_rate": 2.8914927236541706e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 12600488,
      "step": 19270
    },
    {
      "epoch": 10.102201257861635,
      "grad_norm": 0.41422152519226074,
      "learning_rate": 2.8903633612218012e-05,
      "loss": 0.4292,
      "num_input_tokens_seen": 12604040,
      "step": 19275
    },
    {
      "epoch": 10.10482180293501,
      "grad_norm": 0.5524862408638,
      "learning_rate": 2.8892339171298703e-05,
      "loss": 0.5548,
      "num_input_tokens_seen": 12606984,
      "step": 19280
    },
    {
      "epoch": 10.107442348008385,
      "grad_norm": 0.41939064860343933,
      "learning_rate": 2.888104391614645e-05,
      "loss": 0.4264,
      "num_input_tokens_seen": 12609800,
      "step": 19285
    },
    {
      "epoch": 10.11006289308176,
      "grad_norm": 0.36346864700317383,
      "learning_rate": 2.886974784912408e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 12614024,
      "step": 19290
    },
    {
      "epoch": 10.112683438155136,
      "grad_norm": 0.8815901279449463,
      "learning_rate": 2.8858450972594612e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 12617480,
      "step": 19295
    },
    {
      "epoch": 10.115303983228511,
      "grad_norm": 0.656872570514679,
      "learning_rate": 2.8847153288921226e-05,
      "loss": 0.4977,
      "num_input_tokens_seen": 12620904,
      "step": 19300
    },
    {
      "epoch": 10.117924528301886,
      "grad_norm": 0.4150760769844055,
      "learning_rate": 2.883585480046725e-05,
      "loss": 0.4794,
      "num_input_tokens_seen": 12623464,
      "step": 19305
    },
    {
      "epoch": 10.120545073375261,
      "grad_norm": 0.49869900941848755,
      "learning_rate": 2.8824555509596223e-05,
      "loss": 0.5698,
      "num_input_tokens_seen": 12626472,
      "step": 19310
    },
    {
      "epoch": 10.123165618448636,
      "grad_norm": 0.516054093837738,
      "learning_rate": 2.8813255418671815e-05,
      "loss": 0.2857,
      "num_input_tokens_seen": 12629224,
      "step": 19315
    },
    {
      "epoch": 10.125786163522013,
      "grad_norm": 0.3901466727256775,
      "learning_rate": 2.8801954530057874e-05,
      "loss": 0.4515,
      "num_input_tokens_seen": 12633032,
      "step": 19320
    },
    {
      "epoch": 10.128406708595389,
      "grad_norm": 0.4346875250339508,
      "learning_rate": 2.879065284611842e-05,
      "loss": 0.4391,
      "num_input_tokens_seen": 12636072,
      "step": 19325
    },
    {
      "epoch": 10.131027253668764,
      "grad_norm": 0.3660743832588196,
      "learning_rate": 2.8779350369217634e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 12639400,
      "step": 19330
    },
    {
      "epoch": 10.133647798742139,
      "grad_norm": 1.355201244354248,
      "learning_rate": 2.876804710171987e-05,
      "loss": 0.5304,
      "num_input_tokens_seen": 12642920,
      "step": 19335
    },
    {
      "epoch": 10.136268343815514,
      "grad_norm": 0.39573583006858826,
      "learning_rate": 2.8756743045989643e-05,
      "loss": 0.3983,
      "num_input_tokens_seen": 12646312,
      "step": 19340
    },
    {
      "epoch": 10.13888888888889,
      "grad_norm": 0.47253382205963135,
      "learning_rate": 2.8745438204391623e-05,
      "loss": 0.5604,
      "num_input_tokens_seen": 12648648,
      "step": 19345
    },
    {
      "epoch": 10.141509433962264,
      "grad_norm": 0.3017186224460602,
      "learning_rate": 2.8734132579290675e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 12651176,
      "step": 19350
    },
    {
      "epoch": 10.14412997903564,
      "grad_norm": 0.38295504450798035,
      "learning_rate": 2.8722826173051782e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 12653672,
      "step": 19355
    },
    {
      "epoch": 10.146750524109015,
      "grad_norm": 1.0722572803497314,
      "learning_rate": 2.871151898804012e-05,
      "loss": 0.4788,
      "num_input_tokens_seen": 12656744,
      "step": 19360
    },
    {
      "epoch": 10.14937106918239,
      "grad_norm": 0.48572948575019836,
      "learning_rate": 2.8700211026621043e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 12660168,
      "step": 19365
    },
    {
      "epoch": 10.151991614255765,
      "grad_norm": 0.36095112562179565,
      "learning_rate": 2.8688902291160025e-05,
      "loss": 0.5704,
      "num_input_tokens_seen": 12664424,
      "step": 19370
    },
    {
      "epoch": 10.15461215932914,
      "grad_norm": 0.5628170371055603,
      "learning_rate": 2.867759278402273e-05,
      "loss": 0.4093,
      "num_input_tokens_seen": 12667496,
      "step": 19375
    },
    {
      "epoch": 10.157232704402515,
      "grad_norm": 0.30580392479896545,
      "learning_rate": 2.8666282507574994e-05,
      "loss": 0.4361,
      "num_input_tokens_seen": 12671752,
      "step": 19380
    },
    {
      "epoch": 10.15985324947589,
      "grad_norm": 1.0506913661956787,
      "learning_rate": 2.8654971464182773e-05,
      "loss": 0.4777,
      "num_input_tokens_seen": 12674408,
      "step": 19385
    },
    {
      "epoch": 10.162473794549266,
      "grad_norm": 0.42464810609817505,
      "learning_rate": 2.8643659656212222e-05,
      "loss": 0.3962,
      "num_input_tokens_seen": 12676968,
      "step": 19390
    },
    {
      "epoch": 10.165094339622641,
      "grad_norm": 0.45758217573165894,
      "learning_rate": 2.8632347086029642e-05,
      "loss": 0.3182,
      "num_input_tokens_seen": 12680200,
      "step": 19395
    },
    {
      "epoch": 10.167714884696016,
      "grad_norm": 0.623037576675415,
      "learning_rate": 2.8621033756001486e-05,
      "loss": 0.6463,
      "num_input_tokens_seen": 12684104,
      "step": 19400
    },
    {
      "epoch": 10.170335429769391,
      "grad_norm": 0.6356914639472961,
      "learning_rate": 2.8609719668494394e-05,
      "loss": 0.3897,
      "num_input_tokens_seen": 12687688,
      "step": 19405
    },
    {
      "epoch": 10.172955974842766,
      "grad_norm": 0.28359395265579224,
      "learning_rate": 2.8598404825875113e-05,
      "loss": 0.3854,
      "num_input_tokens_seen": 12690728,
      "step": 19410
    },
    {
      "epoch": 10.175576519916143,
      "grad_norm": 0.5584954023361206,
      "learning_rate": 2.8587089230510604e-05,
      "loss": 0.3943,
      "num_input_tokens_seen": 12694568,
      "step": 19415
    },
    {
      "epoch": 10.178197064989519,
      "grad_norm": 0.7293024659156799,
      "learning_rate": 2.857577288476795e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 12697768,
      "step": 19420
    },
    {
      "epoch": 10.180817610062894,
      "grad_norm": 0.4265783429145813,
      "learning_rate": 2.85644557910144e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 12700840,
      "step": 19425
    },
    {
      "epoch": 10.183438155136269,
      "grad_norm": 0.3850974142551422,
      "learning_rate": 2.8553137951617366e-05,
      "loss": 0.449,
      "num_input_tokens_seen": 12703784,
      "step": 19430
    },
    {
      "epoch": 10.186058700209644,
      "grad_norm": 0.24160927534103394,
      "learning_rate": 2.854181936894441e-05,
      "loss": 0.417,
      "num_input_tokens_seen": 12707240,
      "step": 19435
    },
    {
      "epoch": 10.18867924528302,
      "grad_norm": 0.37023502588272095,
      "learning_rate": 2.8530500045363247e-05,
      "loss": 0.3587,
      "num_input_tokens_seen": 12710312,
      "step": 19440
    },
    {
      "epoch": 10.191299790356394,
      "grad_norm": 0.3665752112865448,
      "learning_rate": 2.851917998324175e-05,
      "loss": 0.5557,
      "num_input_tokens_seen": 12713640,
      "step": 19445
    },
    {
      "epoch": 10.19392033542977,
      "grad_norm": 1.0305697917938232,
      "learning_rate": 2.8507859184947957e-05,
      "loss": 0.3821,
      "num_input_tokens_seen": 12716616,
      "step": 19450
    },
    {
      "epoch": 10.196540880503145,
      "grad_norm": 0.48723456263542175,
      "learning_rate": 2.8496537652850043e-05,
      "loss": 0.4057,
      "num_input_tokens_seen": 12719528,
      "step": 19455
    },
    {
      "epoch": 10.19916142557652,
      "grad_norm": 0.5501089096069336,
      "learning_rate": 2.8485215389316333e-05,
      "loss": 0.3708,
      "num_input_tokens_seen": 12722952,
      "step": 19460
    },
    {
      "epoch": 10.201781970649895,
      "grad_norm": 0.27830976247787476,
      "learning_rate": 2.8473892396715328e-05,
      "loss": 0.4056,
      "num_input_tokens_seen": 12726440,
      "step": 19465
    },
    {
      "epoch": 10.20440251572327,
      "grad_norm": 0.5430903434753418,
      "learning_rate": 2.846256867741567e-05,
      "loss": 0.4951,
      "num_input_tokens_seen": 12729096,
      "step": 19470
    },
    {
      "epoch": 10.207023060796645,
      "grad_norm": 0.9888027310371399,
      "learning_rate": 2.845124423378614e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 12731976,
      "step": 19475
    },
    {
      "epoch": 10.20964360587002,
      "grad_norm": 0.6504284143447876,
      "learning_rate": 2.8439919068195697e-05,
      "loss": 0.4618,
      "num_input_tokens_seen": 12734696,
      "step": 19480
    },
    {
      "epoch": 10.212264150943396,
      "grad_norm": 0.5058473348617554,
      "learning_rate": 2.842859318301343e-05,
      "loss": 0.4707,
      "num_input_tokens_seen": 12738184,
      "step": 19485
    },
    {
      "epoch": 10.214884696016771,
      "grad_norm": 0.37484413385391235,
      "learning_rate": 2.841726658060858e-05,
      "loss": 0.354,
      "num_input_tokens_seen": 12741224,
      "step": 19490
    },
    {
      "epoch": 10.217505241090146,
      "grad_norm": 0.33782434463500977,
      "learning_rate": 2.8405939263350555e-05,
      "loss": 0.3963,
      "num_input_tokens_seen": 12743976,
      "step": 19495
    },
    {
      "epoch": 10.220125786163521,
      "grad_norm": 0.5782505869865417,
      "learning_rate": 2.8394611233608886e-05,
      "loss": 0.5328,
      "num_input_tokens_seen": 12747464,
      "step": 19500
    },
    {
      "epoch": 10.222746331236896,
      "grad_norm": 0.6028364896774292,
      "learning_rate": 2.8383282493753283e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 12750440,
      "step": 19505
    },
    {
      "epoch": 10.225366876310272,
      "grad_norm": 0.44381260871887207,
      "learning_rate": 2.8371953046153575e-05,
      "loss": 0.5151,
      "num_input_tokens_seen": 12753448,
      "step": 19510
    },
    {
      "epoch": 10.227987421383649,
      "grad_norm": 0.44346046447753906,
      "learning_rate": 2.8360622893179757e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 12757352,
      "step": 19515
    },
    {
      "epoch": 10.230607966457024,
      "grad_norm": 0.2074626386165619,
      "learning_rate": 2.834929203720197e-05,
      "loss": 0.3295,
      "num_input_tokens_seen": 12760968,
      "step": 19520
    },
    {
      "epoch": 10.233228511530399,
      "grad_norm": 0.6232503056526184,
      "learning_rate": 2.83379604805905e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 12764008,
      "step": 19525
    },
    {
      "epoch": 10.235849056603774,
      "grad_norm": 0.5065596699714661,
      "learning_rate": 2.832662822571578e-05,
      "loss": 0.608,
      "num_input_tokens_seen": 12766600,
      "step": 19530
    },
    {
      "epoch": 10.23846960167715,
      "grad_norm": 0.615449845790863,
      "learning_rate": 2.8315295274948395e-05,
      "loss": 0.3781,
      "num_input_tokens_seen": 12769608,
      "step": 19535
    },
    {
      "epoch": 10.241090146750524,
      "grad_norm": 0.36943626403808594,
      "learning_rate": 2.8303961630659047e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 12773608,
      "step": 19540
    },
    {
      "epoch": 10.2437106918239,
      "grad_norm": 0.3656565248966217,
      "learning_rate": 2.8292627295218627e-05,
      "loss": 0.3969,
      "num_input_tokens_seen": 12776296,
      "step": 19545
    },
    {
      "epoch": 10.246331236897275,
      "grad_norm": 0.28271082043647766,
      "learning_rate": 2.828129227099815e-05,
      "loss": 0.4007,
      "num_input_tokens_seen": 12778984,
      "step": 19550
    },
    {
      "epoch": 10.24895178197065,
      "grad_norm": 0.6997239589691162,
      "learning_rate": 2.826995656036875e-05,
      "loss": 0.4966,
      "num_input_tokens_seen": 12782248,
      "step": 19555
    },
    {
      "epoch": 10.251572327044025,
      "grad_norm": 0.3525499701499939,
      "learning_rate": 2.8258620165701747e-05,
      "loss": 0.3788,
      "num_input_tokens_seen": 12785032,
      "step": 19560
    },
    {
      "epoch": 10.2541928721174,
      "grad_norm": 0.3698577284812927,
      "learning_rate": 2.8247283089368586e-05,
      "loss": 0.4609,
      "num_input_tokens_seen": 12787848,
      "step": 19565
    },
    {
      "epoch": 10.256813417190775,
      "grad_norm": 0.6357420682907104,
      "learning_rate": 2.8235945333740842e-05,
      "loss": 0.3637,
      "num_input_tokens_seen": 12790440,
      "step": 19570
    },
    {
      "epoch": 10.25943396226415,
      "grad_norm": 0.8708006143569946,
      "learning_rate": 2.822460690119026e-05,
      "loss": 0.6336,
      "num_input_tokens_seen": 12793032,
      "step": 19575
    },
    {
      "epoch": 10.262054507337526,
      "grad_norm": 0.3166930675506592,
      "learning_rate": 2.82132677940887e-05,
      "loss": 0.589,
      "num_input_tokens_seen": 12796392,
      "step": 19580
    },
    {
      "epoch": 10.264675052410901,
      "grad_norm": 0.5866371989250183,
      "learning_rate": 2.820192801480817e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 12799240,
      "step": 19585
    },
    {
      "epoch": 10.267295597484276,
      "grad_norm": 0.4665009379386902,
      "learning_rate": 2.8190587565720842e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 12803752,
      "step": 19590
    },
    {
      "epoch": 10.269916142557651,
      "grad_norm": 0.41074228286743164,
      "learning_rate": 2.8179246449198986e-05,
      "loss": 0.3618,
      "num_input_tokens_seen": 12806568,
      "step": 19595
    },
    {
      "epoch": 10.272536687631026,
      "grad_norm": 0.7788923382759094,
      "learning_rate": 2.8167904667615048e-05,
      "loss": 0.3609,
      "num_input_tokens_seen": 12811464,
      "step": 19600
    },
    {
      "epoch": 10.275157232704402,
      "grad_norm": 0.3071759045124054,
      "learning_rate": 2.8156562223341592e-05,
      "loss": 0.5175,
      "num_input_tokens_seen": 12815880,
      "step": 19605
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 0.6871597170829773,
      "learning_rate": 2.814521911875133e-05,
      "loss": 0.474,
      "num_input_tokens_seen": 12818632,
      "step": 19610
    },
    {
      "epoch": 10.280398322851154,
      "grad_norm": 0.2848377525806427,
      "learning_rate": 2.8133875356217114e-05,
      "loss": 0.3649,
      "num_input_tokens_seen": 12821736,
      "step": 19615
    },
    {
      "epoch": 10.283018867924529,
      "grad_norm": 0.3607431650161743,
      "learning_rate": 2.8122530938111923e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 12824584,
      "step": 19620
    },
    {
      "epoch": 10.285639412997904,
      "grad_norm": 0.5478637218475342,
      "learning_rate": 2.8111185866808888e-05,
      "loss": 0.5313,
      "num_input_tokens_seen": 12827624,
      "step": 19625
    },
    {
      "epoch": 10.28825995807128,
      "grad_norm": 0.5405433773994446,
      "learning_rate": 2.809984014468126e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 12830440,
      "step": 19630
    },
    {
      "epoch": 10.290880503144654,
      "grad_norm": 0.3391660749912262,
      "learning_rate": 2.8088493774102443e-05,
      "loss": 0.369,
      "num_input_tokens_seen": 12836456,
      "step": 19635
    },
    {
      "epoch": 10.29350104821803,
      "grad_norm": 0.5787503719329834,
      "learning_rate": 2.807714675744596e-05,
      "loss": 0.5847,
      "num_input_tokens_seen": 12839432,
      "step": 19640
    },
    {
      "epoch": 10.296121593291405,
      "grad_norm": 0.6528051495552063,
      "learning_rate": 2.8065799097085475e-05,
      "loss": 0.3682,
      "num_input_tokens_seen": 12843624,
      "step": 19645
    },
    {
      "epoch": 10.29874213836478,
      "grad_norm": 0.7248002290725708,
      "learning_rate": 2.8054450795394804e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 12845960,
      "step": 19650
    },
    {
      "epoch": 10.301362683438155,
      "grad_norm": 0.4045386016368866,
      "learning_rate": 2.8043101854747878e-05,
      "loss": 0.4257,
      "num_input_tokens_seen": 12850216,
      "step": 19655
    },
    {
      "epoch": 10.30398322851153,
      "grad_norm": 0.46879515051841736,
      "learning_rate": 2.8031752277518748e-05,
      "loss": 0.4188,
      "num_input_tokens_seen": 12853064,
      "step": 19660
    },
    {
      "epoch": 10.306603773584905,
      "grad_norm": 0.4272969663143158,
      "learning_rate": 2.802040206608164e-05,
      "loss": 0.411,
      "num_input_tokens_seen": 12855912,
      "step": 19665
    },
    {
      "epoch": 10.30922431865828,
      "grad_norm": 0.526772141456604,
      "learning_rate": 2.8009051222810877e-05,
      "loss": 0.4439,
      "num_input_tokens_seen": 12858824,
      "step": 19670
    },
    {
      "epoch": 10.311844863731656,
      "grad_norm": 0.6519392728805542,
      "learning_rate": 2.7997699750080918e-05,
      "loss": 0.5801,
      "num_input_tokens_seen": 12861416,
      "step": 19675
    },
    {
      "epoch": 10.314465408805031,
      "grad_norm": 0.5710827708244324,
      "learning_rate": 2.7986347650266386e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 12864968,
      "step": 19680
    },
    {
      "epoch": 10.317085953878406,
      "grad_norm": 0.717348039150238,
      "learning_rate": 2.797499492574199e-05,
      "loss": 0.43,
      "num_input_tokens_seen": 12868040,
      "step": 19685
    },
    {
      "epoch": 10.319706498951781,
      "grad_norm": 0.33248138427734375,
      "learning_rate": 2.7963641578882594e-05,
      "loss": 0.4822,
      "num_input_tokens_seen": 12871336,
      "step": 19690
    },
    {
      "epoch": 10.322327044025156,
      "grad_norm": 0.5226349830627441,
      "learning_rate": 2.7952287612063193e-05,
      "loss": 0.4032,
      "num_input_tokens_seen": 12875336,
      "step": 19695
    },
    {
      "epoch": 10.324947589098532,
      "grad_norm": 0.5708103179931641,
      "learning_rate": 2.79409330276589e-05,
      "loss": 0.4671,
      "num_input_tokens_seen": 12878696,
      "step": 19700
    },
    {
      "epoch": 10.327568134171909,
      "grad_norm": 0.20046088099479675,
      "learning_rate": 2.7929577828044974e-05,
      "loss": 0.4816,
      "num_input_tokens_seen": 12882600,
      "step": 19705
    },
    {
      "epoch": 10.330188679245284,
      "grad_norm": 0.5419148802757263,
      "learning_rate": 2.7918222015596784e-05,
      "loss": 0.3916,
      "num_input_tokens_seen": 12885192,
      "step": 19710
    },
    {
      "epoch": 10.332809224318659,
      "grad_norm": 0.7802144885063171,
      "learning_rate": 2.7906865592689847e-05,
      "loss": 0.3673,
      "num_input_tokens_seen": 12888040,
      "step": 19715
    },
    {
      "epoch": 10.335429769392034,
      "grad_norm": 0.5549254417419434,
      "learning_rate": 2.789550856169979e-05,
      "loss": 0.5147,
      "num_input_tokens_seen": 12890536,
      "step": 19720
    },
    {
      "epoch": 10.33805031446541,
      "grad_norm": 0.8094576597213745,
      "learning_rate": 2.7884150925002374e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 12892872,
      "step": 19725
    },
    {
      "epoch": 10.340670859538784,
      "grad_norm": 0.3030000329017639,
      "learning_rate": 2.787279268497349e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 12896008,
      "step": 19730
    },
    {
      "epoch": 10.34329140461216,
      "grad_norm": 0.6259400248527527,
      "learning_rate": 2.786143384398915e-05,
      "loss": 0.354,
      "num_input_tokens_seen": 12899464,
      "step": 19735
    },
    {
      "epoch": 10.345911949685535,
      "grad_norm": 0.533881664276123,
      "learning_rate": 2.785007440442549e-05,
      "loss": 0.4552,
      "num_input_tokens_seen": 12902312,
      "step": 19740
    },
    {
      "epoch": 10.34853249475891,
      "grad_norm": 0.5992128849029541,
      "learning_rate": 2.7838714368658785e-05,
      "loss": 0.3527,
      "num_input_tokens_seen": 12906760,
      "step": 19745
    },
    {
      "epoch": 10.351153039832285,
      "grad_norm": 0.6449118852615356,
      "learning_rate": 2.7827353739065427e-05,
      "loss": 0.5123,
      "num_input_tokens_seen": 12909576,
      "step": 19750
    },
    {
      "epoch": 10.35377358490566,
      "grad_norm": 0.3231385052204132,
      "learning_rate": 2.781599251802191e-05,
      "loss": 0.3393,
      "num_input_tokens_seen": 12912808,
      "step": 19755
    },
    {
      "epoch": 10.356394129979035,
      "grad_norm": 0.48040857911109924,
      "learning_rate": 2.780463070790489e-05,
      "loss": 0.3113,
      "num_input_tokens_seen": 12916552,
      "step": 19760
    },
    {
      "epoch": 10.35901467505241,
      "grad_norm": 0.2094353288412094,
      "learning_rate": 2.7793268311091125e-05,
      "loss": 0.4528,
      "num_input_tokens_seen": 12919368,
      "step": 19765
    },
    {
      "epoch": 10.361635220125786,
      "grad_norm": 0.4323672652244568,
      "learning_rate": 2.7781905329957496e-05,
      "loss": 0.3071,
      "num_input_tokens_seen": 12923976,
      "step": 19770
    },
    {
      "epoch": 10.364255765199161,
      "grad_norm": 0.4562002718448639,
      "learning_rate": 2.7770541766881008e-05,
      "loss": 0.4763,
      "num_input_tokens_seen": 12926728,
      "step": 19775
    },
    {
      "epoch": 10.366876310272536,
      "grad_norm": 0.6184608340263367,
      "learning_rate": 2.7759177624238786e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 12931176,
      "step": 19780
    },
    {
      "epoch": 10.369496855345911,
      "grad_norm": 0.640728235244751,
      "learning_rate": 2.7747812904408094e-05,
      "loss": 0.5407,
      "num_input_tokens_seen": 12934088,
      "step": 19785
    },
    {
      "epoch": 10.372117400419286,
      "grad_norm": 0.5449515581130981,
      "learning_rate": 2.7736447609766286e-05,
      "loss": 0.3673,
      "num_input_tokens_seen": 12936936,
      "step": 19790
    },
    {
      "epoch": 10.374737945492662,
      "grad_norm": 0.26722216606140137,
      "learning_rate": 2.772508174269085e-05,
      "loss": 0.425,
      "num_input_tokens_seen": 12940968,
      "step": 19795
    },
    {
      "epoch": 10.377358490566039,
      "grad_norm": 0.6963452100753784,
      "learning_rate": 2.7713715305559414e-05,
      "loss": 0.5574,
      "num_input_tokens_seen": 12944840,
      "step": 19800
    },
    {
      "epoch": 10.379979035639414,
      "grad_norm": 0.3919859230518341,
      "learning_rate": 2.7702348300749692e-05,
      "loss": 0.4033,
      "num_input_tokens_seen": 12947848,
      "step": 19805
    },
    {
      "epoch": 10.382599580712789,
      "grad_norm": 0.30638009309768677,
      "learning_rate": 2.7690980730639533e-05,
      "loss": 0.4415,
      "num_input_tokens_seen": 12951496,
      "step": 19810
    },
    {
      "epoch": 10.385220125786164,
      "grad_norm": 0.3934016525745392,
      "learning_rate": 2.7679612597606902e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 12955272,
      "step": 19815
    },
    {
      "epoch": 10.38784067085954,
      "grad_norm": 0.395949125289917,
      "learning_rate": 2.7668243904029894e-05,
      "loss": 0.3774,
      "num_input_tokens_seen": 12958376,
      "step": 19820
    },
    {
      "epoch": 10.390461215932914,
      "grad_norm": 0.7679216265678406,
      "learning_rate": 2.76568746522867e-05,
      "loss": 0.4297,
      "num_input_tokens_seen": 12961192,
      "step": 19825
    },
    {
      "epoch": 10.39308176100629,
      "grad_norm": 0.33430200815200806,
      "learning_rate": 2.7645504844755637e-05,
      "loss": 0.4342,
      "num_input_tokens_seen": 12965256,
      "step": 19830
    },
    {
      "epoch": 10.395702306079665,
      "grad_norm": 0.6010422706604004,
      "learning_rate": 2.7634134483815143e-05,
      "loss": 0.6519,
      "num_input_tokens_seen": 12968456,
      "step": 19835
    },
    {
      "epoch": 10.39832285115304,
      "grad_norm": 0.485422819852829,
      "learning_rate": 2.762276357184377e-05,
      "loss": 0.5106,
      "num_input_tokens_seen": 12972104,
      "step": 19840
    },
    {
      "epoch": 10.400943396226415,
      "grad_norm": 0.7512326836585999,
      "learning_rate": 2.7611392111220176e-05,
      "loss": 0.5074,
      "num_input_tokens_seen": 12975752,
      "step": 19845
    },
    {
      "epoch": 10.40356394129979,
      "grad_norm": 0.326204389333725,
      "learning_rate": 2.7600020104323154e-05,
      "loss": 0.4463,
      "num_input_tokens_seen": 12979656,
      "step": 19850
    },
    {
      "epoch": 10.406184486373165,
      "grad_norm": 0.46898579597473145,
      "learning_rate": 2.7588647553531576e-05,
      "loss": 0.5223,
      "num_input_tokens_seen": 12982696,
      "step": 19855
    },
    {
      "epoch": 10.40880503144654,
      "grad_norm": 0.3642999529838562,
      "learning_rate": 2.7577274461224463e-05,
      "loss": 0.3892,
      "num_input_tokens_seen": 12985576,
      "step": 19860
    },
    {
      "epoch": 10.411425576519916,
      "grad_norm": 0.3742739260196686,
      "learning_rate": 2.756590082978095e-05,
      "loss": 0.3782,
      "num_input_tokens_seen": 12988904,
      "step": 19865
    },
    {
      "epoch": 10.414046121593291,
      "grad_norm": 0.5839921832084656,
      "learning_rate": 2.7554526661580243e-05,
      "loss": 0.3418,
      "num_input_tokens_seen": 12992520,
      "step": 19870
    },
    {
      "epoch": 10.416666666666666,
      "grad_norm": 0.47693368792533875,
      "learning_rate": 2.7543151959001708e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 12996040,
      "step": 19875
    },
    {
      "epoch": 10.419287211740041,
      "grad_norm": 0.5913225412368774,
      "learning_rate": 2.7531776724424797e-05,
      "loss": 0.6023,
      "num_input_tokens_seen": 12999848,
      "step": 19880
    },
    {
      "epoch": 10.421907756813416,
      "grad_norm": 0.2918276786804199,
      "learning_rate": 2.752040096022907e-05,
      "loss": 0.4664,
      "num_input_tokens_seen": 13003144,
      "step": 19885
    },
    {
      "epoch": 10.424528301886792,
      "grad_norm": 0.35061901807785034,
      "learning_rate": 2.750902466879423e-05,
      "loss": 0.3491,
      "num_input_tokens_seen": 13005512,
      "step": 19890
    },
    {
      "epoch": 10.427148846960169,
      "grad_norm": 0.6673669815063477,
      "learning_rate": 2.7497647852500048e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 13008680,
      "step": 19895
    },
    {
      "epoch": 10.429769392033544,
      "grad_norm": 0.5360912680625916,
      "learning_rate": 2.7486270513726424e-05,
      "loss": 0.5763,
      "num_input_tokens_seen": 13011688,
      "step": 19900
    },
    {
      "epoch": 10.432389937106919,
      "grad_norm": 0.23444882035255432,
      "learning_rate": 2.7474892654853384e-05,
      "loss": 0.3968,
      "num_input_tokens_seen": 13014664,
      "step": 19905
    },
    {
      "epoch": 10.435010482180294,
      "grad_norm": 0.4030393362045288,
      "learning_rate": 2.7463514278261034e-05,
      "loss": 0.4636,
      "num_input_tokens_seen": 13019112,
      "step": 19910
    },
    {
      "epoch": 10.43763102725367,
      "grad_norm": 1.075124979019165,
      "learning_rate": 2.7452135386329596e-05,
      "loss": 0.627,
      "num_input_tokens_seen": 13021832,
      "step": 19915
    },
    {
      "epoch": 10.440251572327044,
      "grad_norm": 0.395799458026886,
      "learning_rate": 2.7440755981439414e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 13025128,
      "step": 19920
    },
    {
      "epoch": 10.44287211740042,
      "grad_norm": 0.3518705666065216,
      "learning_rate": 2.7429376065970924e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 13028072,
      "step": 19925
    },
    {
      "epoch": 10.445492662473795,
      "grad_norm": 0.6542487740516663,
      "learning_rate": 2.7417995642304678e-05,
      "loss": 0.5017,
      "num_input_tokens_seen": 13031272,
      "step": 19930
    },
    {
      "epoch": 10.44811320754717,
      "grad_norm": 0.37535560131073,
      "learning_rate": 2.7406614712821333e-05,
      "loss": 0.3611,
      "num_input_tokens_seen": 13034120,
      "step": 19935
    },
    {
      "epoch": 10.450733752620545,
      "grad_norm": 0.3758050799369812,
      "learning_rate": 2.739523327990164e-05,
      "loss": 0.519,
      "num_input_tokens_seen": 13037256,
      "step": 19940
    },
    {
      "epoch": 10.45335429769392,
      "grad_norm": 0.6073431968688965,
      "learning_rate": 2.7383851345926477e-05,
      "loss": 0.3952,
      "num_input_tokens_seen": 13040008,
      "step": 19945
    },
    {
      "epoch": 10.455974842767295,
      "grad_norm": 0.33775007724761963,
      "learning_rate": 2.7372468913276804e-05,
      "loss": 0.4406,
      "num_input_tokens_seen": 13043944,
      "step": 19950
    },
    {
      "epoch": 10.45859538784067,
      "grad_norm": 0.35013729333877563,
      "learning_rate": 2.7361085984333705e-05,
      "loss": 0.4955,
      "num_input_tokens_seen": 13046408,
      "step": 19955
    },
    {
      "epoch": 10.461215932914046,
      "grad_norm": 0.4466109275817871,
      "learning_rate": 2.7349702561478356e-05,
      "loss": 0.4073,
      "num_input_tokens_seen": 13049832,
      "step": 19960
    },
    {
      "epoch": 10.463836477987421,
      "grad_norm": 0.3135738968849182,
      "learning_rate": 2.733831864709203e-05,
      "loss": 0.4499,
      "num_input_tokens_seen": 13053640,
      "step": 19965
    },
    {
      "epoch": 10.466457023060796,
      "grad_norm": 0.4875093400478363,
      "learning_rate": 2.732693424355613e-05,
      "loss": 0.4297,
      "num_input_tokens_seen": 13056680,
      "step": 19970
    },
    {
      "epoch": 10.469077568134171,
      "grad_norm": 0.552892804145813,
      "learning_rate": 2.731554935325213e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 13059560,
      "step": 19975
    },
    {
      "epoch": 10.471698113207546,
      "grad_norm": 0.8099747896194458,
      "learning_rate": 2.7304163978561613e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 13062856,
      "step": 19980
    },
    {
      "epoch": 10.474318658280922,
      "grad_norm": 0.4779302775859833,
      "learning_rate": 2.729277812186629e-05,
      "loss": 0.3464,
      "num_input_tokens_seen": 13068392,
      "step": 19985
    },
    {
      "epoch": 10.476939203354299,
      "grad_norm": 0.3172908425331116,
      "learning_rate": 2.7281391785547945e-05,
      "loss": 0.5342,
      "num_input_tokens_seen": 13071400,
      "step": 19990
    },
    {
      "epoch": 10.479559748427674,
      "grad_norm": 0.9268481135368347,
      "learning_rate": 2.7270004971988456e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 13074760,
      "step": 19995
    },
    {
      "epoch": 10.482180293501049,
      "grad_norm": 1.6284830570220947,
      "learning_rate": 2.7258617683569825e-05,
      "loss": 0.4838,
      "num_input_tokens_seen": 13078664,
      "step": 20000
    },
    {
      "epoch": 10.484800838574424,
      "grad_norm": 0.9463963508605957,
      "learning_rate": 2.7247229922674143e-05,
      "loss": 0.604,
      "num_input_tokens_seen": 13081448,
      "step": 20005
    },
    {
      "epoch": 10.4874213836478,
      "grad_norm": 0.4415876269340515,
      "learning_rate": 2.7235841691683605e-05,
      "loss": 0.3604,
      "num_input_tokens_seen": 13085224,
      "step": 20010
    },
    {
      "epoch": 10.490041928721174,
      "grad_norm": 0.5899642109870911,
      "learning_rate": 2.7224452992980487e-05,
      "loss": 0.585,
      "num_input_tokens_seen": 13088136,
      "step": 20015
    },
    {
      "epoch": 10.49266247379455,
      "grad_norm": 0.6365948915481567,
      "learning_rate": 2.7213063828947182e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 13091112,
      "step": 20020
    },
    {
      "epoch": 10.495283018867925,
      "grad_norm": 0.5628882050514221,
      "learning_rate": 2.7201674201966176e-05,
      "loss": 0.4046,
      "num_input_tokens_seen": 13094376,
      "step": 20025
    },
    {
      "epoch": 10.4979035639413,
      "grad_norm": 0.35888099670410156,
      "learning_rate": 2.7190284114420035e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 13097160,
      "step": 20030
    },
    {
      "epoch": 10.500524109014675,
      "grad_norm": 0.7459242343902588,
      "learning_rate": 2.717889356869146e-05,
      "loss": 0.3842,
      "num_input_tokens_seen": 13101128,
      "step": 20035
    },
    {
      "epoch": 10.50314465408805,
      "grad_norm": 0.5663238167762756,
      "learning_rate": 2.7167502567163194e-05,
      "loss": 0.5395,
      "num_input_tokens_seen": 13104552,
      "step": 20040
    },
    {
      "epoch": 10.505765199161425,
      "grad_norm": 0.47794851660728455,
      "learning_rate": 2.715611111221813e-05,
      "loss": 0.4606,
      "num_input_tokens_seen": 13106920,
      "step": 20045
    },
    {
      "epoch": 10.5083857442348,
      "grad_norm": 0.46338966488838196,
      "learning_rate": 2.7144719206239217e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 13110408,
      "step": 20050
    },
    {
      "epoch": 10.511006289308176,
      "grad_norm": 0.4103107750415802,
      "learning_rate": 2.7133326851609508e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 13112808,
      "step": 20055
    },
    {
      "epoch": 10.51362683438155,
      "grad_norm": 0.31996122002601624,
      "learning_rate": 2.7121934050712166e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 13116680,
      "step": 20060
    },
    {
      "epoch": 10.516247379454926,
      "grad_norm": 0.6278875470161438,
      "learning_rate": 2.711054080593043e-05,
      "loss": 0.5004,
      "num_input_tokens_seen": 13119272,
      "step": 20065
    },
    {
      "epoch": 10.518867924528301,
      "grad_norm": 0.41237902641296387,
      "learning_rate": 2.7099147119647627e-05,
      "loss": 0.4349,
      "num_input_tokens_seen": 13122664,
      "step": 20070
    },
    {
      "epoch": 10.521488469601676,
      "grad_norm": 0.8196198344230652,
      "learning_rate": 2.7087752994247206e-05,
      "loss": 0.4384,
      "num_input_tokens_seen": 13125672,
      "step": 20075
    },
    {
      "epoch": 10.524109014675052,
      "grad_norm": 0.5732174515724182,
      "learning_rate": 2.707635843211267e-05,
      "loss": 0.4948,
      "num_input_tokens_seen": 13128392,
      "step": 20080
    },
    {
      "epoch": 10.526729559748428,
      "grad_norm": 0.588798999786377,
      "learning_rate": 2.7064963435627637e-05,
      "loss": 0.3711,
      "num_input_tokens_seen": 13131624,
      "step": 20085
    },
    {
      "epoch": 10.529350104821804,
      "grad_norm": 0.3165706694126129,
      "learning_rate": 2.7053568007175823e-05,
      "loss": 0.475,
      "num_input_tokens_seen": 13135240,
      "step": 20090
    },
    {
      "epoch": 10.531970649895179,
      "grad_norm": 0.3350890576839447,
      "learning_rate": 2.7042172149140994e-05,
      "loss": 0.5954,
      "num_input_tokens_seen": 13138152,
      "step": 20095
    },
    {
      "epoch": 10.534591194968554,
      "grad_norm": 0.6946736574172974,
      "learning_rate": 2.703077586390706e-05,
      "loss": 0.3955,
      "num_input_tokens_seen": 13144040,
      "step": 20100
    },
    {
      "epoch": 10.53721174004193,
      "grad_norm": 0.48255541920661926,
      "learning_rate": 2.7019379153857982e-05,
      "loss": 0.5488,
      "num_input_tokens_seen": 13146664,
      "step": 20105
    },
    {
      "epoch": 10.539832285115304,
      "grad_norm": 0.5961949825286865,
      "learning_rate": 2.700798202137782e-05,
      "loss": 0.5258,
      "num_input_tokens_seen": 13151080,
      "step": 20110
    },
    {
      "epoch": 10.54245283018868,
      "grad_norm": 0.7215142846107483,
      "learning_rate": 2.6996584468850732e-05,
      "loss": 0.4436,
      "num_input_tokens_seen": 13154600,
      "step": 20115
    },
    {
      "epoch": 10.545073375262055,
      "grad_norm": 0.8529426455497742,
      "learning_rate": 2.6985186498660946e-05,
      "loss": 0.3086,
      "num_input_tokens_seen": 13157928,
      "step": 20120
    },
    {
      "epoch": 10.54769392033543,
      "grad_norm": 0.3170647919178009,
      "learning_rate": 2.69737881131928e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 13161000,
      "step": 20125
    },
    {
      "epoch": 10.550314465408805,
      "grad_norm": 0.622800886631012,
      "learning_rate": 2.6962389314830692e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 13163880,
      "step": 20130
    },
    {
      "epoch": 10.55293501048218,
      "grad_norm": 0.4195508360862732,
      "learning_rate": 2.6950990105959135e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 13166376,
      "step": 20135
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 0.4055559039115906,
      "learning_rate": 2.6939590488962703e-05,
      "loss": 0.3508,
      "num_input_tokens_seen": 13170120,
      "step": 20140
    },
    {
      "epoch": 10.55817610062893,
      "grad_norm": 0.6114981770515442,
      "learning_rate": 2.6928190466226067e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 13172552,
      "step": 20145
    },
    {
      "epoch": 10.560796645702306,
      "grad_norm": 0.4084034562110901,
      "learning_rate": 2.691679004013398e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 13176456,
      "step": 20150
    },
    {
      "epoch": 10.56341719077568,
      "grad_norm": 0.38020357489585876,
      "learning_rate": 2.6905389213071285e-05,
      "loss": 0.4737,
      "num_input_tokens_seen": 13179496,
      "step": 20155
    },
    {
      "epoch": 10.566037735849056,
      "grad_norm": 0.3942023813724518,
      "learning_rate": 2.6893987987422907e-05,
      "loss": 0.5208,
      "num_input_tokens_seen": 13182504,
      "step": 20160
    },
    {
      "epoch": 10.568658280922431,
      "grad_norm": 0.1547868549823761,
      "learning_rate": 2.6882586365573854e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 13185960,
      "step": 20165
    },
    {
      "epoch": 10.571278825995806,
      "grad_norm": 0.42258402705192566,
      "learning_rate": 2.6871184349909206e-05,
      "loss": 0.4239,
      "num_input_tokens_seen": 13188968,
      "step": 20170
    },
    {
      "epoch": 10.573899371069182,
      "grad_norm": 0.3775765299797058,
      "learning_rate": 2.685978194281414e-05,
      "loss": 0.3096,
      "num_input_tokens_seen": 13191944,
      "step": 20175
    },
    {
      "epoch": 10.576519916142558,
      "grad_norm": 0.6971015334129333,
      "learning_rate": 2.6848379146673914e-05,
      "loss": 0.6118,
      "num_input_tokens_seen": 13195496,
      "step": 20180
    },
    {
      "epoch": 10.579140461215934,
      "grad_norm": 0.5473385453224182,
      "learning_rate": 2.6836975963873843e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 13198408,
      "step": 20185
    },
    {
      "epoch": 10.581761006289309,
      "grad_norm": 0.4787271320819855,
      "learning_rate": 2.6825572396799377e-05,
      "loss": 0.3443,
      "num_input_tokens_seen": 13202184,
      "step": 20190
    },
    {
      "epoch": 10.584381551362684,
      "grad_norm": 0.6617447137832642,
      "learning_rate": 2.6814168447835986e-05,
      "loss": 0.5725,
      "num_input_tokens_seen": 13204872,
      "step": 20195
    },
    {
      "epoch": 10.58700209643606,
      "grad_norm": 0.3002294898033142,
      "learning_rate": 2.6802764119369245e-05,
      "loss": 0.3866,
      "num_input_tokens_seen": 13208296,
      "step": 20200
    },
    {
      "epoch": 10.589622641509434,
      "grad_norm": 0.31743323802948,
      "learning_rate": 2.6791359413784828e-05,
      "loss": 0.3483,
      "num_input_tokens_seen": 13211592,
      "step": 20205
    },
    {
      "epoch": 10.59224318658281,
      "grad_norm": 0.46644148230552673,
      "learning_rate": 2.677995433346845e-05,
      "loss": 0.5317,
      "num_input_tokens_seen": 13215048,
      "step": 20210
    },
    {
      "epoch": 10.594863731656185,
      "grad_norm": 0.9369986653327942,
      "learning_rate": 2.6768548880805934e-05,
      "loss": 0.5855,
      "num_input_tokens_seen": 13218760,
      "step": 20215
    },
    {
      "epoch": 10.59748427672956,
      "grad_norm": 0.4840274453163147,
      "learning_rate": 2.6757143058183177e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 13221800,
      "step": 20220
    },
    {
      "epoch": 10.600104821802935,
      "grad_norm": 0.6711065173149109,
      "learning_rate": 2.6745736867986132e-05,
      "loss": 0.5462,
      "num_input_tokens_seen": 13224872,
      "step": 20225
    },
    {
      "epoch": 10.60272536687631,
      "grad_norm": 0.5683155655860901,
      "learning_rate": 2.673433031260085e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 13227464,
      "step": 20230
    },
    {
      "epoch": 10.605345911949685,
      "grad_norm": 0.31659772992134094,
      "learning_rate": 2.6722923394413456e-05,
      "loss": 0.6154,
      "num_input_tokens_seen": 13230952,
      "step": 20235
    },
    {
      "epoch": 10.60796645702306,
      "grad_norm": 0.2818845212459564,
      "learning_rate": 2.6711516115810136e-05,
      "loss": 0.4853,
      "num_input_tokens_seen": 13235784,
      "step": 20240
    },
    {
      "epoch": 10.610587002096436,
      "grad_norm": 0.5000637173652649,
      "learning_rate": 2.6700108479177188e-05,
      "loss": 0.3074,
      "num_input_tokens_seen": 13238824,
      "step": 20245
    },
    {
      "epoch": 10.61320754716981,
      "grad_norm": 0.7680625915527344,
      "learning_rate": 2.6688700486900926e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 13241736,
      "step": 20250
    },
    {
      "epoch": 10.615828092243186,
      "grad_norm": 0.5069543719291687,
      "learning_rate": 2.6677292141367787e-05,
      "loss": 0.5031,
      "num_input_tokens_seen": 13244520,
      "step": 20255
    },
    {
      "epoch": 10.618448637316561,
      "grad_norm": 0.9803436398506165,
      "learning_rate": 2.6665883444964278e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 13248520,
      "step": 20260
    },
    {
      "epoch": 10.621069182389936,
      "grad_norm": 0.355080246925354,
      "learning_rate": 2.665447440007695e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 13251912,
      "step": 20265
    },
    {
      "epoch": 10.623689727463312,
      "grad_norm": 0.6063776016235352,
      "learning_rate": 2.664306500909246e-05,
      "loss": 0.4518,
      "num_input_tokens_seen": 13254280,
      "step": 20270
    },
    {
      "epoch": 10.626310272536688,
      "grad_norm": 0.5508469939231873,
      "learning_rate": 2.663165527439751e-05,
      "loss": 0.5012,
      "num_input_tokens_seen": 13257864,
      "step": 20275
    },
    {
      "epoch": 10.628930817610064,
      "grad_norm": 0.3880937993526459,
      "learning_rate": 2.662024519837889e-05,
      "loss": 0.4353,
      "num_input_tokens_seen": 13260520,
      "step": 20280
    },
    {
      "epoch": 10.631551362683439,
      "grad_norm": 0.5899421572685242,
      "learning_rate": 2.6608834783423465e-05,
      "loss": 0.5024,
      "num_input_tokens_seen": 13263432,
      "step": 20285
    },
    {
      "epoch": 10.634171907756814,
      "grad_norm": 0.6302187442779541,
      "learning_rate": 2.6597424031918156e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 13267112,
      "step": 20290
    },
    {
      "epoch": 10.63679245283019,
      "grad_norm": 0.8420175909996033,
      "learning_rate": 2.6586012946249954e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 13269736,
      "step": 20295
    },
    {
      "epoch": 10.639412997903564,
      "grad_norm": 0.45500680804252625,
      "learning_rate": 2.6574601528805953e-05,
      "loss": 0.6276,
      "num_input_tokens_seen": 13272520,
      "step": 20300
    },
    {
      "epoch": 10.64203354297694,
      "grad_norm": 0.9105333685874939,
      "learning_rate": 2.6563189781973268e-05,
      "loss": 0.5504,
      "num_input_tokens_seen": 13275976,
      "step": 20305
    },
    {
      "epoch": 10.644654088050315,
      "grad_norm": 0.9700660109519958,
      "learning_rate": 2.655177770813912e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 13279304,
      "step": 20310
    },
    {
      "epoch": 10.64727463312369,
      "grad_norm": 0.4896987974643707,
      "learning_rate": 2.654036530969077e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 13282888,
      "step": 20315
    },
    {
      "epoch": 10.649895178197065,
      "grad_norm": 0.3435535728931427,
      "learning_rate": 2.6528952589015576e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 13285512,
      "step": 20320
    },
    {
      "epoch": 10.65251572327044,
      "grad_norm": 0.3131081759929657,
      "learning_rate": 2.6517539548500947e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 13288904,
      "step": 20325
    },
    {
      "epoch": 10.655136268343815,
      "grad_norm": 0.30316469073295593,
      "learning_rate": 2.6506126190534354e-05,
      "loss": 0.377,
      "num_input_tokens_seen": 13291656,
      "step": 20330
    },
    {
      "epoch": 10.65775681341719,
      "grad_norm": 0.3628499209880829,
      "learning_rate": 2.649471251750334e-05,
      "loss": 0.3828,
      "num_input_tokens_seen": 13295816,
      "step": 20335
    },
    {
      "epoch": 10.660377358490566,
      "grad_norm": 0.4301076829433441,
      "learning_rate": 2.6483298531795527e-05,
      "loss": 0.5024,
      "num_input_tokens_seen": 13299112,
      "step": 20340
    },
    {
      "epoch": 10.66299790356394,
      "grad_norm": 0.46987852454185486,
      "learning_rate": 2.6471884235798583e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 13302696,
      "step": 20345
    },
    {
      "epoch": 10.665618448637316,
      "grad_norm": 0.2524541914463043,
      "learning_rate": 2.6460469631900247e-05,
      "loss": 0.413,
      "num_input_tokens_seen": 13306408,
      "step": 20350
    },
    {
      "epoch": 10.668238993710691,
      "grad_norm": 0.4233434498310089,
      "learning_rate": 2.644905472248832e-05,
      "loss": 0.3845,
      "num_input_tokens_seen": 13309160,
      "step": 20355
    },
    {
      "epoch": 10.670859538784066,
      "grad_norm": 0.3996800482273102,
      "learning_rate": 2.6437639509950686e-05,
      "loss": 0.4116,
      "num_input_tokens_seen": 13311752,
      "step": 20360
    },
    {
      "epoch": 10.673480083857442,
      "grad_norm": 0.39242154359817505,
      "learning_rate": 2.6426223996675264e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 13314792,
      "step": 20365
    },
    {
      "epoch": 10.676100628930818,
      "grad_norm": 0.363249808549881,
      "learning_rate": 2.641480818505005e-05,
      "loss": 0.4204,
      "num_input_tokens_seen": 13317832,
      "step": 20370
    },
    {
      "epoch": 10.678721174004194,
      "grad_norm": 0.5651424527168274,
      "learning_rate": 2.6403392077463108e-05,
      "loss": 0.3083,
      "num_input_tokens_seen": 13321096,
      "step": 20375
    },
    {
      "epoch": 10.681341719077569,
      "grad_norm": 0.3670152723789215,
      "learning_rate": 2.639197567630256e-05,
      "loss": 0.435,
      "num_input_tokens_seen": 13324008,
      "step": 20380
    },
    {
      "epoch": 10.683962264150944,
      "grad_norm": 0.4504646956920624,
      "learning_rate": 2.638055898395657e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 13327400,
      "step": 20385
    },
    {
      "epoch": 10.68658280922432,
      "grad_norm": 0.5260940194129944,
      "learning_rate": 2.6369142002813403e-05,
      "loss": 0.4372,
      "num_input_tokens_seen": 13330504,
      "step": 20390
    },
    {
      "epoch": 10.689203354297694,
      "grad_norm": 0.4230835735797882,
      "learning_rate": 2.6357724735261335e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 13334152,
      "step": 20395
    },
    {
      "epoch": 10.69182389937107,
      "grad_norm": 0.3724035620689392,
      "learning_rate": 2.6346307183688752e-05,
      "loss": 0.4021,
      "num_input_tokens_seen": 13337960,
      "step": 20400
    },
    {
      "epoch": 10.694444444444445,
      "grad_norm": 0.4122629761695862,
      "learning_rate": 2.6334889350484064e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 13341096,
      "step": 20405
    },
    {
      "epoch": 10.69706498951782,
      "grad_norm": 0.6640549898147583,
      "learning_rate": 2.632347123803575e-05,
      "loss": 0.3361,
      "num_input_tokens_seen": 13344744,
      "step": 20410
    },
    {
      "epoch": 10.699685534591195,
      "grad_norm": 0.8450330495834351,
      "learning_rate": 2.6312052848732353e-05,
      "loss": 0.5252,
      "num_input_tokens_seen": 13347752,
      "step": 20415
    },
    {
      "epoch": 10.70230607966457,
      "grad_norm": 0.3484645187854767,
      "learning_rate": 2.6300634184962475e-05,
      "loss": 0.4729,
      "num_input_tokens_seen": 13350696,
      "step": 20420
    },
    {
      "epoch": 10.704926624737945,
      "grad_norm": 0.801487922668457,
      "learning_rate": 2.628921524911475e-05,
      "loss": 0.5936,
      "num_input_tokens_seen": 13353032,
      "step": 20425
    },
    {
      "epoch": 10.70754716981132,
      "grad_norm": 0.4471486508846283,
      "learning_rate": 2.6277796043577924e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 13355784,
      "step": 20430
    },
    {
      "epoch": 10.710167714884696,
      "grad_norm": 0.5538091063499451,
      "learning_rate": 2.6266376570740725e-05,
      "loss": 0.4265,
      "num_input_tokens_seen": 13358408,
      "step": 20435
    },
    {
      "epoch": 10.71278825995807,
      "grad_norm": 0.3219187557697296,
      "learning_rate": 2.6254956832992005e-05,
      "loss": 0.3938,
      "num_input_tokens_seen": 13365192,
      "step": 20440
    },
    {
      "epoch": 10.715408805031446,
      "grad_norm": 0.46307799220085144,
      "learning_rate": 2.6243536832720627e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 13368808,
      "step": 20445
    },
    {
      "epoch": 10.718029350104821,
      "grad_norm": 0.41165077686309814,
      "learning_rate": 2.623211657231553e-05,
      "loss": 0.4412,
      "num_input_tokens_seen": 13371880,
      "step": 20450
    },
    {
      "epoch": 10.720649895178196,
      "grad_norm": 0.776960551738739,
      "learning_rate": 2.6220696054165704e-05,
      "loss": 0.7694,
      "num_input_tokens_seen": 13374632,
      "step": 20455
    },
    {
      "epoch": 10.723270440251572,
      "grad_norm": 0.38600462675094604,
      "learning_rate": 2.6209275280660194e-05,
      "loss": 0.5237,
      "num_input_tokens_seen": 13377960,
      "step": 20460
    },
    {
      "epoch": 10.725890985324948,
      "grad_norm": 0.397138774394989,
      "learning_rate": 2.6197854254188085e-05,
      "loss": 0.4489,
      "num_input_tokens_seen": 13381864,
      "step": 20465
    },
    {
      "epoch": 10.728511530398324,
      "grad_norm": 0.5927534699440002,
      "learning_rate": 2.618643297713854e-05,
      "loss": 0.4216,
      "num_input_tokens_seen": 13384232,
      "step": 20470
    },
    {
      "epoch": 10.731132075471699,
      "grad_norm": 0.41029852628707886,
      "learning_rate": 2.617501145190075e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 13387208,
      "step": 20475
    },
    {
      "epoch": 10.733752620545074,
      "grad_norm": 0.8189145922660828,
      "learning_rate": 2.616358968086397e-05,
      "loss": 0.5885,
      "num_input_tokens_seen": 13389992,
      "step": 20480
    },
    {
      "epoch": 10.73637316561845,
      "grad_norm": 0.49116286635398865,
      "learning_rate": 2.6152167666417493e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 13393224,
      "step": 20485
    },
    {
      "epoch": 10.738993710691824,
      "grad_norm": 0.38600051403045654,
      "learning_rate": 2.6140745410950694e-05,
      "loss": 0.6367,
      "num_input_tokens_seen": 13395880,
      "step": 20490
    },
    {
      "epoch": 10.7416142557652,
      "grad_norm": 0.7325577139854431,
      "learning_rate": 2.6129322916852973e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 13399432,
      "step": 20495
    },
    {
      "epoch": 10.744234800838575,
      "grad_norm": 0.39294251799583435,
      "learning_rate": 2.611790018651377e-05,
      "loss": 0.5216,
      "num_input_tokens_seen": 13402248,
      "step": 20500
    },
    {
      "epoch": 10.74685534591195,
      "grad_norm": 0.5060191750526428,
      "learning_rate": 2.6106477222322613e-05,
      "loss": 0.3818,
      "num_input_tokens_seen": 13404648,
      "step": 20505
    },
    {
      "epoch": 10.749475890985325,
      "grad_norm": 0.35207054018974304,
      "learning_rate": 2.6095054026669037e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 13408296,
      "step": 20510
    },
    {
      "epoch": 10.7520964360587,
      "grad_norm": 0.6919803023338318,
      "learning_rate": 2.608363060194265e-05,
      "loss": 0.4256,
      "num_input_tokens_seen": 13411560,
      "step": 20515
    },
    {
      "epoch": 10.754716981132075,
      "grad_norm": 0.5215184092521667,
      "learning_rate": 2.6072206950533117e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 13414536,
      "step": 20520
    },
    {
      "epoch": 10.75733752620545,
      "grad_norm": 0.4810032248497009,
      "learning_rate": 2.606078307483011e-05,
      "loss": 0.5174,
      "num_input_tokens_seen": 13417736,
      "step": 20525
    },
    {
      "epoch": 10.759958071278826,
      "grad_norm": 0.5793569087982178,
      "learning_rate": 2.6049358977223392e-05,
      "loss": 0.476,
      "num_input_tokens_seen": 13420744,
      "step": 20530
    },
    {
      "epoch": 10.7625786163522,
      "grad_norm": 0.46610480546951294,
      "learning_rate": 2.6037934660102747e-05,
      "loss": 0.441,
      "num_input_tokens_seen": 13425032,
      "step": 20535
    },
    {
      "epoch": 10.765199161425576,
      "grad_norm": 0.3112085163593292,
      "learning_rate": 2.6026510125858006e-05,
      "loss": 0.3857,
      "num_input_tokens_seen": 13429544,
      "step": 20540
    },
    {
      "epoch": 10.767819706498951,
      "grad_norm": 0.3544626235961914,
      "learning_rate": 2.6015085376879063e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 13434088,
      "step": 20545
    },
    {
      "epoch": 10.770440251572326,
      "grad_norm": 0.37287652492523193,
      "learning_rate": 2.600366041555584e-05,
      "loss": 0.4451,
      "num_input_tokens_seen": 13437096,
      "step": 20550
    },
    {
      "epoch": 10.773060796645701,
      "grad_norm": 0.5929825305938721,
      "learning_rate": 2.5992235244278302e-05,
      "loss": 0.4365,
      "num_input_tokens_seen": 13440680,
      "step": 20555
    },
    {
      "epoch": 10.775681341719078,
      "grad_norm": 0.35764142870903015,
      "learning_rate": 2.5980809865436485e-05,
      "loss": 0.3626,
      "num_input_tokens_seen": 13443496,
      "step": 20560
    },
    {
      "epoch": 10.778301886792454,
      "grad_norm": 0.352378785610199,
      "learning_rate": 2.5969384281420424e-05,
      "loss": 0.5239,
      "num_input_tokens_seen": 13446920,
      "step": 20565
    },
    {
      "epoch": 10.780922431865829,
      "grad_norm": 0.6557778716087341,
      "learning_rate": 2.5957958494620227e-05,
      "loss": 0.5281,
      "num_input_tokens_seen": 13449544,
      "step": 20570
    },
    {
      "epoch": 10.783542976939204,
      "grad_norm": 0.43576085567474365,
      "learning_rate": 2.594653250742605e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 13451848,
      "step": 20575
    },
    {
      "epoch": 10.786163522012579,
      "grad_norm": 0.5196729898452759,
      "learning_rate": 2.5935106322228065e-05,
      "loss": 0.4988,
      "num_input_tokens_seen": 13455272,
      "step": 20580
    },
    {
      "epoch": 10.788784067085954,
      "grad_norm": 0.5307499170303345,
      "learning_rate": 2.5923679941416514e-05,
      "loss": 0.5448,
      "num_input_tokens_seen": 13457992,
      "step": 20585
    },
    {
      "epoch": 10.79140461215933,
      "grad_norm": 0.2827240824699402,
      "learning_rate": 2.591225336738165e-05,
      "loss": 0.3447,
      "num_input_tokens_seen": 13461480,
      "step": 20590
    },
    {
      "epoch": 10.794025157232705,
      "grad_norm": 0.44392555952072144,
      "learning_rate": 2.5900826602513782e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 13464072,
      "step": 20595
    },
    {
      "epoch": 10.79664570230608,
      "grad_norm": 0.3568318486213684,
      "learning_rate": 2.5889399649203277e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 13466888,
      "step": 20600
    },
    {
      "epoch": 10.799266247379455,
      "grad_norm": 0.4711682200431824,
      "learning_rate": 2.5877972509840514e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 13469736,
      "step": 20605
    },
    {
      "epoch": 10.80188679245283,
      "grad_norm": 0.4585939943790436,
      "learning_rate": 2.5866545186815912e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 13472424,
      "step": 20610
    },
    {
      "epoch": 10.804507337526205,
      "grad_norm": 0.32347142696380615,
      "learning_rate": 2.5855117682519954e-05,
      "loss": 0.3729,
      "num_input_tokens_seen": 13476040,
      "step": 20615
    },
    {
      "epoch": 10.80712788259958,
      "grad_norm": 0.30387502908706665,
      "learning_rate": 2.584368999934312e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 13481160,
      "step": 20620
    },
    {
      "epoch": 10.809748427672956,
      "grad_norm": 0.5372463464736938,
      "learning_rate": 2.5832262139675966e-05,
      "loss": 0.3927,
      "num_input_tokens_seen": 13484904,
      "step": 20625
    },
    {
      "epoch": 10.81236897274633,
      "grad_norm": 0.47995299100875854,
      "learning_rate": 2.582083410590907e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 13487688,
      "step": 20630
    },
    {
      "epoch": 10.814989517819706,
      "grad_norm": 0.37060827016830444,
      "learning_rate": 2.5809405900433036e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 13491112,
      "step": 20635
    },
    {
      "epoch": 10.817610062893081,
      "grad_norm": 0.5367863178253174,
      "learning_rate": 2.579797752563853e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 13494248,
      "step": 20640
    },
    {
      "epoch": 10.820230607966456,
      "grad_norm": 0.3879699110984802,
      "learning_rate": 2.5786548983916232e-05,
      "loss": 0.5076,
      "num_input_tokens_seen": 13497960,
      "step": 20645
    },
    {
      "epoch": 10.822851153039831,
      "grad_norm": 0.5134609341621399,
      "learning_rate": 2.577512027765685e-05,
      "loss": 0.4745,
      "num_input_tokens_seen": 13501256,
      "step": 20650
    },
    {
      "epoch": 10.825471698113208,
      "grad_norm": 0.7596473097801208,
      "learning_rate": 2.5763691409251155e-05,
      "loss": 0.4248,
      "num_input_tokens_seen": 13503592,
      "step": 20655
    },
    {
      "epoch": 10.828092243186584,
      "grad_norm": 0.7286575436592102,
      "learning_rate": 2.5752262381089932e-05,
      "loss": 0.5803,
      "num_input_tokens_seen": 13505704,
      "step": 20660
    },
    {
      "epoch": 10.830712788259959,
      "grad_norm": 0.884445071220398,
      "learning_rate": 2.5740833195563996e-05,
      "loss": 0.4304,
      "num_input_tokens_seen": 13508712,
      "step": 20665
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 0.40457966923713684,
      "learning_rate": 2.572940385506421e-05,
      "loss": 0.5106,
      "num_input_tokens_seen": 13512136,
      "step": 20670
    },
    {
      "epoch": 10.835953878406709,
      "grad_norm": 0.4974706768989563,
      "learning_rate": 2.5717974361981457e-05,
      "loss": 0.482,
      "num_input_tokens_seen": 13514760,
      "step": 20675
    },
    {
      "epoch": 10.838574423480084,
      "grad_norm": 0.46999284625053406,
      "learning_rate": 2.5706544718706665e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 13517384,
      "step": 20680
    },
    {
      "epoch": 10.84119496855346,
      "grad_norm": 0.889163613319397,
      "learning_rate": 2.569511492763077e-05,
      "loss": 0.5073,
      "num_input_tokens_seen": 13519624,
      "step": 20685
    },
    {
      "epoch": 10.843815513626835,
      "grad_norm": 0.7135959267616272,
      "learning_rate": 2.5683684991144764e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 13522152,
      "step": 20690
    },
    {
      "epoch": 10.84643605870021,
      "grad_norm": 0.5915601849555969,
      "learning_rate": 2.5672254911639666e-05,
      "loss": 0.4262,
      "num_input_tokens_seen": 13525256,
      "step": 20695
    },
    {
      "epoch": 10.849056603773585,
      "grad_norm": 0.4696350395679474,
      "learning_rate": 2.5660824691506496e-05,
      "loss": 0.4412,
      "num_input_tokens_seen": 13528744,
      "step": 20700
    },
    {
      "epoch": 10.85167714884696,
      "grad_norm": 0.7639276385307312,
      "learning_rate": 2.5649394333136356e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 13531304,
      "step": 20705
    },
    {
      "epoch": 10.854297693920335,
      "grad_norm": 0.43019938468933105,
      "learning_rate": 2.5637963838920316e-05,
      "loss": 0.5089,
      "num_input_tokens_seen": 13534696,
      "step": 20710
    },
    {
      "epoch": 10.85691823899371,
      "grad_norm": 0.48807382583618164,
      "learning_rate": 2.5626533211249525e-05,
      "loss": 0.4751,
      "num_input_tokens_seen": 13537672,
      "step": 20715
    },
    {
      "epoch": 10.859538784067086,
      "grad_norm": 0.44855833053588867,
      "learning_rate": 2.5615102452515138e-05,
      "loss": 0.4248,
      "num_input_tokens_seen": 13542504,
      "step": 20720
    },
    {
      "epoch": 10.86215932914046,
      "grad_norm": 0.4930512607097626,
      "learning_rate": 2.5603671565108328e-05,
      "loss": 0.3319,
      "num_input_tokens_seen": 13545544,
      "step": 20725
    },
    {
      "epoch": 10.864779874213836,
      "grad_norm": 0.38510820269584656,
      "learning_rate": 2.5592240551420317e-05,
      "loss": 0.4605,
      "num_input_tokens_seen": 13551336,
      "step": 20730
    },
    {
      "epoch": 10.867400419287211,
      "grad_norm": 0.6858940720558167,
      "learning_rate": 2.5580809413842343e-05,
      "loss": 0.4349,
      "num_input_tokens_seen": 13554952,
      "step": 20735
    },
    {
      "epoch": 10.870020964360586,
      "grad_norm": 0.428423672914505,
      "learning_rate": 2.5569378154765655e-05,
      "loss": 0.4306,
      "num_input_tokens_seen": 13557864,
      "step": 20740
    },
    {
      "epoch": 10.872641509433961,
      "grad_norm": 0.4056078791618347,
      "learning_rate": 2.5557946776581572e-05,
      "loss": 0.5553,
      "num_input_tokens_seen": 13561096,
      "step": 20745
    },
    {
      "epoch": 10.875262054507338,
      "grad_norm": 0.3163509666919708,
      "learning_rate": 2.554651528168137e-05,
      "loss": 0.5008,
      "num_input_tokens_seen": 13565000,
      "step": 20750
    },
    {
      "epoch": 10.877882599580714,
      "grad_norm": 0.4463200271129608,
      "learning_rate": 2.5535083672456407e-05,
      "loss": 0.3746,
      "num_input_tokens_seen": 13568360,
      "step": 20755
    },
    {
      "epoch": 10.880503144654089,
      "grad_norm": 0.4553689658641815,
      "learning_rate": 2.5523651951298055e-05,
      "loss": 0.494,
      "num_input_tokens_seen": 13572168,
      "step": 20760
    },
    {
      "epoch": 10.883123689727464,
      "grad_norm": 0.33224183320999146,
      "learning_rate": 2.5512220120597686e-05,
      "loss": 0.387,
      "num_input_tokens_seen": 13575144,
      "step": 20765
    },
    {
      "epoch": 10.885744234800839,
      "grad_norm": 0.49584221839904785,
      "learning_rate": 2.550078818274671e-05,
      "loss": 0.4007,
      "num_input_tokens_seen": 13581832,
      "step": 20770
    },
    {
      "epoch": 10.888364779874214,
      "grad_norm": 0.31802690029144287,
      "learning_rate": 2.548935614013656e-05,
      "loss": 0.4047,
      "num_input_tokens_seen": 13585096,
      "step": 20775
    },
    {
      "epoch": 10.89098532494759,
      "grad_norm": 0.8480201363563538,
      "learning_rate": 2.5477923995158683e-05,
      "loss": 0.4864,
      "num_input_tokens_seen": 13588328,
      "step": 20780
    },
    {
      "epoch": 10.893605870020965,
      "grad_norm": 0.3817046880722046,
      "learning_rate": 2.5466491750204562e-05,
      "loss": 0.4423,
      "num_input_tokens_seen": 13592744,
      "step": 20785
    },
    {
      "epoch": 10.89622641509434,
      "grad_norm": 0.4295139014720917,
      "learning_rate": 2.5455059407665693e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 13596232,
      "step": 20790
    },
    {
      "epoch": 10.898846960167715,
      "grad_norm": 0.7668960690498352,
      "learning_rate": 2.544362696993358e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 13599880,
      "step": 20795
    },
    {
      "epoch": 10.90146750524109,
      "grad_norm": 0.37957754731178284,
      "learning_rate": 2.5432194439399775e-05,
      "loss": 0.5734,
      "num_input_tokens_seen": 13603560,
      "step": 20800
    },
    {
      "epoch": 10.904088050314465,
      "grad_norm": 0.4107905328273773,
      "learning_rate": 2.542076181845582e-05,
      "loss": 0.416,
      "num_input_tokens_seen": 13606152,
      "step": 20805
    },
    {
      "epoch": 10.90670859538784,
      "grad_norm": 0.49019134044647217,
      "learning_rate": 2.540932910949329e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 13609864,
      "step": 20810
    },
    {
      "epoch": 10.909329140461216,
      "grad_norm": 0.39410334825515747,
      "learning_rate": 2.5397896314903785e-05,
      "loss": 0.4455,
      "num_input_tokens_seen": 13612424,
      "step": 20815
    },
    {
      "epoch": 10.91194968553459,
      "grad_norm": 0.4568338096141815,
      "learning_rate": 2.5386463437078905e-05,
      "loss": 0.5764,
      "num_input_tokens_seen": 13615496,
      "step": 20820
    },
    {
      "epoch": 10.914570230607966,
      "grad_norm": 0.4492649435997009,
      "learning_rate": 2.5375030478410282e-05,
      "loss": 0.5767,
      "num_input_tokens_seen": 13618248,
      "step": 20825
    },
    {
      "epoch": 10.917190775681341,
      "grad_norm": 0.32180219888687134,
      "learning_rate": 2.536359744128957e-05,
      "loss": 0.6639,
      "num_input_tokens_seen": 13622088,
      "step": 20830
    },
    {
      "epoch": 10.919811320754716,
      "grad_norm": 0.40607011318206787,
      "learning_rate": 2.5352164328108413e-05,
      "loss": 0.3779,
      "num_input_tokens_seen": 13625544,
      "step": 20835
    },
    {
      "epoch": 10.922431865828091,
      "grad_norm": 0.3740234076976776,
      "learning_rate": 2.53407311412585e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 13629896,
      "step": 20840
    },
    {
      "epoch": 10.925052410901468,
      "grad_norm": 0.39685776829719543,
      "learning_rate": 2.5329297883131526e-05,
      "loss": 0.4828,
      "num_input_tokens_seen": 13632936,
      "step": 20845
    },
    {
      "epoch": 10.927672955974844,
      "grad_norm": 0.6589999794960022,
      "learning_rate": 2.5317864556119182e-05,
      "loss": 0.3989,
      "num_input_tokens_seen": 13635176,
      "step": 20850
    },
    {
      "epoch": 10.930293501048219,
      "grad_norm": 0.3550412952899933,
      "learning_rate": 2.53064311626132e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 13638376,
      "step": 20855
    },
    {
      "epoch": 10.932914046121594,
      "grad_norm": 0.8531113862991333,
      "learning_rate": 2.5294997705005318e-05,
      "loss": 0.4,
      "num_input_tokens_seen": 13641832,
      "step": 20860
    },
    {
      "epoch": 10.935534591194969,
      "grad_norm": 0.4743978977203369,
      "learning_rate": 2.5283564185687282e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 13644872,
      "step": 20865
    },
    {
      "epoch": 10.938155136268344,
      "grad_norm": 0.6039063334465027,
      "learning_rate": 2.5272130607050848e-05,
      "loss": 0.441,
      "num_input_tokens_seen": 13648776,
      "step": 20870
    },
    {
      "epoch": 10.94077568134172,
      "grad_norm": 1.0456072092056274,
      "learning_rate": 2.5260696971487802e-05,
      "loss": 0.4714,
      "num_input_tokens_seen": 13652648,
      "step": 20875
    },
    {
      "epoch": 10.943396226415095,
      "grad_norm": 0.5034814476966858,
      "learning_rate": 2.5249263281389923e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 13656040,
      "step": 20880
    },
    {
      "epoch": 10.94601677148847,
      "grad_norm": 0.7915456891059875,
      "learning_rate": 2.5237829539149004e-05,
      "loss": 0.4456,
      "num_input_tokens_seen": 13658888,
      "step": 20885
    },
    {
      "epoch": 10.948637316561845,
      "grad_norm": 0.32817384600639343,
      "learning_rate": 2.5226395747156866e-05,
      "loss": 0.3752,
      "num_input_tokens_seen": 13661736,
      "step": 20890
    },
    {
      "epoch": 10.95125786163522,
      "grad_norm": 0.30811986327171326,
      "learning_rate": 2.5214961907805317e-05,
      "loss": 0.3673,
      "num_input_tokens_seen": 13664872,
      "step": 20895
    },
    {
      "epoch": 10.953878406708595,
      "grad_norm": 0.6384792923927307,
      "learning_rate": 2.520352802348619e-05,
      "loss": 0.3975,
      "num_input_tokens_seen": 13668488,
      "step": 20900
    },
    {
      "epoch": 10.95649895178197,
      "grad_norm": 0.4754042327404022,
      "learning_rate": 2.5192094096591323e-05,
      "loss": 0.4061,
      "num_input_tokens_seen": 13671528,
      "step": 20905
    },
    {
      "epoch": 10.959119496855346,
      "grad_norm": 0.41762667894363403,
      "learning_rate": 2.5180660129512556e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 13674280,
      "step": 20910
    },
    {
      "epoch": 10.96174004192872,
      "grad_norm": 0.5638988018035889,
      "learning_rate": 2.516922612464176e-05,
      "loss": 0.363,
      "num_input_tokens_seen": 13677832,
      "step": 20915
    },
    {
      "epoch": 10.964360587002096,
      "grad_norm": 0.47292977571487427,
      "learning_rate": 2.5157792084370783e-05,
      "loss": 0.4064,
      "num_input_tokens_seen": 13680392,
      "step": 20920
    },
    {
      "epoch": 10.966981132075471,
      "grad_norm": 0.4025336503982544,
      "learning_rate": 2.5146358011091502e-05,
      "loss": 0.5655,
      "num_input_tokens_seen": 13684584,
      "step": 20925
    },
    {
      "epoch": 10.969601677148846,
      "grad_norm": 0.38097384572029114,
      "learning_rate": 2.5134923907195805e-05,
      "loss": 0.4155,
      "num_input_tokens_seen": 13687592,
      "step": 20930
    },
    {
      "epoch": 10.972222222222221,
      "grad_norm": 0.6717157959938049,
      "learning_rate": 2.5123489775075558e-05,
      "loss": 0.495,
      "num_input_tokens_seen": 13691816,
      "step": 20935
    },
    {
      "epoch": 10.974842767295598,
      "grad_norm": 0.3980361819267273,
      "learning_rate": 2.511205561712266e-05,
      "loss": 0.358,
      "num_input_tokens_seen": 13695432,
      "step": 20940
    },
    {
      "epoch": 10.977463312368974,
      "grad_norm": 0.4171771705150604,
      "learning_rate": 2.5100621435729004e-05,
      "loss": 0.4396,
      "num_input_tokens_seen": 13699272,
      "step": 20945
    },
    {
      "epoch": 10.980083857442349,
      "grad_norm": 0.7612091898918152,
      "learning_rate": 2.5089187233286492e-05,
      "loss": 0.4636,
      "num_input_tokens_seen": 13702568,
      "step": 20950
    },
    {
      "epoch": 10.982704402515724,
      "grad_norm": 0.5036285519599915,
      "learning_rate": 2.5077753012187026e-05,
      "loss": 0.683,
      "num_input_tokens_seen": 13706216,
      "step": 20955
    },
    {
      "epoch": 10.985324947589099,
      "grad_norm": 0.4211614429950714,
      "learning_rate": 2.506631877482252e-05,
      "loss": 0.4641,
      "num_input_tokens_seen": 13709832,
      "step": 20960
    },
    {
      "epoch": 10.987945492662474,
      "grad_norm": 0.408279150724411,
      "learning_rate": 2.505488452358488e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 13713128,
      "step": 20965
    },
    {
      "epoch": 10.99056603773585,
      "grad_norm": 0.23200656473636627,
      "learning_rate": 2.504345026086603e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 13716008,
      "step": 20970
    },
    {
      "epoch": 10.993186582809225,
      "grad_norm": 1.4718784093856812,
      "learning_rate": 2.5032015989057877e-05,
      "loss": 0.5701,
      "num_input_tokens_seen": 13719304,
      "step": 20975
    },
    {
      "epoch": 10.9958071278826,
      "grad_norm": 0.4373818039894104,
      "learning_rate": 2.5020581710552348e-05,
      "loss": 0.6561,
      "num_input_tokens_seen": 13721704,
      "step": 20980
    },
    {
      "epoch": 10.998427672955975,
      "grad_norm": 0.42297282814979553,
      "learning_rate": 2.5009147427741358e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 13724456,
      "step": 20985
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.4486265480518341,
      "eval_runtime": 9.6525,
      "eval_samples_per_second": 87.853,
      "eval_steps_per_second": 21.963,
      "num_input_tokens_seen": 13725560,
      "step": 20988
    },
    {
      "epoch": 11.00104821802935,
      "grad_norm": 0.8105343580245972,
      "learning_rate": 2.4997713143016822e-05,
      "loss": 0.3526,
      "num_input_tokens_seen": 13726872,
      "step": 20990
    },
    {
      "epoch": 11.003668763102725,
      "grad_norm": 0.3353980481624603,
      "learning_rate": 2.498627885877068e-05,
      "loss": 0.3468,
      "num_input_tokens_seen": 13729944,
      "step": 20995
    },
    {
      "epoch": 11.0062893081761,
      "grad_norm": 0.5170789957046509,
      "learning_rate": 2.4974844577394846e-05,
      "loss": 0.4101,
      "num_input_tokens_seen": 13732472,
      "step": 21000
    },
    {
      "epoch": 11.008909853249476,
      "grad_norm": 0.5579884052276611,
      "learning_rate": 2.496341030128123e-05,
      "loss": 0.5199,
      "num_input_tokens_seen": 13735704,
      "step": 21005
    },
    {
      "epoch": 11.01153039832285,
      "grad_norm": 0.8781086802482605,
      "learning_rate": 2.495197603282177e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 13738392,
      "step": 21010
    },
    {
      "epoch": 11.014150943396226,
      "grad_norm": 0.5206854343414307,
      "learning_rate": 2.4940541774408368e-05,
      "loss": 0.3926,
      "num_input_tokens_seen": 13741336,
      "step": 21015
    },
    {
      "epoch": 11.016771488469601,
      "grad_norm": 0.43987515568733215,
      "learning_rate": 2.4929107528432953e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 13744888,
      "step": 21020
    },
    {
      "epoch": 11.019392033542976,
      "grad_norm": 0.7437693476676941,
      "learning_rate": 2.4917673297287446e-05,
      "loss": 0.428,
      "num_input_tokens_seen": 13747512,
      "step": 21025
    },
    {
      "epoch": 11.022012578616351,
      "grad_norm": 0.7565900087356567,
      "learning_rate": 2.4906239083363728e-05,
      "loss": 0.3628,
      "num_input_tokens_seen": 13750456,
      "step": 21030
    },
    {
      "epoch": 11.024633123689727,
      "grad_norm": 0.47771918773651123,
      "learning_rate": 2.4894804889053732e-05,
      "loss": 0.4152,
      "num_input_tokens_seen": 13753912,
      "step": 21035
    },
    {
      "epoch": 11.027253668763104,
      "grad_norm": 0.34147825837135315,
      "learning_rate": 2.4883370716749347e-05,
      "loss": 0.3915,
      "num_input_tokens_seen": 13758232,
      "step": 21040
    },
    {
      "epoch": 11.029874213836479,
      "grad_norm": 0.5235142111778259,
      "learning_rate": 2.4871936568842486e-05,
      "loss": 0.489,
      "num_input_tokens_seen": 13760728,
      "step": 21045
    },
    {
      "epoch": 11.032494758909854,
      "grad_norm": 0.8741412162780762,
      "learning_rate": 2.4860502447725033e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 13763640,
      "step": 21050
    },
    {
      "epoch": 11.035115303983229,
      "grad_norm": 0.418516606092453,
      "learning_rate": 2.4849068355788873e-05,
      "loss": 0.5605,
      "num_input_tokens_seen": 13766968,
      "step": 21055
    },
    {
      "epoch": 11.037735849056604,
      "grad_norm": 0.7003382444381714,
      "learning_rate": 2.4837634295425892e-05,
      "loss": 0.4248,
      "num_input_tokens_seen": 13770104,
      "step": 21060
    },
    {
      "epoch": 11.04035639412998,
      "grad_norm": 0.4081780016422272,
      "learning_rate": 2.4826200269027957e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 13774392,
      "step": 21065
    },
    {
      "epoch": 11.042976939203355,
      "grad_norm": 0.2965068817138672,
      "learning_rate": 2.481476627898695e-05,
      "loss": 0.4133,
      "num_input_tokens_seen": 13778360,
      "step": 21070
    },
    {
      "epoch": 11.04559748427673,
      "grad_norm": 0.29688313603401184,
      "learning_rate": 2.4803332327694724e-05,
      "loss": 0.3361,
      "num_input_tokens_seen": 13782360,
      "step": 21075
    },
    {
      "epoch": 11.048218029350105,
      "grad_norm": 0.28591692447662354,
      "learning_rate": 2.4791898417543136e-05,
      "loss": 0.3465,
      "num_input_tokens_seen": 13785848,
      "step": 21080
    },
    {
      "epoch": 11.05083857442348,
      "grad_norm": 0.8859177231788635,
      "learning_rate": 2.4780464550924014e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 13788504,
      "step": 21085
    },
    {
      "epoch": 11.053459119496855,
      "grad_norm": 0.4439575672149658,
      "learning_rate": 2.4769030730229213e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 13791064,
      "step": 21090
    },
    {
      "epoch": 11.05607966457023,
      "grad_norm": 0.6570337414741516,
      "learning_rate": 2.475759695785054e-05,
      "loss": 0.4242,
      "num_input_tokens_seen": 13794200,
      "step": 21095
    },
    {
      "epoch": 11.058700209643606,
      "grad_norm": 0.34323084354400635,
      "learning_rate": 2.4746163236179822e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 13798232,
      "step": 21100
    },
    {
      "epoch": 11.06132075471698,
      "grad_norm": 0.3950141370296478,
      "learning_rate": 2.4734729567608858e-05,
      "loss": 0.5061,
      "num_input_tokens_seen": 13802040,
      "step": 21105
    },
    {
      "epoch": 11.063941299790356,
      "grad_norm": 0.31708452105522156,
      "learning_rate": 2.472329595452945e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 13805048,
      "step": 21110
    },
    {
      "epoch": 11.066561844863731,
      "grad_norm": 0.3507179915904999,
      "learning_rate": 2.4711862399333367e-05,
      "loss": 0.368,
      "num_input_tokens_seen": 13807608,
      "step": 21115
    },
    {
      "epoch": 11.069182389937106,
      "grad_norm": 0.5191909670829773,
      "learning_rate": 2.4700428904412377e-05,
      "loss": 0.4931,
      "num_input_tokens_seen": 13811576,
      "step": 21120
    },
    {
      "epoch": 11.071802935010481,
      "grad_norm": 0.6722913384437561,
      "learning_rate": 2.4688995472158245e-05,
      "loss": 0.4566,
      "num_input_tokens_seen": 13814360,
      "step": 21125
    },
    {
      "epoch": 11.074423480083857,
      "grad_norm": 0.42823395133018494,
      "learning_rate": 2.4677562104962712e-05,
      "loss": 0.3492,
      "num_input_tokens_seen": 13817688,
      "step": 21130
    },
    {
      "epoch": 11.077044025157234,
      "grad_norm": 0.7380668520927429,
      "learning_rate": 2.4666128805217513e-05,
      "loss": 0.3336,
      "num_input_tokens_seen": 13821016,
      "step": 21135
    },
    {
      "epoch": 11.079664570230609,
      "grad_norm": 0.46253761649131775,
      "learning_rate": 2.465469557531436e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 13823640,
      "step": 21140
    },
    {
      "epoch": 11.082285115303984,
      "grad_norm": 0.5993592739105225,
      "learning_rate": 2.4643262417644944e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 13826328,
      "step": 21145
    },
    {
      "epoch": 11.084905660377359,
      "grad_norm": 0.42436057329177856,
      "learning_rate": 2.463182933460096e-05,
      "loss": 0.4495,
      "num_input_tokens_seen": 13829528,
      "step": 21150
    },
    {
      "epoch": 11.087526205450734,
      "grad_norm": 0.47993364930152893,
      "learning_rate": 2.4620396328574087e-05,
      "loss": 0.4997,
      "num_input_tokens_seen": 13832760,
      "step": 21155
    },
    {
      "epoch": 11.09014675052411,
      "grad_norm": 0.2829667627811432,
      "learning_rate": 2.4608963401955965e-05,
      "loss": 0.3821,
      "num_input_tokens_seen": 13836664,
      "step": 21160
    },
    {
      "epoch": 11.092767295597485,
      "grad_norm": 0.4064702093601227,
      "learning_rate": 2.4597530557138247e-05,
      "loss": 0.5269,
      "num_input_tokens_seen": 13840408,
      "step": 21165
    },
    {
      "epoch": 11.09538784067086,
      "grad_norm": 0.3680547773838043,
      "learning_rate": 2.4586097796512538e-05,
      "loss": 0.4119,
      "num_input_tokens_seen": 13842968,
      "step": 21170
    },
    {
      "epoch": 11.098008385744235,
      "grad_norm": 0.4980504512786865,
      "learning_rate": 2.457466512247045e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 13845496,
      "step": 21175
    },
    {
      "epoch": 11.10062893081761,
      "grad_norm": 0.7911339998245239,
      "learning_rate": 2.4563232537403563e-05,
      "loss": 0.4863,
      "num_input_tokens_seen": 13848120,
      "step": 21180
    },
    {
      "epoch": 11.103249475890985,
      "grad_norm": 0.3665432333946228,
      "learning_rate": 2.4551800043703442e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 13851576,
      "step": 21185
    },
    {
      "epoch": 11.10587002096436,
      "grad_norm": 0.6179501414299011,
      "learning_rate": 2.454036764376165e-05,
      "loss": 0.4354,
      "num_input_tokens_seen": 13855064,
      "step": 21190
    },
    {
      "epoch": 11.108490566037736,
      "grad_norm": 0.30876657366752625,
      "learning_rate": 2.4528935339969705e-05,
      "loss": 0.3525,
      "num_input_tokens_seen": 13859864,
      "step": 21195
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 0.3695818781852722,
      "learning_rate": 2.45175031347191e-05,
      "loss": 0.3726,
      "num_input_tokens_seen": 13863000,
      "step": 21200
    },
    {
      "epoch": 11.113731656184486,
      "grad_norm": 0.29139378666877747,
      "learning_rate": 2.4506071030401344e-05,
      "loss": 0.4577,
      "num_input_tokens_seen": 13868280,
      "step": 21205
    },
    {
      "epoch": 11.116352201257861,
      "grad_norm": 0.3925284743309021,
      "learning_rate": 2.4494639029407884e-05,
      "loss": 0.5267,
      "num_input_tokens_seen": 13871640,
      "step": 21210
    },
    {
      "epoch": 11.118972746331236,
      "grad_norm": 0.721402108669281,
      "learning_rate": 2.448320713413018e-05,
      "loss": 0.5222,
      "num_input_tokens_seen": 13873752,
      "step": 21215
    },
    {
      "epoch": 11.121593291404611,
      "grad_norm": 0.5665021538734436,
      "learning_rate": 2.4471775346959654e-05,
      "loss": 0.4006,
      "num_input_tokens_seen": 13876888,
      "step": 21220
    },
    {
      "epoch": 11.124213836477987,
      "grad_norm": 0.6237539052963257,
      "learning_rate": 2.446034367028769e-05,
      "loss": 0.428,
      "num_input_tokens_seen": 13879416,
      "step": 21225
    },
    {
      "epoch": 11.126834381551364,
      "grad_norm": 0.8390718102455139,
      "learning_rate": 2.444891210650567e-05,
      "loss": 0.43,
      "num_input_tokens_seen": 13883736,
      "step": 21230
    },
    {
      "epoch": 11.129454926624739,
      "grad_norm": 0.634036123752594,
      "learning_rate": 2.4437480658004958e-05,
      "loss": 0.51,
      "num_input_tokens_seen": 13886392,
      "step": 21235
    },
    {
      "epoch": 11.132075471698114,
      "grad_norm": 0.3231872618198395,
      "learning_rate": 2.4426049327176868e-05,
      "loss": 0.4065,
      "num_input_tokens_seen": 13890680,
      "step": 21240
    },
    {
      "epoch": 11.134696016771489,
      "grad_norm": 0.2554110586643219,
      "learning_rate": 2.4414618116412716e-05,
      "loss": 0.3802,
      "num_input_tokens_seen": 13894200,
      "step": 21245
    },
    {
      "epoch": 11.137316561844864,
      "grad_norm": 0.37737682461738586,
      "learning_rate": 2.440318702810378e-05,
      "loss": 0.3237,
      "num_input_tokens_seen": 13896728,
      "step": 21250
    },
    {
      "epoch": 11.13993710691824,
      "grad_norm": 0.6159225702285767,
      "learning_rate": 2.43917560646413e-05,
      "loss": 0.401,
      "num_input_tokens_seen": 13899288,
      "step": 21255
    },
    {
      "epoch": 11.142557651991615,
      "grad_norm": 0.47337421774864197,
      "learning_rate": 2.4380325228416518e-05,
      "loss": 0.5929,
      "num_input_tokens_seen": 13903384,
      "step": 21260
    },
    {
      "epoch": 11.14517819706499,
      "grad_norm": 0.3776896893978119,
      "learning_rate": 2.4368894521820624e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 13906488,
      "step": 21265
    },
    {
      "epoch": 11.147798742138365,
      "grad_norm": 0.3487642705440521,
      "learning_rate": 2.43574639472448e-05,
      "loss": 0.3823,
      "num_input_tokens_seen": 13909720,
      "step": 21270
    },
    {
      "epoch": 11.15041928721174,
      "grad_norm": 0.5166603922843933,
      "learning_rate": 2.4346033507080178e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 13913496,
      "step": 21275
    },
    {
      "epoch": 11.153039832285115,
      "grad_norm": 0.6477637887001038,
      "learning_rate": 2.4334603203717903e-05,
      "loss": 0.6044,
      "num_input_tokens_seen": 13915928,
      "step": 21280
    },
    {
      "epoch": 11.15566037735849,
      "grad_norm": 0.8033901453018188,
      "learning_rate": 2.432317303954904e-05,
      "loss": 0.5871,
      "num_input_tokens_seen": 13919512,
      "step": 21285
    },
    {
      "epoch": 11.158280922431866,
      "grad_norm": 0.7134532332420349,
      "learning_rate": 2.4311743016964648e-05,
      "loss": 0.474,
      "num_input_tokens_seen": 13922232,
      "step": 21290
    },
    {
      "epoch": 11.16090146750524,
      "grad_norm": 0.45899108052253723,
      "learning_rate": 2.4300313138355772e-05,
      "loss": 0.5279,
      "num_input_tokens_seen": 13925624,
      "step": 21295
    },
    {
      "epoch": 11.163522012578616,
      "grad_norm": 0.444997102022171,
      "learning_rate": 2.4288883406113396e-05,
      "loss": 0.5412,
      "num_input_tokens_seen": 13929208,
      "step": 21300
    },
    {
      "epoch": 11.166142557651991,
      "grad_norm": 0.43944060802459717,
      "learning_rate": 2.4277453822628512e-05,
      "loss": 0.3957,
      "num_input_tokens_seen": 13932952,
      "step": 21305
    },
    {
      "epoch": 11.168763102725366,
      "grad_norm": 0.4586166441440582,
      "learning_rate": 2.4266024390292037e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 13935416,
      "step": 21310
    },
    {
      "epoch": 11.171383647798741,
      "grad_norm": 0.5581742525100708,
      "learning_rate": 2.425459511149489e-05,
      "loss": 0.4388,
      "num_input_tokens_seen": 13938424,
      "step": 21315
    },
    {
      "epoch": 11.174004192872117,
      "grad_norm": 0.5389298796653748,
      "learning_rate": 2.424316598862793e-05,
      "loss": 0.384,
      "num_input_tokens_seen": 13941112,
      "step": 21320
    },
    {
      "epoch": 11.176624737945493,
      "grad_norm": 0.3352477550506592,
      "learning_rate": 2.4231737024082023e-05,
      "loss": 0.576,
      "num_input_tokens_seen": 13944344,
      "step": 21325
    },
    {
      "epoch": 11.179245283018869,
      "grad_norm": 0.48184242844581604,
      "learning_rate": 2.4220308220247954e-05,
      "loss": 0.5412,
      "num_input_tokens_seen": 13947096,
      "step": 21330
    },
    {
      "epoch": 11.181865828092244,
      "grad_norm": 0.5673599243164062,
      "learning_rate": 2.4208879579516528e-05,
      "loss": 0.3949,
      "num_input_tokens_seen": 13950264,
      "step": 21335
    },
    {
      "epoch": 11.184486373165619,
      "grad_norm": 0.3152553141117096,
      "learning_rate": 2.419745110427846e-05,
      "loss": 0.4077,
      "num_input_tokens_seen": 13953656,
      "step": 21340
    },
    {
      "epoch": 11.187106918238994,
      "grad_norm": 0.6524897217750549,
      "learning_rate": 2.418602279692446e-05,
      "loss": 0.4158,
      "num_input_tokens_seen": 13956920,
      "step": 21345
    },
    {
      "epoch": 11.18972746331237,
      "grad_norm": 0.44026097655296326,
      "learning_rate": 2.417459465984521e-05,
      "loss": 0.4231,
      "num_input_tokens_seen": 13961208,
      "step": 21350
    },
    {
      "epoch": 11.192348008385745,
      "grad_norm": 0.6389864683151245,
      "learning_rate": 2.4163166695431336e-05,
      "loss": 0.6357,
      "num_input_tokens_seen": 13964216,
      "step": 21355
    },
    {
      "epoch": 11.19496855345912,
      "grad_norm": 0.4416287839412689,
      "learning_rate": 2.4151738906073446e-05,
      "loss": 0.5006,
      "num_input_tokens_seen": 13967192,
      "step": 21360
    },
    {
      "epoch": 11.197589098532495,
      "grad_norm": 2.1418368816375732,
      "learning_rate": 2.4140311294162105e-05,
      "loss": 0.5529,
      "num_input_tokens_seen": 13969624,
      "step": 21365
    },
    {
      "epoch": 11.20020964360587,
      "grad_norm": 0.5414151549339294,
      "learning_rate": 2.412888386208782e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 13972312,
      "step": 21370
    },
    {
      "epoch": 11.202830188679245,
      "grad_norm": 0.4531608819961548,
      "learning_rate": 2.4117456612241103e-05,
      "loss": 0.3166,
      "num_input_tokens_seen": 13974712,
      "step": 21375
    },
    {
      "epoch": 11.20545073375262,
      "grad_norm": 0.4580676555633545,
      "learning_rate": 2.410602954701239e-05,
      "loss": 0.5057,
      "num_input_tokens_seen": 13978552,
      "step": 21380
    },
    {
      "epoch": 11.208071278825996,
      "grad_norm": 0.6601299047470093,
      "learning_rate": 2.409460266879209e-05,
      "loss": 0.4875,
      "num_input_tokens_seen": 13982424,
      "step": 21385
    },
    {
      "epoch": 11.21069182389937,
      "grad_norm": 0.8558114171028137,
      "learning_rate": 2.4083175979970592e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 13985112,
      "step": 21390
    },
    {
      "epoch": 11.213312368972746,
      "grad_norm": 1.0250320434570312,
      "learning_rate": 2.4071749482938226e-05,
      "loss": 0.3213,
      "num_input_tokens_seen": 13987640,
      "step": 21395
    },
    {
      "epoch": 11.215932914046121,
      "grad_norm": 0.6420779824256897,
      "learning_rate": 2.4060323180085266e-05,
      "loss": 0.4086,
      "num_input_tokens_seen": 13990648,
      "step": 21400
    },
    {
      "epoch": 11.218553459119496,
      "grad_norm": 0.47111162543296814,
      "learning_rate": 2.404889707380198e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 13994360,
      "step": 21405
    },
    {
      "epoch": 11.221174004192871,
      "grad_norm": 0.4430495500564575,
      "learning_rate": 2.403747116647857e-05,
      "loss": 0.4001,
      "num_input_tokens_seen": 13997592,
      "step": 21410
    },
    {
      "epoch": 11.223794549266247,
      "grad_norm": 0.5407798886299133,
      "learning_rate": 2.4026045460505216e-05,
      "loss": 0.4218,
      "num_input_tokens_seen": 14000600,
      "step": 21415
    },
    {
      "epoch": 11.226415094339623,
      "grad_norm": 0.3052321970462799,
      "learning_rate": 2.4014619958272047e-05,
      "loss": 0.4742,
      "num_input_tokens_seen": 14005560,
      "step": 21420
    },
    {
      "epoch": 11.229035639412999,
      "grad_norm": 0.4770394265651703,
      "learning_rate": 2.4003194662169132e-05,
      "loss": 0.4591,
      "num_input_tokens_seen": 14008920,
      "step": 21425
    },
    {
      "epoch": 11.231656184486374,
      "grad_norm": 0.48779863119125366,
      "learning_rate": 2.3991769574586522e-05,
      "loss": 0.5609,
      "num_input_tokens_seen": 14011448,
      "step": 21430
    },
    {
      "epoch": 11.234276729559749,
      "grad_norm": 0.48449981212615967,
      "learning_rate": 2.3980344697914215e-05,
      "loss": 0.3584,
      "num_input_tokens_seen": 14015000,
      "step": 21435
    },
    {
      "epoch": 11.236897274633124,
      "grad_norm": 0.3596650958061218,
      "learning_rate": 2.3968920034542168e-05,
      "loss": 0.4523,
      "num_input_tokens_seen": 14019640,
      "step": 21440
    },
    {
      "epoch": 11.2395178197065,
      "grad_norm": 0.42961645126342773,
      "learning_rate": 2.3957495586860278e-05,
      "loss": 0.4854,
      "num_input_tokens_seen": 14023032,
      "step": 21445
    },
    {
      "epoch": 11.242138364779874,
      "grad_norm": 0.3888043165206909,
      "learning_rate": 2.394607135725843e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 14026040,
      "step": 21450
    },
    {
      "epoch": 11.24475890985325,
      "grad_norm": 0.38228127360343933,
      "learning_rate": 2.3934647348126426e-05,
      "loss": 0.4662,
      "num_input_tokens_seen": 14029400,
      "step": 21455
    },
    {
      "epoch": 11.247379454926625,
      "grad_norm": 0.523240864276886,
      "learning_rate": 2.3923223561854035e-05,
      "loss": 0.5407,
      "num_input_tokens_seen": 14032664,
      "step": 21460
    },
    {
      "epoch": 11.25,
      "grad_norm": 0.6023306250572205,
      "learning_rate": 2.391180000083099e-05,
      "loss": 0.4997,
      "num_input_tokens_seen": 14035320,
      "step": 21465
    },
    {
      "epoch": 11.252620545073375,
      "grad_norm": 0.2604683041572571,
      "learning_rate": 2.390037666744698e-05,
      "loss": 0.5396,
      "num_input_tokens_seen": 14038840,
      "step": 21470
    },
    {
      "epoch": 11.25524109014675,
      "grad_norm": 0.567579984664917,
      "learning_rate": 2.388895356409162e-05,
      "loss": 0.5124,
      "num_input_tokens_seen": 14042872,
      "step": 21475
    },
    {
      "epoch": 11.257861635220126,
      "grad_norm": 0.5125670433044434,
      "learning_rate": 2.3877530693154506e-05,
      "loss": 0.5137,
      "num_input_tokens_seen": 14046200,
      "step": 21480
    },
    {
      "epoch": 11.2604821802935,
      "grad_norm": 0.43837055563926697,
      "learning_rate": 2.3866108057025162e-05,
      "loss": 0.3778,
      "num_input_tokens_seen": 14048984,
      "step": 21485
    },
    {
      "epoch": 11.263102725366876,
      "grad_norm": 0.3408815562725067,
      "learning_rate": 2.385468565809307e-05,
      "loss": 0.3972,
      "num_input_tokens_seen": 14051704,
      "step": 21490
    },
    {
      "epoch": 11.265723270440251,
      "grad_norm": 0.3889835476875305,
      "learning_rate": 2.3843263498747675e-05,
      "loss": 0.3746,
      "num_input_tokens_seen": 14054264,
      "step": 21495
    },
    {
      "epoch": 11.268343815513626,
      "grad_norm": 0.34739547967910767,
      "learning_rate": 2.383184158137836e-05,
      "loss": 0.47,
      "num_input_tokens_seen": 14057336,
      "step": 21500
    },
    {
      "epoch": 11.270964360587001,
      "grad_norm": 0.4489154815673828,
      "learning_rate": 2.3820419908374465e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 14061272,
      "step": 21505
    },
    {
      "epoch": 11.273584905660377,
      "grad_norm": 0.6062411665916443,
      "learning_rate": 2.3808998482125265e-05,
      "loss": 0.3866,
      "num_input_tokens_seen": 14064824,
      "step": 21510
    },
    {
      "epoch": 11.276205450733753,
      "grad_norm": 1.1279995441436768,
      "learning_rate": 2.3797577305019987e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 14068280,
      "step": 21515
    },
    {
      "epoch": 11.278825995807129,
      "grad_norm": 0.5564109683036804,
      "learning_rate": 2.378615637944783e-05,
      "loss": 0.5573,
      "num_input_tokens_seen": 14071864,
      "step": 21520
    },
    {
      "epoch": 11.281446540880504,
      "grad_norm": 0.5589863657951355,
      "learning_rate": 2.3774735707797895e-05,
      "loss": 0.5208,
      "num_input_tokens_seen": 14075064,
      "step": 21525
    },
    {
      "epoch": 11.284067085953879,
      "grad_norm": 0.4495850205421448,
      "learning_rate": 2.376331529245928e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 14078360,
      "step": 21530
    },
    {
      "epoch": 11.286687631027254,
      "grad_norm": 0.2761431336402893,
      "learning_rate": 2.3751895135821006e-05,
      "loss": 0.4453,
      "num_input_tokens_seen": 14081816,
      "step": 21535
    },
    {
      "epoch": 11.28930817610063,
      "grad_norm": 0.2695862054824829,
      "learning_rate": 2.374047524027202e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 14085176,
      "step": 21540
    },
    {
      "epoch": 11.291928721174004,
      "grad_norm": 0.2954392433166504,
      "learning_rate": 2.3729055608201243e-05,
      "loss": 0.3456,
      "num_input_tokens_seen": 14089624,
      "step": 21545
    },
    {
      "epoch": 11.29454926624738,
      "grad_norm": 1.1526752710342407,
      "learning_rate": 2.3717636241997537e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 14092088,
      "step": 21550
    },
    {
      "epoch": 11.297169811320755,
      "grad_norm": 0.3957562744617462,
      "learning_rate": 2.3706217144049696e-05,
      "loss": 0.3268,
      "num_input_tokens_seen": 14095512,
      "step": 21555
    },
    {
      "epoch": 11.29979035639413,
      "grad_norm": 0.3653569221496582,
      "learning_rate": 2.369479831674647e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 14099192,
      "step": 21560
    },
    {
      "epoch": 11.302410901467505,
      "grad_norm": 0.6366908550262451,
      "learning_rate": 2.3683379762476555e-05,
      "loss": 0.7018,
      "num_input_tokens_seen": 14102392,
      "step": 21565
    },
    {
      "epoch": 11.30503144654088,
      "grad_norm": 0.38945573568344116,
      "learning_rate": 2.3671961483628562e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 14106712,
      "step": 21570
    },
    {
      "epoch": 11.307651991614255,
      "grad_norm": 0.39683997631073,
      "learning_rate": 2.3660543482591078e-05,
      "loss": 0.4112,
      "num_input_tokens_seen": 14110008,
      "step": 21575
    },
    {
      "epoch": 11.31027253668763,
      "grad_norm": 0.6013399958610535,
      "learning_rate": 2.364912576175261e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 14112632,
      "step": 21580
    },
    {
      "epoch": 11.312893081761006,
      "grad_norm": 0.3194696307182312,
      "learning_rate": 2.363770832350163e-05,
      "loss": 0.3708,
      "num_input_tokens_seen": 14117080,
      "step": 21585
    },
    {
      "epoch": 11.315513626834381,
      "grad_norm": 0.29793068766593933,
      "learning_rate": 2.362629117022653e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 14120568,
      "step": 21590
    },
    {
      "epoch": 11.318134171907756,
      "grad_norm": 0.4356348514556885,
      "learning_rate": 2.3614874304315633e-05,
      "loss": 0.5397,
      "num_input_tokens_seen": 14123448,
      "step": 21595
    },
    {
      "epoch": 11.320754716981131,
      "grad_norm": 0.4087345004081726,
      "learning_rate": 2.3603457728157237e-05,
      "loss": 0.5656,
      "num_input_tokens_seen": 14128056,
      "step": 21600
    },
    {
      "epoch": 11.323375262054507,
      "grad_norm": 0.5093610286712646,
      "learning_rate": 2.3592041444139547e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 14130968,
      "step": 21605
    },
    {
      "epoch": 11.325995807127883,
      "grad_norm": 0.8276435136795044,
      "learning_rate": 2.358062545465073e-05,
      "loss": 0.4656,
      "num_input_tokens_seen": 14133400,
      "step": 21610
    },
    {
      "epoch": 11.328616352201259,
      "grad_norm": 0.5154696106910706,
      "learning_rate": 2.3569209762078866e-05,
      "loss": 0.4441,
      "num_input_tokens_seen": 14136056,
      "step": 21615
    },
    {
      "epoch": 11.331236897274634,
      "grad_norm": 0.48929545283317566,
      "learning_rate": 2.355779436881201e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 14139896,
      "step": 21620
    },
    {
      "epoch": 11.333857442348009,
      "grad_norm": 0.3993755280971527,
      "learning_rate": 2.3546379277238107e-05,
      "loss": 0.5412,
      "num_input_tokens_seen": 14142840,
      "step": 21625
    },
    {
      "epoch": 11.336477987421384,
      "grad_norm": 0.37862035632133484,
      "learning_rate": 2.353496448974508e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 14145688,
      "step": 21630
    },
    {
      "epoch": 11.33909853249476,
      "grad_norm": 0.5416076183319092,
      "learning_rate": 2.3523550008720763e-05,
      "loss": 0.4996,
      "num_input_tokens_seen": 14148376,
      "step": 21635
    },
    {
      "epoch": 11.341719077568134,
      "grad_norm": 0.8296611905097961,
      "learning_rate": 2.3512135836552945e-05,
      "loss": 0.4546,
      "num_input_tokens_seen": 14150808,
      "step": 21640
    },
    {
      "epoch": 11.34433962264151,
      "grad_norm": 0.4064152240753174,
      "learning_rate": 2.350072197562933e-05,
      "loss": 0.4449,
      "num_input_tokens_seen": 14153848,
      "step": 21645
    },
    {
      "epoch": 11.346960167714885,
      "grad_norm": 0.5276814103126526,
      "learning_rate": 2.3489308428337586e-05,
      "loss": 0.3968,
      "num_input_tokens_seen": 14157016,
      "step": 21650
    },
    {
      "epoch": 11.34958071278826,
      "grad_norm": 0.3461691439151764,
      "learning_rate": 2.3477895197065275e-05,
      "loss": 0.3175,
      "num_input_tokens_seen": 14161208,
      "step": 21655
    },
    {
      "epoch": 11.352201257861635,
      "grad_norm": 0.4138348400592804,
      "learning_rate": 2.346648228419992e-05,
      "loss": 0.3806,
      "num_input_tokens_seen": 14164344,
      "step": 21660
    },
    {
      "epoch": 11.35482180293501,
      "grad_norm": 0.6223554015159607,
      "learning_rate": 2.3455069692128982e-05,
      "loss": 0.3306,
      "num_input_tokens_seen": 14166680,
      "step": 21665
    },
    {
      "epoch": 11.357442348008385,
      "grad_norm": 0.37782931327819824,
      "learning_rate": 2.3443657423239832e-05,
      "loss": 0.2697,
      "num_input_tokens_seen": 14169304,
      "step": 21670
    },
    {
      "epoch": 11.36006289308176,
      "grad_norm": 0.4892382323741913,
      "learning_rate": 2.34322454799198e-05,
      "loss": 0.5103,
      "num_input_tokens_seen": 14172376,
      "step": 21675
    },
    {
      "epoch": 11.362683438155136,
      "grad_norm": 0.2307601124048233,
      "learning_rate": 2.3420833864556133e-05,
      "loss": 0.4569,
      "num_input_tokens_seen": 14178104,
      "step": 21680
    },
    {
      "epoch": 11.365303983228511,
      "grad_norm": 0.5795360803604126,
      "learning_rate": 2.3409422579535995e-05,
      "loss": 0.3702,
      "num_input_tokens_seen": 14181144,
      "step": 21685
    },
    {
      "epoch": 11.367924528301886,
      "grad_norm": 0.9988833069801331,
      "learning_rate": 2.339801162724651e-05,
      "loss": 0.4343,
      "num_input_tokens_seen": 14183416,
      "step": 21690
    },
    {
      "epoch": 11.370545073375261,
      "grad_norm": 0.92573481798172,
      "learning_rate": 2.338660101007471e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 14186104,
      "step": 21695
    },
    {
      "epoch": 11.373165618448636,
      "grad_norm": 0.5839728116989136,
      "learning_rate": 2.3375190730407578e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 14190168,
      "step": 21700
    },
    {
      "epoch": 11.375786163522013,
      "grad_norm": 0.6464874148368835,
      "learning_rate": 2.336378079063202e-05,
      "loss": 0.4718,
      "num_input_tokens_seen": 14192856,
      "step": 21705
    },
    {
      "epoch": 11.378406708595389,
      "grad_norm": 0.41472309827804565,
      "learning_rate": 2.335237119313484e-05,
      "loss": 0.3211,
      "num_input_tokens_seen": 14197112,
      "step": 21710
    },
    {
      "epoch": 11.381027253668764,
      "grad_norm": 0.3906993269920349,
      "learning_rate": 2.3340961940302807e-05,
      "loss": 0.3518,
      "num_input_tokens_seen": 14200024,
      "step": 21715
    },
    {
      "epoch": 11.383647798742139,
      "grad_norm": 0.7621853351593018,
      "learning_rate": 2.3329553034522616e-05,
      "loss": 0.5865,
      "num_input_tokens_seen": 14203800,
      "step": 21720
    },
    {
      "epoch": 11.386268343815514,
      "grad_norm": 0.4845782518386841,
      "learning_rate": 2.3318144478180866e-05,
      "loss": 0.6635,
      "num_input_tokens_seen": 14207128,
      "step": 21725
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 0.3927880525588989,
      "learning_rate": 2.330673627366411e-05,
      "loss": 0.4451,
      "num_input_tokens_seen": 14210168,
      "step": 21730
    },
    {
      "epoch": 11.391509433962264,
      "grad_norm": 0.6840922832489014,
      "learning_rate": 2.329532842335881e-05,
      "loss": 0.4839,
      "num_input_tokens_seen": 14216184,
      "step": 21735
    },
    {
      "epoch": 11.39412997903564,
      "grad_norm": 0.2567172944545746,
      "learning_rate": 2.328392092965135e-05,
      "loss": 0.3891,
      "num_input_tokens_seen": 14219576,
      "step": 21740
    },
    {
      "epoch": 11.396750524109015,
      "grad_norm": 0.30237001180648804,
      "learning_rate": 2.3272513794928054e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 14223544,
      "step": 21745
    },
    {
      "epoch": 11.39937106918239,
      "grad_norm": 0.47170352935791016,
      "learning_rate": 2.326110702157516e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 14227384,
      "step": 21750
    },
    {
      "epoch": 11.401991614255765,
      "grad_norm": 0.38940122723579407,
      "learning_rate": 2.3249700611978847e-05,
      "loss": 0.4118,
      "num_input_tokens_seen": 14229816,
      "step": 21755
    },
    {
      "epoch": 11.40461215932914,
      "grad_norm": 0.4973548352718353,
      "learning_rate": 2.323829456852519e-05,
      "loss": 0.5633,
      "num_input_tokens_seen": 14233176,
      "step": 21760
    },
    {
      "epoch": 11.407232704402515,
      "grad_norm": 0.5219002366065979,
      "learning_rate": 2.3226888893600223e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 14236728,
      "step": 21765
    },
    {
      "epoch": 11.40985324947589,
      "grad_norm": 0.316537082195282,
      "learning_rate": 2.3215483589589865e-05,
      "loss": 0.3523,
      "num_input_tokens_seen": 14240184,
      "step": 21770
    },
    {
      "epoch": 11.412473794549266,
      "grad_norm": 0.6035292744636536,
      "learning_rate": 2.3204078658879974e-05,
      "loss": 0.3723,
      "num_input_tokens_seen": 14242808,
      "step": 21775
    },
    {
      "epoch": 11.415094339622641,
      "grad_norm": 1.008292555809021,
      "learning_rate": 2.3192674103856337e-05,
      "loss": 0.5309,
      "num_input_tokens_seen": 14245400,
      "step": 21780
    },
    {
      "epoch": 11.417714884696016,
      "grad_norm": 0.48371583223342896,
      "learning_rate": 2.318126992690467e-05,
      "loss": 0.4199,
      "num_input_tokens_seen": 14247768,
      "step": 21785
    },
    {
      "epoch": 11.420335429769391,
      "grad_norm": 0.4414806067943573,
      "learning_rate": 2.3169866130410585e-05,
      "loss": 0.3446,
      "num_input_tokens_seen": 14254520,
      "step": 21790
    },
    {
      "epoch": 11.422955974842766,
      "grad_norm": 1.4259238243103027,
      "learning_rate": 2.315846271675962e-05,
      "loss": 0.4546,
      "num_input_tokens_seen": 14257880,
      "step": 21795
    },
    {
      "epoch": 11.425576519916142,
      "grad_norm": 0.3120133876800537,
      "learning_rate": 2.3147059688337244e-05,
      "loss": 0.3525,
      "num_input_tokens_seen": 14261432,
      "step": 21800
    },
    {
      "epoch": 11.428197064989519,
      "grad_norm": 0.225173220038414,
      "learning_rate": 2.313565704752884e-05,
      "loss": 0.3753,
      "num_input_tokens_seen": 14264856,
      "step": 21805
    },
    {
      "epoch": 11.430817610062894,
      "grad_norm": 0.37742191553115845,
      "learning_rate": 2.3124254796719718e-05,
      "loss": 0.3567,
      "num_input_tokens_seen": 14267672,
      "step": 21810
    },
    {
      "epoch": 11.433438155136269,
      "grad_norm": 0.36499541997909546,
      "learning_rate": 2.3112852938295088e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 14271480,
      "step": 21815
    },
    {
      "epoch": 11.436058700209644,
      "grad_norm": 0.42812874913215637,
      "learning_rate": 2.3101451474640104e-05,
      "loss": 0.4231,
      "num_input_tokens_seen": 14277400,
      "step": 21820
    },
    {
      "epoch": 11.43867924528302,
      "grad_norm": 0.3705236315727234,
      "learning_rate": 2.3090050408139808e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 14280664,
      "step": 21825
    },
    {
      "epoch": 11.441299790356394,
      "grad_norm": 0.7457108497619629,
      "learning_rate": 2.307864974117917e-05,
      "loss": 0.4099,
      "num_input_tokens_seen": 14284696,
      "step": 21830
    },
    {
      "epoch": 11.44392033542977,
      "grad_norm": 0.3122042119503021,
      "learning_rate": 2.3067249476143087e-05,
      "loss": 0.5202,
      "num_input_tokens_seen": 14287800,
      "step": 21835
    },
    {
      "epoch": 11.446540880503145,
      "grad_norm": 0.7271007299423218,
      "learning_rate": 2.3055849615416363e-05,
      "loss": 0.5152,
      "num_input_tokens_seen": 14289848,
      "step": 21840
    },
    {
      "epoch": 11.44916142557652,
      "grad_norm": 0.557297945022583,
      "learning_rate": 2.304445016138372e-05,
      "loss": 0.3737,
      "num_input_tokens_seen": 14292920,
      "step": 21845
    },
    {
      "epoch": 11.451781970649895,
      "grad_norm": 0.8333563804626465,
      "learning_rate": 2.303305111642981e-05,
      "loss": 0.3978,
      "num_input_tokens_seen": 14296120,
      "step": 21850
    },
    {
      "epoch": 11.45440251572327,
      "grad_norm": 0.507063090801239,
      "learning_rate": 2.3021652482939144e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 14299384,
      "step": 21855
    },
    {
      "epoch": 11.457023060796645,
      "grad_norm": 0.46750831604003906,
      "learning_rate": 2.3010254263296214e-05,
      "loss": 0.4076,
      "num_input_tokens_seen": 14302968,
      "step": 21860
    },
    {
      "epoch": 11.45964360587002,
      "grad_norm": 0.6922519207000732,
      "learning_rate": 2.2998856459885396e-05,
      "loss": 0.3861,
      "num_input_tokens_seen": 14305400,
      "step": 21865
    },
    {
      "epoch": 11.462264150943396,
      "grad_norm": 0.6348571181297302,
      "learning_rate": 2.298745907509097e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 14309368,
      "step": 21870
    },
    {
      "epoch": 11.464884696016771,
      "grad_norm": 0.28867730498313904,
      "learning_rate": 2.2976062111297156e-05,
      "loss": 0.3931,
      "num_input_tokens_seen": 14312632,
      "step": 21875
    },
    {
      "epoch": 11.467505241090146,
      "grad_norm": 0.5222015976905823,
      "learning_rate": 2.296466557088805e-05,
      "loss": 0.3537,
      "num_input_tokens_seen": 14315480,
      "step": 21880
    },
    {
      "epoch": 11.470125786163521,
      "grad_norm": 1.018367886543274,
      "learning_rate": 2.2953269456247688e-05,
      "loss": 0.5202,
      "num_input_tokens_seen": 14318264,
      "step": 21885
    },
    {
      "epoch": 11.472746331236896,
      "grad_norm": 0.3958718180656433,
      "learning_rate": 2.2941873769760006e-05,
      "loss": 0.4984,
      "num_input_tokens_seen": 14320952,
      "step": 21890
    },
    {
      "epoch": 11.475366876310272,
      "grad_norm": 0.4166046977043152,
      "learning_rate": 2.2930478513808844e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 14324280,
      "step": 21895
    },
    {
      "epoch": 11.477987421383649,
      "grad_norm": 0.5159749388694763,
      "learning_rate": 2.2919083690777976e-05,
      "loss": 0.3803,
      "num_input_tokens_seen": 14326520,
      "step": 21900
    },
    {
      "epoch": 11.480607966457024,
      "grad_norm": 0.4767584502696991,
      "learning_rate": 2.2907689303051062e-05,
      "loss": 0.4743,
      "num_input_tokens_seen": 14329272,
      "step": 21905
    },
    {
      "epoch": 11.483228511530399,
      "grad_norm": 0.3555453419685364,
      "learning_rate": 2.289629535301167e-05,
      "loss": 0.4439,
      "num_input_tokens_seen": 14332312,
      "step": 21910
    },
    {
      "epoch": 11.485849056603774,
      "grad_norm": 0.9797846078872681,
      "learning_rate": 2.288490184304329e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 14334744,
      "step": 21915
    },
    {
      "epoch": 11.48846960167715,
      "grad_norm": 0.48763611912727356,
      "learning_rate": 2.2873508775529308e-05,
      "loss": 0.4504,
      "num_input_tokens_seen": 14338872,
      "step": 21920
    },
    {
      "epoch": 11.491090146750524,
      "grad_norm": 0.5498354434967041,
      "learning_rate": 2.2862116152853034e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 14342232,
      "step": 21925
    },
    {
      "epoch": 11.4937106918239,
      "grad_norm": 0.3834898769855499,
      "learning_rate": 2.285072397739767e-05,
      "loss": 0.4747,
      "num_input_tokens_seen": 14345432,
      "step": 21930
    },
    {
      "epoch": 11.496331236897275,
      "grad_norm": 0.5493269562721252,
      "learning_rate": 2.283933225154634e-05,
      "loss": 0.5954,
      "num_input_tokens_seen": 14348792,
      "step": 21935
    },
    {
      "epoch": 11.49895178197065,
      "grad_norm": 0.4421940743923187,
      "learning_rate": 2.2827940977682045e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 14351736,
      "step": 21940
    },
    {
      "epoch": 11.501572327044025,
      "grad_norm": 0.958691418170929,
      "learning_rate": 2.2816550158187717e-05,
      "loss": 0.4914,
      "num_input_tokens_seen": 14354680,
      "step": 21945
    },
    {
      "epoch": 11.5041928721174,
      "grad_norm": 0.5668125748634338,
      "learning_rate": 2.2805159795446182e-05,
      "loss": 0.5035,
      "num_input_tokens_seen": 14357432,
      "step": 21950
    },
    {
      "epoch": 11.506813417190775,
      "grad_norm": 0.22778953611850739,
      "learning_rate": 2.2793769891840184e-05,
      "loss": 0.4419,
      "num_input_tokens_seen": 14360728,
      "step": 21955
    },
    {
      "epoch": 11.50943396226415,
      "grad_norm": 0.4929937720298767,
      "learning_rate": 2.278238044975236e-05,
      "loss": 0.5359,
      "num_input_tokens_seen": 14364376,
      "step": 21960
    },
    {
      "epoch": 11.512054507337526,
      "grad_norm": 0.6157088875770569,
      "learning_rate": 2.2770991471565236e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 14367960,
      "step": 21965
    },
    {
      "epoch": 11.514675052410901,
      "grad_norm": 0.4210047721862793,
      "learning_rate": 2.2759602959661265e-05,
      "loss": 0.4895,
      "num_input_tokens_seen": 14372696,
      "step": 21970
    },
    {
      "epoch": 11.517295597484276,
      "grad_norm": 0.3386904299259186,
      "learning_rate": 2.2748214916422796e-05,
      "loss": 0.3904,
      "num_input_tokens_seen": 14376280,
      "step": 21975
    },
    {
      "epoch": 11.519916142557651,
      "grad_norm": 0.5493919253349304,
      "learning_rate": 2.2736827344232082e-05,
      "loss": 0.3606,
      "num_input_tokens_seen": 14379960,
      "step": 21980
    },
    {
      "epoch": 11.522536687631026,
      "grad_norm": 0.5037553310394287,
      "learning_rate": 2.272544024547126e-05,
      "loss": 0.4546,
      "num_input_tokens_seen": 14383000,
      "step": 21985
    },
    {
      "epoch": 11.525157232704402,
      "grad_norm": 0.3798147737979889,
      "learning_rate": 2.27140536225224e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 14386488,
      "step": 21990
    },
    {
      "epoch": 11.527777777777779,
      "grad_norm": 0.47235527634620667,
      "learning_rate": 2.270266747776744e-05,
      "loss": 0.3307,
      "num_input_tokens_seen": 14390264,
      "step": 21995
    },
    {
      "epoch": 11.530398322851154,
      "grad_norm": 0.44712236523628235,
      "learning_rate": 2.2691281813588226e-05,
      "loss": 0.3989,
      "num_input_tokens_seen": 14392920,
      "step": 22000
    },
    {
      "epoch": 11.533018867924529,
      "grad_norm": 0.4199274182319641,
      "learning_rate": 2.267989663236652e-05,
      "loss": 0.4232,
      "num_input_tokens_seen": 14396568,
      "step": 22005
    },
    {
      "epoch": 11.535639412997904,
      "grad_norm": 0.47553035616874695,
      "learning_rate": 2.2668511936483975e-05,
      "loss": 0.4789,
      "num_input_tokens_seen": 14400120,
      "step": 22010
    },
    {
      "epoch": 11.53825995807128,
      "grad_norm": 0.45857080817222595,
      "learning_rate": 2.2657127728322124e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 14403704,
      "step": 22015
    },
    {
      "epoch": 11.540880503144654,
      "grad_norm": 0.5157606601715088,
      "learning_rate": 2.264574401026244e-05,
      "loss": 0.3665,
      "num_input_tokens_seen": 14405976,
      "step": 22020
    },
    {
      "epoch": 11.54350104821803,
      "grad_norm": 0.5081661343574524,
      "learning_rate": 2.2634360784686242e-05,
      "loss": 0.3808,
      "num_input_tokens_seen": 14408600,
      "step": 22025
    },
    {
      "epoch": 11.546121593291405,
      "grad_norm": 0.38892701268196106,
      "learning_rate": 2.2622978053974776e-05,
      "loss": 0.3361,
      "num_input_tokens_seen": 14411352,
      "step": 22030
    },
    {
      "epoch": 11.54874213836478,
      "grad_norm": 0.8082579970359802,
      "learning_rate": 2.2611595820509186e-05,
      "loss": 0.618,
      "num_input_tokens_seen": 14414552,
      "step": 22035
    },
    {
      "epoch": 11.551362683438155,
      "grad_norm": 0.31690889596939087,
      "learning_rate": 2.2600214086670494e-05,
      "loss": 0.3795,
      "num_input_tokens_seen": 14418680,
      "step": 22040
    },
    {
      "epoch": 11.55398322851153,
      "grad_norm": 0.5013225078582764,
      "learning_rate": 2.2588832854839644e-05,
      "loss": 0.5063,
      "num_input_tokens_seen": 14421432,
      "step": 22045
    },
    {
      "epoch": 11.556603773584905,
      "grad_norm": 0.5753999948501587,
      "learning_rate": 2.2577452127397463e-05,
      "loss": 0.5231,
      "num_input_tokens_seen": 14424280,
      "step": 22050
    },
    {
      "epoch": 11.55922431865828,
      "grad_norm": 0.6047230362892151,
      "learning_rate": 2.256607190672464e-05,
      "loss": 0.5202,
      "num_input_tokens_seen": 14427064,
      "step": 22055
    },
    {
      "epoch": 11.561844863731656,
      "grad_norm": 0.4393067955970764,
      "learning_rate": 2.2554692195201817e-05,
      "loss": 0.4087,
      "num_input_tokens_seen": 14430584,
      "step": 22060
    },
    {
      "epoch": 11.564465408805031,
      "grad_norm": 0.5732552409172058,
      "learning_rate": 2.254331299520948e-05,
      "loss": 0.4742,
      "num_input_tokens_seen": 14433496,
      "step": 22065
    },
    {
      "epoch": 11.567085953878406,
      "grad_norm": 0.8410393595695496,
      "learning_rate": 2.253193430912805e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 14436632,
      "step": 22070
    },
    {
      "epoch": 11.569706498951781,
      "grad_norm": 0.9354249835014343,
      "learning_rate": 2.2520556139337803e-05,
      "loss": 0.3926,
      "num_input_tokens_seen": 14439448,
      "step": 22075
    },
    {
      "epoch": 11.572327044025156,
      "grad_norm": 0.5942001342773438,
      "learning_rate": 2.250917848821892e-05,
      "loss": 0.4579,
      "num_input_tokens_seen": 14442328,
      "step": 22080
    },
    {
      "epoch": 11.574947589098532,
      "grad_norm": 0.45374825596809387,
      "learning_rate": 2.2497801358151472e-05,
      "loss": 0.4258,
      "num_input_tokens_seen": 14444984,
      "step": 22085
    },
    {
      "epoch": 11.577568134171909,
      "grad_norm": 0.29925110936164856,
      "learning_rate": 2.248642475151545e-05,
      "loss": 0.5882,
      "num_input_tokens_seen": 14448280,
      "step": 22090
    },
    {
      "epoch": 11.580188679245284,
      "grad_norm": 0.5239774584770203,
      "learning_rate": 2.2475048670690678e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 14451864,
      "step": 22095
    },
    {
      "epoch": 11.582809224318659,
      "grad_norm": 0.7624487280845642,
      "learning_rate": 2.2463673118056927e-05,
      "loss": 0.4758,
      "num_input_tokens_seen": 14454392,
      "step": 22100
    },
    {
      "epoch": 11.585429769392034,
      "grad_norm": 0.5997311472892761,
      "learning_rate": 2.245229809599383e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 14459928,
      "step": 22105
    },
    {
      "epoch": 11.58805031446541,
      "grad_norm": 0.4307173192501068,
      "learning_rate": 2.2440923606880896e-05,
      "loss": 0.4567,
      "num_input_tokens_seen": 14462616,
      "step": 22110
    },
    {
      "epoch": 11.590670859538784,
      "grad_norm": 0.28718242049217224,
      "learning_rate": 2.242954965309755e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 14466072,
      "step": 22115
    },
    {
      "epoch": 11.59329140461216,
      "grad_norm": 0.6460992693901062,
      "learning_rate": 2.2418176237023088e-05,
      "loss": 0.394,
      "num_input_tokens_seen": 14468728,
      "step": 22120
    },
    {
      "epoch": 11.595911949685535,
      "grad_norm": 0.4948977828025818,
      "learning_rate": 2.2406803361036706e-05,
      "loss": 0.4776,
      "num_input_tokens_seen": 14471672,
      "step": 22125
    },
    {
      "epoch": 11.59853249475891,
      "grad_norm": 0.7431184649467468,
      "learning_rate": 2.2395431027517475e-05,
      "loss": 0.3504,
      "num_input_tokens_seen": 14474840,
      "step": 22130
    },
    {
      "epoch": 11.601153039832285,
      "grad_norm": 0.45607462525367737,
      "learning_rate": 2.2384059238844364e-05,
      "loss": 0.5179,
      "num_input_tokens_seen": 14478648,
      "step": 22135
    },
    {
      "epoch": 11.60377358490566,
      "grad_norm": 0.6440909504890442,
      "learning_rate": 2.2372687997396215e-05,
      "loss": 0.3277,
      "num_input_tokens_seen": 14482968,
      "step": 22140
    },
    {
      "epoch": 11.606394129979035,
      "grad_norm": 0.36168068647384644,
      "learning_rate": 2.2361317305551756e-05,
      "loss": 0.3307,
      "num_input_tokens_seen": 14486712,
      "step": 22145
    },
    {
      "epoch": 11.60901467505241,
      "grad_norm": 0.5438834428787231,
      "learning_rate": 2.2349947165689615e-05,
      "loss": 0.4934,
      "num_input_tokens_seen": 14489624,
      "step": 22150
    },
    {
      "epoch": 11.611635220125786,
      "grad_norm": 0.5733595490455627,
      "learning_rate": 2.2338577580188296e-05,
      "loss": 0.4425,
      "num_input_tokens_seen": 14493016,
      "step": 22155
    },
    {
      "epoch": 11.614255765199161,
      "grad_norm": 0.6962007880210876,
      "learning_rate": 2.2327208551426188e-05,
      "loss": 0.3315,
      "num_input_tokens_seen": 14495928,
      "step": 22160
    },
    {
      "epoch": 11.616876310272536,
      "grad_norm": 0.4199713468551636,
      "learning_rate": 2.2315840081781546e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 14499416,
      "step": 22165
    },
    {
      "epoch": 11.619496855345911,
      "grad_norm": 0.4935166835784912,
      "learning_rate": 2.230447217363255e-05,
      "loss": 0.6566,
      "num_input_tokens_seen": 14502232,
      "step": 22170
    },
    {
      "epoch": 11.622117400419286,
      "grad_norm": 0.5987290740013123,
      "learning_rate": 2.2293104829357212e-05,
      "loss": 0.5584,
      "num_input_tokens_seen": 14505144,
      "step": 22175
    },
    {
      "epoch": 11.624737945492662,
      "grad_norm": 0.31061387062072754,
      "learning_rate": 2.2281738051333467e-05,
      "loss": 0.487,
      "num_input_tokens_seen": 14509208,
      "step": 22180
    },
    {
      "epoch": 11.627358490566039,
      "grad_norm": 0.4713976979255676,
      "learning_rate": 2.2270371841939105e-05,
      "loss": 0.4161,
      "num_input_tokens_seen": 14512312,
      "step": 22185
    },
    {
      "epoch": 11.629979035639414,
      "grad_norm": 0.27758800983428955,
      "learning_rate": 2.2259006203551824e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 14515960,
      "step": 22190
    },
    {
      "epoch": 11.632599580712789,
      "grad_norm": 0.5071659088134766,
      "learning_rate": 2.2247641138549168e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 14518968,
      "step": 22195
    },
    {
      "epoch": 11.635220125786164,
      "grad_norm": 0.3747135400772095,
      "learning_rate": 2.2236276649308583e-05,
      "loss": 0.5336,
      "num_input_tokens_seen": 14522936,
      "step": 22200
    },
    {
      "epoch": 11.63784067085954,
      "grad_norm": 0.44412699341773987,
      "learning_rate": 2.2224912738207395e-05,
      "loss": 0.6065,
      "num_input_tokens_seen": 14525592,
      "step": 22205
    },
    {
      "epoch": 11.640461215932914,
      "grad_norm": 0.6607299447059631,
      "learning_rate": 2.2213549407622796e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 14529112,
      "step": 22210
    },
    {
      "epoch": 11.64308176100629,
      "grad_norm": 0.8456178307533264,
      "learning_rate": 2.220218665993188e-05,
      "loss": 0.3747,
      "num_input_tokens_seen": 14532536,
      "step": 22215
    },
    {
      "epoch": 11.645702306079665,
      "grad_norm": 0.701853334903717,
      "learning_rate": 2.21908244975116e-05,
      "loss": 0.4097,
      "num_input_tokens_seen": 14535672,
      "step": 22220
    },
    {
      "epoch": 11.64832285115304,
      "grad_norm": 0.4575992822647095,
      "learning_rate": 2.2179462922738776e-05,
      "loss": 0.4749,
      "num_input_tokens_seen": 14538712,
      "step": 22225
    },
    {
      "epoch": 11.650943396226415,
      "grad_norm": 0.4946471154689789,
      "learning_rate": 2.2168101937990135e-05,
      "loss": 0.4072,
      "num_input_tokens_seen": 14541624,
      "step": 22230
    },
    {
      "epoch": 11.65356394129979,
      "grad_norm": 0.4166795611381531,
      "learning_rate": 2.2156741545642253e-05,
      "loss": 0.3951,
      "num_input_tokens_seen": 14544600,
      "step": 22235
    },
    {
      "epoch": 11.656184486373165,
      "grad_norm": 0.41446036100387573,
      "learning_rate": 2.2145381748071606e-05,
      "loss": 0.3557,
      "num_input_tokens_seen": 14547800,
      "step": 22240
    },
    {
      "epoch": 11.65880503144654,
      "grad_norm": 0.4670511782169342,
      "learning_rate": 2.2134022547654538e-05,
      "loss": 0.5899,
      "num_input_tokens_seen": 14550072,
      "step": 22245
    },
    {
      "epoch": 11.661425576519916,
      "grad_norm": 0.518021285533905,
      "learning_rate": 2.212266394676725e-05,
      "loss": 0.4404,
      "num_input_tokens_seen": 14554584,
      "step": 22250
    },
    {
      "epoch": 11.664046121593291,
      "grad_norm": 0.3064095079898834,
      "learning_rate": 2.2111305947785836e-05,
      "loss": 0.435,
      "num_input_tokens_seen": 14557912,
      "step": 22255
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 0.3731432855129242,
      "learning_rate": 2.2099948553086263e-05,
      "loss": 0.4067,
      "num_input_tokens_seen": 14561304,
      "step": 22260
    },
    {
      "epoch": 11.669287211740041,
      "grad_norm": 0.43466755747795105,
      "learning_rate": 2.2088591765044365e-05,
      "loss": 0.5274,
      "num_input_tokens_seen": 14564856,
      "step": 22265
    },
    {
      "epoch": 11.671907756813416,
      "grad_norm": 0.44143226742744446,
      "learning_rate": 2.207723558603586e-05,
      "loss": 0.4299,
      "num_input_tokens_seen": 14567960,
      "step": 22270
    },
    {
      "epoch": 11.674528301886792,
      "grad_norm": 0.42794620990753174,
      "learning_rate": 2.206588001843633e-05,
      "loss": 0.4143,
      "num_input_tokens_seen": 14570840,
      "step": 22275
    },
    {
      "epoch": 11.677148846960169,
      "grad_norm": 0.49188950657844543,
      "learning_rate": 2.2054525064621216e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 14575128,
      "step": 22280
    },
    {
      "epoch": 11.679769392033544,
      "grad_norm": 0.48581641912460327,
      "learning_rate": 2.2043170726965858e-05,
      "loss": 0.5362,
      "num_input_tokens_seen": 14578680,
      "step": 22285
    },
    {
      "epoch": 11.682389937106919,
      "grad_norm": 0.7111713886260986,
      "learning_rate": 2.2031817007845447e-05,
      "loss": 0.5234,
      "num_input_tokens_seen": 14581400,
      "step": 22290
    },
    {
      "epoch": 11.685010482180294,
      "grad_norm": 0.5827444195747375,
      "learning_rate": 2.202046390963506e-05,
      "loss": 0.4164,
      "num_input_tokens_seen": 14584600,
      "step": 22295
    },
    {
      "epoch": 11.68763102725367,
      "grad_norm": 0.5371038317680359,
      "learning_rate": 2.2009111434709627e-05,
      "loss": 0.4283,
      "num_input_tokens_seen": 14587704,
      "step": 22300
    },
    {
      "epoch": 11.690251572327044,
      "grad_norm": 0.4843544661998749,
      "learning_rate": 2.1997759585443968e-05,
      "loss": 0.4288,
      "num_input_tokens_seen": 14590264,
      "step": 22305
    },
    {
      "epoch": 11.69287211740042,
      "grad_norm": 0.5661599040031433,
      "learning_rate": 2.198640836421275e-05,
      "loss": 0.4399,
      "num_input_tokens_seen": 14593528,
      "step": 22310
    },
    {
      "epoch": 11.695492662473795,
      "grad_norm": 0.4956153333187103,
      "learning_rate": 2.1975057773390515e-05,
      "loss": 0.4953,
      "num_input_tokens_seen": 14596568,
      "step": 22315
    },
    {
      "epoch": 11.69811320754717,
      "grad_norm": 0.9553708434104919,
      "learning_rate": 2.1963707815351683e-05,
      "loss": 0.5766,
      "num_input_tokens_seen": 14599288,
      "step": 22320
    },
    {
      "epoch": 11.700733752620545,
      "grad_norm": 0.692236602306366,
      "learning_rate": 2.1952358492470544e-05,
      "loss": 0.5276,
      "num_input_tokens_seen": 14602648,
      "step": 22325
    },
    {
      "epoch": 11.70335429769392,
      "grad_norm": 0.36477839946746826,
      "learning_rate": 2.194100980712124e-05,
      "loss": 0.383,
      "num_input_tokens_seen": 14605848,
      "step": 22330
    },
    {
      "epoch": 11.705974842767295,
      "grad_norm": 0.2751690447330475,
      "learning_rate": 2.1929661761677774e-05,
      "loss": 0.326,
      "num_input_tokens_seen": 14609112,
      "step": 22335
    },
    {
      "epoch": 11.70859538784067,
      "grad_norm": 0.547255277633667,
      "learning_rate": 2.191831435851405e-05,
      "loss": 0.7276,
      "num_input_tokens_seen": 14613208,
      "step": 22340
    },
    {
      "epoch": 11.711215932914046,
      "grad_norm": 0.24156633019447327,
      "learning_rate": 2.1906967600003793e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 14616152,
      "step": 22345
    },
    {
      "epoch": 11.713836477987421,
      "grad_norm": 0.2986123859882355,
      "learning_rate": 2.1895621488520636e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 14619128,
      "step": 22350
    },
    {
      "epoch": 11.716457023060796,
      "grad_norm": 0.9610064625740051,
      "learning_rate": 2.1884276026438044e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 14623192,
      "step": 22355
    },
    {
      "epoch": 11.719077568134171,
      "grad_norm": 0.39493700861930847,
      "learning_rate": 2.187293121612937e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 14626488,
      "step": 22360
    },
    {
      "epoch": 11.721698113207546,
      "grad_norm": 0.4181646406650543,
      "learning_rate": 2.1861587059967804e-05,
      "loss": 0.5682,
      "num_input_tokens_seen": 14629496,
      "step": 22365
    },
    {
      "epoch": 11.724318658280922,
      "grad_norm": 0.34643635153770447,
      "learning_rate": 2.1850243560326414e-05,
      "loss": 0.417,
      "num_input_tokens_seen": 14632504,
      "step": 22370
    },
    {
      "epoch": 11.726939203354299,
      "grad_norm": 0.44222909212112427,
      "learning_rate": 2.1838900719578143e-05,
      "loss": 0.4416,
      "num_input_tokens_seen": 14635576,
      "step": 22375
    },
    {
      "epoch": 11.729559748427674,
      "grad_norm": 0.3061479330062866,
      "learning_rate": 2.1827558540095772e-05,
      "loss": 0.3947,
      "num_input_tokens_seen": 14638488,
      "step": 22380
    },
    {
      "epoch": 11.732180293501049,
      "grad_norm": 0.3539217412471771,
      "learning_rate": 2.1816217024251966e-05,
      "loss": 0.5535,
      "num_input_tokens_seen": 14641720,
      "step": 22385
    },
    {
      "epoch": 11.734800838574424,
      "grad_norm": 0.3694368302822113,
      "learning_rate": 2.1804876174419242e-05,
      "loss": 0.4709,
      "num_input_tokens_seen": 14644792,
      "step": 22390
    },
    {
      "epoch": 11.7374213836478,
      "grad_norm": 0.4531247615814209,
      "learning_rate": 2.179353599296996e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 14647384,
      "step": 22395
    },
    {
      "epoch": 11.740041928721174,
      "grad_norm": 0.5676102638244629,
      "learning_rate": 2.1782196482276368e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 14650584,
      "step": 22400
    },
    {
      "epoch": 11.74266247379455,
      "grad_norm": 0.6486446261405945,
      "learning_rate": 2.1770857644710567e-05,
      "loss": 0.4093,
      "num_input_tokens_seen": 14653464,
      "step": 22405
    },
    {
      "epoch": 11.745283018867925,
      "grad_norm": 0.5295445322990417,
      "learning_rate": 2.1759519482644495e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 14656376,
      "step": 22410
    },
    {
      "epoch": 11.7479035639413,
      "grad_norm": 0.6174098253250122,
      "learning_rate": 2.174818199844999e-05,
      "loss": 0.3064,
      "num_input_tokens_seen": 14659416,
      "step": 22415
    },
    {
      "epoch": 11.750524109014675,
      "grad_norm": 0.5623379349708557,
      "learning_rate": 2.173684519449872e-05,
      "loss": 0.3933,
      "num_input_tokens_seen": 14662232,
      "step": 22420
    },
    {
      "epoch": 11.75314465408805,
      "grad_norm": 0.430928111076355,
      "learning_rate": 2.1725509073162188e-05,
      "loss": 0.4512,
      "num_input_tokens_seen": 14664920,
      "step": 22425
    },
    {
      "epoch": 11.755765199161425,
      "grad_norm": 0.5929812788963318,
      "learning_rate": 2.171417363681181e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 14668056,
      "step": 22430
    },
    {
      "epoch": 11.7583857442348,
      "grad_norm": 0.4256919026374817,
      "learning_rate": 2.1702838887818816e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 14671032,
      "step": 22435
    },
    {
      "epoch": 11.761006289308176,
      "grad_norm": 0.2681903839111328,
      "learning_rate": 2.1691504828554315e-05,
      "loss": 0.3838,
      "num_input_tokens_seen": 14674360,
      "step": 22440
    },
    {
      "epoch": 11.76362683438155,
      "grad_norm": 0.39133384823799133,
      "learning_rate": 2.1680171461389265e-05,
      "loss": 0.535,
      "num_input_tokens_seen": 14677720,
      "step": 22445
    },
    {
      "epoch": 11.766247379454926,
      "grad_norm": 0.5352754592895508,
      "learning_rate": 2.166883878869446e-05,
      "loss": 0.3987,
      "num_input_tokens_seen": 14680216,
      "step": 22450
    },
    {
      "epoch": 11.768867924528301,
      "grad_norm": 0.39802226424217224,
      "learning_rate": 2.1657506812840582e-05,
      "loss": 0.4758,
      "num_input_tokens_seen": 14683608,
      "step": 22455
    },
    {
      "epoch": 11.771488469601676,
      "grad_norm": 0.6016162633895874,
      "learning_rate": 2.164617553619814e-05,
      "loss": 0.4934,
      "num_input_tokens_seen": 14686872,
      "step": 22460
    },
    {
      "epoch": 11.774109014675052,
      "grad_norm": 0.5404260754585266,
      "learning_rate": 2.1634844961137517e-05,
      "loss": 0.3988,
      "num_input_tokens_seen": 14688824,
      "step": 22465
    },
    {
      "epoch": 11.776729559748428,
      "grad_norm": 0.6064445972442627,
      "learning_rate": 2.162351509002893e-05,
      "loss": 0.3379,
      "num_input_tokens_seen": 14691192,
      "step": 22470
    },
    {
      "epoch": 11.779350104821804,
      "grad_norm": 0.6038731932640076,
      "learning_rate": 2.161218592524248e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 14694072,
      "step": 22475
    },
    {
      "epoch": 11.781970649895179,
      "grad_norm": 0.3626493513584137,
      "learning_rate": 2.160085746914807e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 14697976,
      "step": 22480
    },
    {
      "epoch": 11.784591194968554,
      "grad_norm": 0.6789829730987549,
      "learning_rate": 2.1589529724115503e-05,
      "loss": 0.4455,
      "num_input_tokens_seen": 14701368,
      "step": 22485
    },
    {
      "epoch": 11.78721174004193,
      "grad_norm": 0.7592145204544067,
      "learning_rate": 2.1578202692514407e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 14704888,
      "step": 22490
    },
    {
      "epoch": 11.789832285115304,
      "grad_norm": 0.5370805263519287,
      "learning_rate": 2.1566876376714274e-05,
      "loss": 0.4023,
      "num_input_tokens_seen": 14707928,
      "step": 22495
    },
    {
      "epoch": 11.79245283018868,
      "grad_norm": 0.33214595913887024,
      "learning_rate": 2.155555077908443e-05,
      "loss": 0.5615,
      "num_input_tokens_seen": 14711800,
      "step": 22500
    },
    {
      "epoch": 11.795073375262055,
      "grad_norm": 0.2785264849662781,
      "learning_rate": 2.1544225901994077e-05,
      "loss": 0.5419,
      "num_input_tokens_seen": 14719000,
      "step": 22505
    },
    {
      "epoch": 11.79769392033543,
      "grad_norm": 0.3387577533721924,
      "learning_rate": 2.1532901747812234e-05,
      "loss": 0.3346,
      "num_input_tokens_seen": 14723192,
      "step": 22510
    },
    {
      "epoch": 11.800314465408805,
      "grad_norm": 0.7628806233406067,
      "learning_rate": 2.1521578318907787e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 14725496,
      "step": 22515
    },
    {
      "epoch": 11.80293501048218,
      "grad_norm": 0.3920528292655945,
      "learning_rate": 2.1510255617649476e-05,
      "loss": 0.4377,
      "num_input_tokens_seen": 14729272,
      "step": 22520
    },
    {
      "epoch": 11.805555555555555,
      "grad_norm": 0.4239419102668762,
      "learning_rate": 2.149893364640587e-05,
      "loss": 0.5318,
      "num_input_tokens_seen": 14732280,
      "step": 22525
    },
    {
      "epoch": 11.80817610062893,
      "grad_norm": 0.2460634410381317,
      "learning_rate": 2.1487612407545417e-05,
      "loss": 0.3507,
      "num_input_tokens_seen": 14737784,
      "step": 22530
    },
    {
      "epoch": 11.810796645702306,
      "grad_norm": 0.3057226836681366,
      "learning_rate": 2.147629190343637e-05,
      "loss": 0.5089,
      "num_input_tokens_seen": 14741048,
      "step": 22535
    },
    {
      "epoch": 11.81341719077568,
      "grad_norm": 0.3706178665161133,
      "learning_rate": 2.1464972136446852e-05,
      "loss": 0.4067,
      "num_input_tokens_seen": 14744440,
      "step": 22540
    },
    {
      "epoch": 11.816037735849056,
      "grad_norm": 0.4723772406578064,
      "learning_rate": 2.1453653108944838e-05,
      "loss": 0.6067,
      "num_input_tokens_seen": 14747032,
      "step": 22545
    },
    {
      "epoch": 11.818658280922431,
      "grad_norm": 0.758939266204834,
      "learning_rate": 2.1442334823298128e-05,
      "loss": 0.514,
      "num_input_tokens_seen": 14749528,
      "step": 22550
    },
    {
      "epoch": 11.821278825995806,
      "grad_norm": 0.415795236825943,
      "learning_rate": 2.1431017281874386e-05,
      "loss": 0.5176,
      "num_input_tokens_seen": 14752696,
      "step": 22555
    },
    {
      "epoch": 11.823899371069182,
      "grad_norm": 0.43836069107055664,
      "learning_rate": 2.1419700487041123e-05,
      "loss": 0.4858,
      "num_input_tokens_seen": 14755800,
      "step": 22560
    },
    {
      "epoch": 11.826519916142558,
      "grad_norm": 0.36292141675949097,
      "learning_rate": 2.140838444116566e-05,
      "loss": 0.3685,
      "num_input_tokens_seen": 14758264,
      "step": 22565
    },
    {
      "epoch": 11.829140461215934,
      "grad_norm": 0.47386810183525085,
      "learning_rate": 2.13970691466152e-05,
      "loss": 0.3922,
      "num_input_tokens_seen": 14760984,
      "step": 22570
    },
    {
      "epoch": 11.831761006289309,
      "grad_norm": 1.1377826929092407,
      "learning_rate": 2.1385754605756773e-05,
      "loss": 0.4166,
      "num_input_tokens_seen": 14764472,
      "step": 22575
    },
    {
      "epoch": 11.834381551362684,
      "grad_norm": 0.44184258580207825,
      "learning_rate": 2.137444082095724e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 14767960,
      "step": 22580
    },
    {
      "epoch": 11.83700209643606,
      "grad_norm": 0.3041076958179474,
      "learning_rate": 2.1363127794583335e-05,
      "loss": 0.4608,
      "num_input_tokens_seen": 14770360,
      "step": 22585
    },
    {
      "epoch": 11.839622641509434,
      "grad_norm": 0.456404447555542,
      "learning_rate": 2.1351815529001605e-05,
      "loss": 0.527,
      "num_input_tokens_seen": 14772920,
      "step": 22590
    },
    {
      "epoch": 11.84224318658281,
      "grad_norm": 0.389553964138031,
      "learning_rate": 2.1340504026578438e-05,
      "loss": 0.4435,
      "num_input_tokens_seen": 14776024,
      "step": 22595
    },
    {
      "epoch": 11.844863731656185,
      "grad_norm": 0.4950973391532898,
      "learning_rate": 2.1329193289680084e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 14778776,
      "step": 22600
    },
    {
      "epoch": 11.84748427672956,
      "grad_norm": 0.6405431032180786,
      "learning_rate": 2.1317883320672608e-05,
      "loss": 0.3988,
      "num_input_tokens_seen": 14781176,
      "step": 22605
    },
    {
      "epoch": 11.850104821802935,
      "grad_norm": 0.7900686860084534,
      "learning_rate": 2.1306574121921942e-05,
      "loss": 0.5317,
      "num_input_tokens_seen": 14784280,
      "step": 22610
    },
    {
      "epoch": 11.85272536687631,
      "grad_norm": 0.30113062262535095,
      "learning_rate": 2.1295265695793836e-05,
      "loss": 0.3931,
      "num_input_tokens_seen": 14788184,
      "step": 22615
    },
    {
      "epoch": 11.855345911949685,
      "grad_norm": 0.3826873302459717,
      "learning_rate": 2.128395804465387e-05,
      "loss": 0.5079,
      "num_input_tokens_seen": 14792472,
      "step": 22620
    },
    {
      "epoch": 11.85796645702306,
      "grad_norm": 0.3733917474746704,
      "learning_rate": 2.1272651170867487e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 14795416,
      "step": 22625
    },
    {
      "epoch": 11.860587002096436,
      "grad_norm": 0.4572017788887024,
      "learning_rate": 2.1261345076799955e-05,
      "loss": 0.4494,
      "num_input_tokens_seen": 14799864,
      "step": 22630
    },
    {
      "epoch": 11.86320754716981,
      "grad_norm": 0.5351228713989258,
      "learning_rate": 2.125003976481638e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 14803384,
      "step": 22635
    },
    {
      "epoch": 11.865828092243186,
      "grad_norm": 0.5400452017784119,
      "learning_rate": 2.123873523728171e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 14807224,
      "step": 22640
    },
    {
      "epoch": 11.868448637316561,
      "grad_norm": 0.3880796730518341,
      "learning_rate": 2.122743149656072e-05,
      "loss": 0.4539,
      "num_input_tokens_seen": 14811096,
      "step": 22645
    },
    {
      "epoch": 11.871069182389936,
      "grad_norm": 0.4274163842201233,
      "learning_rate": 2.1216128545018012e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 14814776,
      "step": 22650
    },
    {
      "epoch": 11.873689727463312,
      "grad_norm": 0.4945424497127533,
      "learning_rate": 2.1204826385018052e-05,
      "loss": 0.5232,
      "num_input_tokens_seen": 14818424,
      "step": 22655
    },
    {
      "epoch": 11.876310272536688,
      "grad_norm": 0.33546265959739685,
      "learning_rate": 2.1193525018925106e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 14822456,
      "step": 22660
    },
    {
      "epoch": 11.878930817610064,
      "grad_norm": 0.6929407119750977,
      "learning_rate": 2.1182224449103304e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 14824760,
      "step": 22665
    },
    {
      "epoch": 11.881551362683439,
      "grad_norm": 0.39971014857292175,
      "learning_rate": 2.117092467791659e-05,
      "loss": 0.3383,
      "num_input_tokens_seen": 14826968,
      "step": 22670
    },
    {
      "epoch": 11.884171907756814,
      "grad_norm": 0.49189552664756775,
      "learning_rate": 2.115962570772876e-05,
      "loss": 0.4914,
      "num_input_tokens_seen": 14830328,
      "step": 22675
    },
    {
      "epoch": 11.88679245283019,
      "grad_norm": 0.49160152673721313,
      "learning_rate": 2.1148327540903408e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 14833944,
      "step": 22680
    },
    {
      "epoch": 11.889412997903564,
      "grad_norm": 0.6312667727470398,
      "learning_rate": 2.113703017980399e-05,
      "loss": 0.492,
      "num_input_tokens_seen": 14840888,
      "step": 22685
    },
    {
      "epoch": 11.89203354297694,
      "grad_norm": 0.4189237952232361,
      "learning_rate": 2.1125733626793793e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 14843832,
      "step": 22690
    },
    {
      "epoch": 11.894654088050315,
      "grad_norm": 0.4335697591304779,
      "learning_rate": 2.1114437884235915e-05,
      "loss": 0.3844,
      "num_input_tokens_seen": 14847352,
      "step": 22695
    },
    {
      "epoch": 11.89727463312369,
      "grad_norm": 0.26981040835380554,
      "learning_rate": 2.1103142954493313e-05,
      "loss": 0.3533,
      "num_input_tokens_seen": 14850712,
      "step": 22700
    },
    {
      "epoch": 11.899895178197065,
      "grad_norm": 1.2011817693710327,
      "learning_rate": 2.1091848839928745e-05,
      "loss": 0.5233,
      "num_input_tokens_seen": 14853432,
      "step": 22705
    },
    {
      "epoch": 11.90251572327044,
      "grad_norm": 0.6172798871994019,
      "learning_rate": 2.1080555542904805e-05,
      "loss": 0.4809,
      "num_input_tokens_seen": 14856280,
      "step": 22710
    },
    {
      "epoch": 11.905136268343815,
      "grad_norm": 0.5802691578865051,
      "learning_rate": 2.106926306578393e-05,
      "loss": 0.5067,
      "num_input_tokens_seen": 14858840,
      "step": 22715
    },
    {
      "epoch": 11.90775681341719,
      "grad_norm": 0.7603439092636108,
      "learning_rate": 2.1057971410928383e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 14861880,
      "step": 22720
    },
    {
      "epoch": 11.910377358490566,
      "grad_norm": 0.49144837260246277,
      "learning_rate": 2.104668058070024e-05,
      "loss": 0.5082,
      "num_input_tokens_seen": 14865016,
      "step": 22725
    },
    {
      "epoch": 11.91299790356394,
      "grad_norm": 0.31017136573791504,
      "learning_rate": 2.1035390577461427e-05,
      "loss": 0.5426,
      "num_input_tokens_seen": 14868760,
      "step": 22730
    },
    {
      "epoch": 11.915618448637316,
      "grad_norm": 0.3201880156993866,
      "learning_rate": 2.102410140357367e-05,
      "loss": 0.3574,
      "num_input_tokens_seen": 14872152,
      "step": 22735
    },
    {
      "epoch": 11.918238993710691,
      "grad_norm": 0.5254794359207153,
      "learning_rate": 2.1012813061398534e-05,
      "loss": 0.5017,
      "num_input_tokens_seen": 14875224,
      "step": 22740
    },
    {
      "epoch": 11.920859538784066,
      "grad_norm": 0.4355519115924835,
      "learning_rate": 2.1001525553297422e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 14878552,
      "step": 22745
    },
    {
      "epoch": 11.923480083857442,
      "grad_norm": 0.5316250920295715,
      "learning_rate": 2.099023888163154e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 14881304,
      "step": 22750
    },
    {
      "epoch": 11.926100628930818,
      "grad_norm": 0.6321207880973816,
      "learning_rate": 2.0978953048761947e-05,
      "loss": 0.5767,
      "num_input_tokens_seen": 14884696,
      "step": 22755
    },
    {
      "epoch": 11.928721174004194,
      "grad_norm": 0.377369225025177,
      "learning_rate": 2.0967668057049507e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 14887640,
      "step": 22760
    },
    {
      "epoch": 11.931341719077569,
      "grad_norm": 0.2926960587501526,
      "learning_rate": 2.0956383908854895e-05,
      "loss": 0.4813,
      "num_input_tokens_seen": 14890840,
      "step": 22765
    },
    {
      "epoch": 11.933962264150944,
      "grad_norm": 0.5826810002326965,
      "learning_rate": 2.0945100606538644e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 14894712,
      "step": 22770
    },
    {
      "epoch": 11.93658280922432,
      "grad_norm": 1.3386695384979248,
      "learning_rate": 2.0933818152461077e-05,
      "loss": 0.5133,
      "num_input_tokens_seen": 14897624,
      "step": 22775
    },
    {
      "epoch": 11.939203354297694,
      "grad_norm": 0.2814685106277466,
      "learning_rate": 2.092253654898237e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 14900632,
      "step": 22780
    },
    {
      "epoch": 11.94182389937107,
      "grad_norm": 0.3142450749874115,
      "learning_rate": 2.0911255798462494e-05,
      "loss": 0.4154,
      "num_input_tokens_seen": 14903992,
      "step": 22785
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 0.42494282126426697,
      "learning_rate": 2.0899975903261266e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 14907032,
      "step": 22790
    },
    {
      "epoch": 11.94706498951782,
      "grad_norm": 0.3094229996204376,
      "learning_rate": 2.0888696865738295e-05,
      "loss": 0.4521,
      "num_input_tokens_seen": 14909656,
      "step": 22795
    },
    {
      "epoch": 11.949685534591195,
      "grad_norm": 0.4146997928619385,
      "learning_rate": 2.087741868825304e-05,
      "loss": 0.2454,
      "num_input_tokens_seen": 14915608,
      "step": 22800
    },
    {
      "epoch": 11.95230607966457,
      "grad_norm": 0.8166505098342896,
      "learning_rate": 2.0866141373164758e-05,
      "loss": 0.6129,
      "num_input_tokens_seen": 14919256,
      "step": 22805
    },
    {
      "epoch": 11.954926624737945,
      "grad_norm": 0.5547004342079163,
      "learning_rate": 2.0854864922832544e-05,
      "loss": 0.6985,
      "num_input_tokens_seen": 14922520,
      "step": 22810
    },
    {
      "epoch": 11.95754716981132,
      "grad_norm": 0.8764386177062988,
      "learning_rate": 2.0843589339615304e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 14925336,
      "step": 22815
    },
    {
      "epoch": 11.960167714884696,
      "grad_norm": 0.323799192905426,
      "learning_rate": 2.0832314625871742e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 14928056,
      "step": 22820
    },
    {
      "epoch": 11.96278825995807,
      "grad_norm": 0.7400440573692322,
      "learning_rate": 2.0821040783960423e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 14931992,
      "step": 22825
    },
    {
      "epoch": 11.965408805031446,
      "grad_norm": 0.5716137290000916,
      "learning_rate": 2.0809767816239686e-05,
      "loss": 0.4991,
      "num_input_tokens_seen": 14935800,
      "step": 22830
    },
    {
      "epoch": 11.968029350104821,
      "grad_norm": 0.6207628846168518,
      "learning_rate": 2.0798495725067725e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 14938968,
      "step": 22835
    },
    {
      "epoch": 11.970649895178196,
      "grad_norm": 0.4699094593524933,
      "learning_rate": 2.078722451280252e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 14941400,
      "step": 22840
    },
    {
      "epoch": 11.973270440251572,
      "grad_norm": 0.46275362372398376,
      "learning_rate": 2.0775954181801893e-05,
      "loss": 0.4517,
      "num_input_tokens_seen": 14945080,
      "step": 22845
    },
    {
      "epoch": 11.975890985324948,
      "grad_norm": 0.9586379528045654,
      "learning_rate": 2.0764684734423455e-05,
      "loss": 0.3906,
      "num_input_tokens_seen": 14947448,
      "step": 22850
    },
    {
      "epoch": 11.978511530398324,
      "grad_norm": 0.3665708601474762,
      "learning_rate": 2.075341617302465e-05,
      "loss": 0.4494,
      "num_input_tokens_seen": 14950712,
      "step": 22855
    },
    {
      "epoch": 11.981132075471699,
      "grad_norm": 0.7014532089233398,
      "learning_rate": 2.074214849996274e-05,
      "loss": 0.4094,
      "num_input_tokens_seen": 14954072,
      "step": 22860
    },
    {
      "epoch": 11.983752620545074,
      "grad_norm": 0.37376752495765686,
      "learning_rate": 2.0730881717594776e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 14957592,
      "step": 22865
    },
    {
      "epoch": 11.98637316561845,
      "grad_norm": 0.42889681458473206,
      "learning_rate": 2.0719615828277657e-05,
      "loss": 0.4319,
      "num_input_tokens_seen": 14961080,
      "step": 22870
    },
    {
      "epoch": 11.988993710691824,
      "grad_norm": 0.8276797533035278,
      "learning_rate": 2.0708350834368083e-05,
      "loss": 0.5336,
      "num_input_tokens_seen": 14964248,
      "step": 22875
    },
    {
      "epoch": 11.9916142557652,
      "grad_norm": 0.41930145025253296,
      "learning_rate": 2.069708673822255e-05,
      "loss": 0.4752,
      "num_input_tokens_seen": 14967800,
      "step": 22880
    },
    {
      "epoch": 11.994234800838575,
      "grad_norm": 0.5744986534118652,
      "learning_rate": 2.068582354219737e-05,
      "loss": 0.3716,
      "num_input_tokens_seen": 14971448,
      "step": 22885
    },
    {
      "epoch": 11.99685534591195,
      "grad_norm": 0.8636305928230286,
      "learning_rate": 2.0674561248648697e-05,
      "loss": 0.5151,
      "num_input_tokens_seen": 14974840,
      "step": 22890
    },
    {
      "epoch": 11.999475890985325,
      "grad_norm": 0.38103294372558594,
      "learning_rate": 2.0663299859932456e-05,
      "loss": 0.4387,
      "num_input_tokens_seen": 14977912,
      "step": 22895
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.4477379322052002,
      "eval_runtime": 9.6922,
      "eval_samples_per_second": 87.493,
      "eval_steps_per_second": 21.873,
      "num_input_tokens_seen": 14977976,
      "step": 22896
    },
    {
      "epoch": 12.0020964360587,
      "grad_norm": 0.6778239607810974,
      "learning_rate": 2.065203937840442e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 14979736,
      "step": 22900
    },
    {
      "epoch": 12.004716981132075,
      "grad_norm": 0.4148780107498169,
      "learning_rate": 2.0640779806420136e-05,
      "loss": 0.4518,
      "num_input_tokens_seen": 14983864,
      "step": 22905
    },
    {
      "epoch": 12.00733752620545,
      "grad_norm": 0.4833752512931824,
      "learning_rate": 2.0629521146334978e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 14988152,
      "step": 22910
    },
    {
      "epoch": 12.009958071278826,
      "grad_norm": 0.4405442774295807,
      "learning_rate": 2.061826340050414e-05,
      "loss": 0.4306,
      "num_input_tokens_seen": 14991096,
      "step": 22915
    },
    {
      "epoch": 12.0125786163522,
      "grad_norm": 0.668635904788971,
      "learning_rate": 2.0607006571282605e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 14993816,
      "step": 22920
    },
    {
      "epoch": 12.015199161425576,
      "grad_norm": 0.8119996786117554,
      "learning_rate": 2.0595750661025183e-05,
      "loss": 0.3528,
      "num_input_tokens_seen": 14996824,
      "step": 22925
    },
    {
      "epoch": 12.017819706498951,
      "grad_norm": 0.5713549852371216,
      "learning_rate": 2.0584495672086483e-05,
      "loss": 0.3481,
      "num_input_tokens_seen": 14999544,
      "step": 22930
    },
    {
      "epoch": 12.020440251572326,
      "grad_norm": 0.5554434657096863,
      "learning_rate": 2.05732416068209e-05,
      "loss": 0.5277,
      "num_input_tokens_seen": 15002840,
      "step": 22935
    },
    {
      "epoch": 12.023060796645701,
      "grad_norm": 0.77901291847229,
      "learning_rate": 2.0561988467582676e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 15005080,
      "step": 22940
    },
    {
      "epoch": 12.025681341719078,
      "grad_norm": 0.3638509213924408,
      "learning_rate": 2.0550736256725828e-05,
      "loss": 0.4839,
      "num_input_tokens_seen": 15008312,
      "step": 22945
    },
    {
      "epoch": 12.028301886792454,
      "grad_norm": 0.4899712800979614,
      "learning_rate": 2.0539484976604196e-05,
      "loss": 0.5245,
      "num_input_tokens_seen": 15011608,
      "step": 22950
    },
    {
      "epoch": 12.030922431865829,
      "grad_norm": 0.49946504831314087,
      "learning_rate": 2.052823462957142e-05,
      "loss": 0.4461,
      "num_input_tokens_seen": 15014488,
      "step": 22955
    },
    {
      "epoch": 12.033542976939204,
      "grad_norm": 0.6526178121566772,
      "learning_rate": 2.0516985217980953e-05,
      "loss": 0.4069,
      "num_input_tokens_seen": 15017912,
      "step": 22960
    },
    {
      "epoch": 12.036163522012579,
      "grad_norm": 0.31920090317726135,
      "learning_rate": 2.050573674418602e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 15020728,
      "step": 22965
    },
    {
      "epoch": 12.038784067085954,
      "grad_norm": 0.2511536180973053,
      "learning_rate": 2.0494489210539693e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 15024408,
      "step": 22970
    },
    {
      "epoch": 12.04140461215933,
      "grad_norm": 0.35165074467658997,
      "learning_rate": 2.0483242619394817e-05,
      "loss": 0.3824,
      "num_input_tokens_seen": 15027928,
      "step": 22975
    },
    {
      "epoch": 12.044025157232705,
      "grad_norm": 0.44531816244125366,
      "learning_rate": 2.0471996973104063e-05,
      "loss": 0.413,
      "num_input_tokens_seen": 15031480,
      "step": 22980
    },
    {
      "epoch": 12.04664570230608,
      "grad_norm": 0.5256910920143127,
      "learning_rate": 2.0460752274019886e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 15034040,
      "step": 22985
    },
    {
      "epoch": 12.049266247379455,
      "grad_norm": 0.3783099353313446,
      "learning_rate": 2.0449508524494537e-05,
      "loss": 0.3174,
      "num_input_tokens_seen": 15037464,
      "step": 22990
    },
    {
      "epoch": 12.05188679245283,
      "grad_norm": 0.5276262164115906,
      "learning_rate": 2.0438265726880092e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 15040728,
      "step": 22995
    },
    {
      "epoch": 12.054507337526205,
      "grad_norm": 0.25252866744995117,
      "learning_rate": 2.042702388352841e-05,
      "loss": 0.3563,
      "num_input_tokens_seen": 15044632,
      "step": 23000
    },
    {
      "epoch": 12.05712788259958,
      "grad_norm": 0.4856376349925995,
      "learning_rate": 2.041578299679117e-05,
      "loss": 0.4218,
      "num_input_tokens_seen": 15047352,
      "step": 23005
    },
    {
      "epoch": 12.059748427672956,
      "grad_norm": 0.466333270072937,
      "learning_rate": 2.0404543069019818e-05,
      "loss": 0.387,
      "num_input_tokens_seen": 15050520,
      "step": 23010
    },
    {
      "epoch": 12.06236897274633,
      "grad_norm": 0.3294273614883423,
      "learning_rate": 2.0393304102565642e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 15053208,
      "step": 23015
    },
    {
      "epoch": 12.064989517819706,
      "grad_norm": 0.5724198222160339,
      "learning_rate": 2.0382066099779685e-05,
      "loss": 0.472,
      "num_input_tokens_seen": 15055736,
      "step": 23020
    },
    {
      "epoch": 12.067610062893081,
      "grad_norm": 0.6366543173789978,
      "learning_rate": 2.0370829063012808e-05,
      "loss": 0.3556,
      "num_input_tokens_seen": 15058648,
      "step": 23025
    },
    {
      "epoch": 12.070230607966456,
      "grad_norm": 0.8003712892532349,
      "learning_rate": 2.035959299461568e-05,
      "loss": 0.4737,
      "num_input_tokens_seen": 15061208,
      "step": 23030
    },
    {
      "epoch": 12.072851153039831,
      "grad_norm": 0.5352208018302917,
      "learning_rate": 2.0348357896938762e-05,
      "loss": 0.3529,
      "num_input_tokens_seen": 15065208,
      "step": 23035
    },
    {
      "epoch": 12.075471698113208,
      "grad_norm": 0.432028591632843,
      "learning_rate": 2.03371237723323e-05,
      "loss": 0.4657,
      "num_input_tokens_seen": 15067928,
      "step": 23040
    },
    {
      "epoch": 12.078092243186584,
      "grad_norm": 0.2717757821083069,
      "learning_rate": 2.032589062314636e-05,
      "loss": 0.4294,
      "num_input_tokens_seen": 15071192,
      "step": 23045
    },
    {
      "epoch": 12.080712788259959,
      "grad_norm": 0.6237751841545105,
      "learning_rate": 2.0314658451730767e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 15074456,
      "step": 23050
    },
    {
      "epoch": 12.083333333333334,
      "grad_norm": 0.55213463306427,
      "learning_rate": 2.030342726043517e-05,
      "loss": 0.5767,
      "num_input_tokens_seen": 15078232,
      "step": 23055
    },
    {
      "epoch": 12.085953878406709,
      "grad_norm": 0.43584877252578735,
      "learning_rate": 2.0292197051609013e-05,
      "loss": 0.4528,
      "num_input_tokens_seen": 15081336,
      "step": 23060
    },
    {
      "epoch": 12.088574423480084,
      "grad_norm": 1.6795251369476318,
      "learning_rate": 2.0280967827601518e-05,
      "loss": 0.5489,
      "num_input_tokens_seen": 15085048,
      "step": 23065
    },
    {
      "epoch": 12.09119496855346,
      "grad_norm": 0.27029263973236084,
      "learning_rate": 2.0269739590761723e-05,
      "loss": 0.4688,
      "num_input_tokens_seen": 15092280,
      "step": 23070
    },
    {
      "epoch": 12.093815513626835,
      "grad_norm": 0.25408825278282166,
      "learning_rate": 2.0258512343438442e-05,
      "loss": 0.4091,
      "num_input_tokens_seen": 15096568,
      "step": 23075
    },
    {
      "epoch": 12.09643605870021,
      "grad_norm": 0.319792240858078,
      "learning_rate": 2.0247286087980276e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 15099224,
      "step": 23080
    },
    {
      "epoch": 12.099056603773585,
      "grad_norm": 0.33205491304397583,
      "learning_rate": 2.0236060826735644e-05,
      "loss": 0.4693,
      "num_input_tokens_seen": 15103096,
      "step": 23085
    },
    {
      "epoch": 12.10167714884696,
      "grad_norm": 0.4292576313018799,
      "learning_rate": 2.0224836562052732e-05,
      "loss": 0.4424,
      "num_input_tokens_seen": 15105656,
      "step": 23090
    },
    {
      "epoch": 12.104297693920335,
      "grad_norm": 0.39549487829208374,
      "learning_rate": 2.0213613296279533e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 15108696,
      "step": 23095
    },
    {
      "epoch": 12.10691823899371,
      "grad_norm": 0.4153401851654053,
      "learning_rate": 2.0202391031763844e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 15111576,
      "step": 23100
    },
    {
      "epoch": 12.109538784067086,
      "grad_norm": 0.8193977475166321,
      "learning_rate": 2.0191169770853195e-05,
      "loss": 0.4114,
      "num_input_tokens_seen": 15115192,
      "step": 23105
    },
    {
      "epoch": 12.11215932914046,
      "grad_norm": 0.3908481299877167,
      "learning_rate": 2.0179949515894975e-05,
      "loss": 0.4938,
      "num_input_tokens_seen": 15119160,
      "step": 23110
    },
    {
      "epoch": 12.114779874213836,
      "grad_norm": 0.8740268349647522,
      "learning_rate": 2.016873026923633e-05,
      "loss": 0.3468,
      "num_input_tokens_seen": 15122136,
      "step": 23115
    },
    {
      "epoch": 12.117400419287211,
      "grad_norm": 0.44033345580101013,
      "learning_rate": 2.0157512033224187e-05,
      "loss": 0.3793,
      "num_input_tokens_seen": 15126328,
      "step": 23120
    },
    {
      "epoch": 12.120020964360586,
      "grad_norm": 0.4239853024482727,
      "learning_rate": 2.0146294810205287e-05,
      "loss": 0.3021,
      "num_input_tokens_seen": 15129304,
      "step": 23125
    },
    {
      "epoch": 12.122641509433961,
      "grad_norm": 0.3632930815219879,
      "learning_rate": 2.0135078602526146e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 15131768,
      "step": 23130
    },
    {
      "epoch": 12.125262054507338,
      "grad_norm": 0.5153638124465942,
      "learning_rate": 2.0123863412533046e-05,
      "loss": 0.4074,
      "num_input_tokens_seen": 15135128,
      "step": 23135
    },
    {
      "epoch": 12.127882599580714,
      "grad_norm": 0.5265026092529297,
      "learning_rate": 2.0112649242572095e-05,
      "loss": 0.4069,
      "num_input_tokens_seen": 15137752,
      "step": 23140
    },
    {
      "epoch": 12.130503144654089,
      "grad_norm": 0.47207605838775635,
      "learning_rate": 2.0101436094989164e-05,
      "loss": 0.3216,
      "num_input_tokens_seen": 15141912,
      "step": 23145
    },
    {
      "epoch": 12.133123689727464,
      "grad_norm": 0.36253029108047485,
      "learning_rate": 2.0090223972129918e-05,
      "loss": 0.3672,
      "num_input_tokens_seen": 15144792,
      "step": 23150
    },
    {
      "epoch": 12.135744234800839,
      "grad_norm": 0.31033554673194885,
      "learning_rate": 2.0079012876339797e-05,
      "loss": 0.4427,
      "num_input_tokens_seen": 15149624,
      "step": 23155
    },
    {
      "epoch": 12.138364779874214,
      "grad_norm": 0.596943736076355,
      "learning_rate": 2.0067802809964055e-05,
      "loss": 0.3685,
      "num_input_tokens_seen": 15152856,
      "step": 23160
    },
    {
      "epoch": 12.14098532494759,
      "grad_norm": 0.8997293710708618,
      "learning_rate": 2.0056593775347686e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 15155256,
      "step": 23165
    },
    {
      "epoch": 12.143605870020965,
      "grad_norm": 0.3487790524959564,
      "learning_rate": 2.00453857748355e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 15157976,
      "step": 23170
    },
    {
      "epoch": 12.14622641509434,
      "grad_norm": 0.43841463327407837,
      "learning_rate": 2.0034178810772082e-05,
      "loss": 0.6484,
      "num_input_tokens_seen": 15160344,
      "step": 23175
    },
    {
      "epoch": 12.148846960167715,
      "grad_norm": 0.4472303092479706,
      "learning_rate": 2.002297288550181e-05,
      "loss": 0.5717,
      "num_input_tokens_seen": 15163032,
      "step": 23180
    },
    {
      "epoch": 12.15146750524109,
      "grad_norm": 0.3212181031703949,
      "learning_rate": 2.0011768001368835e-05,
      "loss": 0.4986,
      "num_input_tokens_seen": 15165784,
      "step": 23185
    },
    {
      "epoch": 12.154088050314465,
      "grad_norm": 0.5450606942176819,
      "learning_rate": 2.0000564160717076e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 15168664,
      "step": 23190
    },
    {
      "epoch": 12.15670859538784,
      "grad_norm": 0.5352720618247986,
      "learning_rate": 1.9989361365890263e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 15171576,
      "step": 23195
    },
    {
      "epoch": 12.159329140461216,
      "grad_norm": 0.2763305902481079,
      "learning_rate": 1.997815961923188e-05,
      "loss": 0.3695,
      "num_input_tokens_seen": 15175160,
      "step": 23200
    },
    {
      "epoch": 12.16194968553459,
      "grad_norm": 0.40318453311920166,
      "learning_rate": 1.996695892308522e-05,
      "loss": 0.518,
      "num_input_tokens_seen": 15178744,
      "step": 23205
    },
    {
      "epoch": 12.164570230607966,
      "grad_norm": 0.4986784756183624,
      "learning_rate": 1.9955759279793327e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 15182264,
      "step": 23210
    },
    {
      "epoch": 12.167190775681341,
      "grad_norm": 0.6820975542068481,
      "learning_rate": 1.9944560691699057e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 15184984,
      "step": 23215
    },
    {
      "epoch": 12.169811320754716,
      "grad_norm": 0.8392009735107422,
      "learning_rate": 1.9933363161145006e-05,
      "loss": 0.4894,
      "num_input_tokens_seen": 15187448,
      "step": 23220
    },
    {
      "epoch": 12.172431865828091,
      "grad_norm": 0.4292386472225189,
      "learning_rate": 1.9922166690473575e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 15190872,
      "step": 23225
    },
    {
      "epoch": 12.175052410901468,
      "grad_norm": 0.5317952632904053,
      "learning_rate": 1.991097128202695e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 15193400,
      "step": 23230
    },
    {
      "epoch": 12.177672955974844,
      "grad_norm": 0.7746534943580627,
      "learning_rate": 1.9899776938147065e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 15197336,
      "step": 23235
    },
    {
      "epoch": 12.180293501048219,
      "grad_norm": 0.7267304062843323,
      "learning_rate": 1.9888583661175665e-05,
      "loss": 0.5393,
      "num_input_tokens_seen": 15200504,
      "step": 23240
    },
    {
      "epoch": 12.182914046121594,
      "grad_norm": 0.4323164224624634,
      "learning_rate": 1.9877391453454253e-05,
      "loss": 0.5173,
      "num_input_tokens_seen": 15203640,
      "step": 23245
    },
    {
      "epoch": 12.185534591194969,
      "grad_norm": 0.4001292288303375,
      "learning_rate": 1.98662003173241e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 15206744,
      "step": 23250
    },
    {
      "epoch": 12.188155136268344,
      "grad_norm": 0.4307399392127991,
      "learning_rate": 1.9855010255126266e-05,
      "loss": 0.352,
      "num_input_tokens_seen": 15211064,
      "step": 23255
    },
    {
      "epoch": 12.19077568134172,
      "grad_norm": 0.35579928755760193,
      "learning_rate": 1.9843821269201598e-05,
      "loss": 0.516,
      "num_input_tokens_seen": 15214040,
      "step": 23260
    },
    {
      "epoch": 12.193396226415095,
      "grad_norm": 0.34073954820632935,
      "learning_rate": 1.983263336189069e-05,
      "loss": 0.5963,
      "num_input_tokens_seen": 15216440,
      "step": 23265
    },
    {
      "epoch": 12.19601677148847,
      "grad_norm": 0.612821638584137,
      "learning_rate": 1.9821446535533937e-05,
      "loss": 0.4507,
      "num_input_tokens_seen": 15219352,
      "step": 23270
    },
    {
      "epoch": 12.198637316561845,
      "grad_norm": 0.5653204917907715,
      "learning_rate": 1.9810260792471492e-05,
      "loss": 0.4602,
      "num_input_tokens_seen": 15225432,
      "step": 23275
    },
    {
      "epoch": 12.20125786163522,
      "grad_norm": 0.2402520328760147,
      "learning_rate": 1.979907613504327e-05,
      "loss": 0.487,
      "num_input_tokens_seen": 15229528,
      "step": 23280
    },
    {
      "epoch": 12.203878406708595,
      "grad_norm": 0.6158778071403503,
      "learning_rate": 1.978789256558899e-05,
      "loss": 0.4635,
      "num_input_tokens_seen": 15233208,
      "step": 23285
    },
    {
      "epoch": 12.20649895178197,
      "grad_norm": 0.40884822607040405,
      "learning_rate": 1.977671008644812e-05,
      "loss": 0.3619,
      "num_input_tokens_seen": 15236504,
      "step": 23290
    },
    {
      "epoch": 12.209119496855346,
      "grad_norm": 0.47746819257736206,
      "learning_rate": 1.9765528699959912e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 15239768,
      "step": 23295
    },
    {
      "epoch": 12.21174004192872,
      "grad_norm": 0.6902632117271423,
      "learning_rate": 1.975434840846339e-05,
      "loss": 0.5498,
      "num_input_tokens_seen": 15243032,
      "step": 23300
    },
    {
      "epoch": 12.214360587002096,
      "grad_norm": 0.5971096754074097,
      "learning_rate": 1.9743169214297326e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 15245944,
      "step": 23305
    },
    {
      "epoch": 12.216981132075471,
      "grad_norm": 0.2795611023902893,
      "learning_rate": 1.973199111980029e-05,
      "loss": 0.3829,
      "num_input_tokens_seen": 15249880,
      "step": 23310
    },
    {
      "epoch": 12.219601677148846,
      "grad_norm": 0.6339534521102905,
      "learning_rate": 1.9720814127310608e-05,
      "loss": 0.3896,
      "num_input_tokens_seen": 15252984,
      "step": 23315
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 0.5054576396942139,
      "learning_rate": 1.9709638239166386e-05,
      "loss": 0.4377,
      "num_input_tokens_seen": 15256024,
      "step": 23320
    },
    {
      "epoch": 12.224842767295598,
      "grad_norm": 0.4241565465927124,
      "learning_rate": 1.9698463457705485e-05,
      "loss": 0.3301,
      "num_input_tokens_seen": 15259608,
      "step": 23325
    },
    {
      "epoch": 12.227463312368974,
      "grad_norm": 1.2187694311141968,
      "learning_rate": 1.968728978526556e-05,
      "loss": 0.3297,
      "num_input_tokens_seen": 15262200,
      "step": 23330
    },
    {
      "epoch": 12.230083857442349,
      "grad_norm": 0.4586414098739624,
      "learning_rate": 1.9676117224183986e-05,
      "loss": 0.353,
      "num_input_tokens_seen": 15265336,
      "step": 23335
    },
    {
      "epoch": 12.232704402515724,
      "grad_norm": 0.5088330507278442,
      "learning_rate": 1.9664945776797956e-05,
      "loss": 0.5614,
      "num_input_tokens_seen": 15268024,
      "step": 23340
    },
    {
      "epoch": 12.235324947589099,
      "grad_norm": 0.5173563361167908,
      "learning_rate": 1.96537754454444e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 15271480,
      "step": 23345
    },
    {
      "epoch": 12.237945492662474,
      "grad_norm": 0.3880672752857208,
      "learning_rate": 1.964260623246003e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 15274296,
      "step": 23350
    },
    {
      "epoch": 12.24056603773585,
      "grad_norm": 0.45236077904701233,
      "learning_rate": 1.9631438140181324e-05,
      "loss": 0.3974,
      "num_input_tokens_seen": 15276952,
      "step": 23355
    },
    {
      "epoch": 12.243186582809225,
      "grad_norm": 0.45240044593811035,
      "learning_rate": 1.9620271170944503e-05,
      "loss": 0.3768,
      "num_input_tokens_seen": 15280056,
      "step": 23360
    },
    {
      "epoch": 12.2458071278826,
      "grad_norm": 0.573871910572052,
      "learning_rate": 1.960910532708558e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 15283256,
      "step": 23365
    },
    {
      "epoch": 12.248427672955975,
      "grad_norm": 0.6832507848739624,
      "learning_rate": 1.9597940610940315e-05,
      "loss": 0.4018,
      "num_input_tokens_seen": 15286808,
      "step": 23370
    },
    {
      "epoch": 12.25104821802935,
      "grad_norm": 0.44480493664741516,
      "learning_rate": 1.958677702484425e-05,
      "loss": 0.4887,
      "num_input_tokens_seen": 15291000,
      "step": 23375
    },
    {
      "epoch": 12.253668763102725,
      "grad_norm": 0.3270132541656494,
      "learning_rate": 1.957561457113267e-05,
      "loss": 0.4939,
      "num_input_tokens_seen": 15294744,
      "step": 23380
    },
    {
      "epoch": 12.2562893081761,
      "grad_norm": 0.7501173615455627,
      "learning_rate": 1.9564453252140648e-05,
      "loss": 0.4675,
      "num_input_tokens_seen": 15297752,
      "step": 23385
    },
    {
      "epoch": 12.258909853249476,
      "grad_norm": 0.3362756371498108,
      "learning_rate": 1.9553293070202984e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 15301208,
      "step": 23390
    },
    {
      "epoch": 12.26153039832285,
      "grad_norm": 0.32430458068847656,
      "learning_rate": 1.9542134027654267e-05,
      "loss": 0.4914,
      "num_input_tokens_seen": 15304984,
      "step": 23395
    },
    {
      "epoch": 12.264150943396226,
      "grad_norm": 0.47620439529418945,
      "learning_rate": 1.953097612682885e-05,
      "loss": 0.3927,
      "num_input_tokens_seen": 15307896,
      "step": 23400
    },
    {
      "epoch": 12.266771488469601,
      "grad_norm": 0.25919947028160095,
      "learning_rate": 1.9519819370060832e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 15311064,
      "step": 23405
    },
    {
      "epoch": 12.269392033542976,
      "grad_norm": 0.3687993586063385,
      "learning_rate": 1.9508663759684074e-05,
      "loss": 0.4054,
      "num_input_tokens_seen": 15314136,
      "step": 23410
    },
    {
      "epoch": 12.272012578616351,
      "grad_norm": 0.37905558943748474,
      "learning_rate": 1.9497509298032224e-05,
      "loss": 0.6302,
      "num_input_tokens_seen": 15317464,
      "step": 23415
    },
    {
      "epoch": 12.274633123689728,
      "grad_norm": 0.4430273771286011,
      "learning_rate": 1.9486355987438647e-05,
      "loss": 0.4298,
      "num_input_tokens_seen": 15320920,
      "step": 23420
    },
    {
      "epoch": 12.277253668763104,
      "grad_norm": 1.010634183883667,
      "learning_rate": 1.947520383023649e-05,
      "loss": 0.3135,
      "num_input_tokens_seen": 15323800,
      "step": 23425
    },
    {
      "epoch": 12.279874213836479,
      "grad_norm": 0.33926013112068176,
      "learning_rate": 1.9464052828758666e-05,
      "loss": 0.5327,
      "num_input_tokens_seen": 15327512,
      "step": 23430
    },
    {
      "epoch": 12.282494758909854,
      "grad_norm": 0.4183402955532074,
      "learning_rate": 1.945290298533783e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 15331064,
      "step": 23435
    },
    {
      "epoch": 12.285115303983229,
      "grad_norm": 0.42083740234375,
      "learning_rate": 1.9441754302306406e-05,
      "loss": 0.4718,
      "num_input_tokens_seen": 15334424,
      "step": 23440
    },
    {
      "epoch": 12.287735849056604,
      "grad_norm": 0.5597988367080688,
      "learning_rate": 1.9430606781996578e-05,
      "loss": 0.4247,
      "num_input_tokens_seen": 15337720,
      "step": 23445
    },
    {
      "epoch": 12.29035639412998,
      "grad_norm": 0.5735397338867188,
      "learning_rate": 1.941946042674026e-05,
      "loss": 0.5255,
      "num_input_tokens_seen": 15340216,
      "step": 23450
    },
    {
      "epoch": 12.292976939203355,
      "grad_norm": 0.4843824505805969,
      "learning_rate": 1.940831523886916e-05,
      "loss": 0.3898,
      "num_input_tokens_seen": 15343736,
      "step": 23455
    },
    {
      "epoch": 12.29559748427673,
      "grad_norm": 0.42437487840652466,
      "learning_rate": 1.9397171220714714e-05,
      "loss": 0.5693,
      "num_input_tokens_seen": 15347000,
      "step": 23460
    },
    {
      "epoch": 12.298218029350105,
      "grad_norm": 0.47786813974380493,
      "learning_rate": 1.9386028374608134e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 15349848,
      "step": 23465
    },
    {
      "epoch": 12.30083857442348,
      "grad_norm": 0.2990286648273468,
      "learning_rate": 1.9374886702880372e-05,
      "loss": 0.4869,
      "num_input_tokens_seen": 15353272,
      "step": 23470
    },
    {
      "epoch": 12.303459119496855,
      "grad_norm": 0.5045241117477417,
      "learning_rate": 1.936374620786213e-05,
      "loss": 0.3091,
      "num_input_tokens_seen": 15355960,
      "step": 23475
    },
    {
      "epoch": 12.30607966457023,
      "grad_norm": 0.5545404553413391,
      "learning_rate": 1.9352606891883885e-05,
      "loss": 0.5274,
      "num_input_tokens_seen": 15359800,
      "step": 23480
    },
    {
      "epoch": 12.308700209643606,
      "grad_norm": 0.46521905064582825,
      "learning_rate": 1.9341468757275844e-05,
      "loss": 0.307,
      "num_input_tokens_seen": 15363640,
      "step": 23485
    },
    {
      "epoch": 12.31132075471698,
      "grad_norm": 0.19623707234859467,
      "learning_rate": 1.933033180636798e-05,
      "loss": 0.3919,
      "num_input_tokens_seen": 15367768,
      "step": 23490
    },
    {
      "epoch": 12.313941299790356,
      "grad_norm": 0.4252423048019409,
      "learning_rate": 1.9319196041490027e-05,
      "loss": 0.3381,
      "num_input_tokens_seen": 15370680,
      "step": 23495
    },
    {
      "epoch": 12.316561844863731,
      "grad_norm": 0.44250887632369995,
      "learning_rate": 1.930806146497146e-05,
      "loss": 0.4007,
      "num_input_tokens_seen": 15374008,
      "step": 23500
    },
    {
      "epoch": 12.319182389937106,
      "grad_norm": 0.4611261487007141,
      "learning_rate": 1.9296928079141485e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 15377976,
      "step": 23505
    },
    {
      "epoch": 12.321802935010481,
      "grad_norm": 0.35687628388404846,
      "learning_rate": 1.9285795886329093e-05,
      "loss": 0.5548,
      "num_input_tokens_seen": 15381144,
      "step": 23510
    },
    {
      "epoch": 12.324423480083858,
      "grad_norm": 0.49778300523757935,
      "learning_rate": 1.927466488886301e-05,
      "loss": 0.4985,
      "num_input_tokens_seen": 15385144,
      "step": 23515
    },
    {
      "epoch": 12.327044025157234,
      "grad_norm": 0.271230548620224,
      "learning_rate": 1.926353508907172e-05,
      "loss": 0.4623,
      "num_input_tokens_seen": 15388504,
      "step": 23520
    },
    {
      "epoch": 12.329664570230609,
      "grad_norm": 0.7499226331710815,
      "learning_rate": 1.9252406489283437e-05,
      "loss": 0.5221,
      "num_input_tokens_seen": 15391448,
      "step": 23525
    },
    {
      "epoch": 12.332285115303984,
      "grad_norm": 0.5872433185577393,
      "learning_rate": 1.9241279091826155e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 15393816,
      "step": 23530
    },
    {
      "epoch": 12.334905660377359,
      "grad_norm": 0.43811753392219543,
      "learning_rate": 1.923015289902758e-05,
      "loss": 0.3435,
      "num_input_tokens_seen": 15396888,
      "step": 23535
    },
    {
      "epoch": 12.337526205450734,
      "grad_norm": 0.43912413716316223,
      "learning_rate": 1.9219027913215193e-05,
      "loss": 0.3513,
      "num_input_tokens_seen": 15399352,
      "step": 23540
    },
    {
      "epoch": 12.34014675052411,
      "grad_norm": 0.2487337440252304,
      "learning_rate": 1.9207904136716214e-05,
      "loss": 0.5669,
      "num_input_tokens_seen": 15404152,
      "step": 23545
    },
    {
      "epoch": 12.342767295597485,
      "grad_norm": 0.5253121852874756,
      "learning_rate": 1.9196781571857608e-05,
      "loss": 0.3887,
      "num_input_tokens_seen": 15407000,
      "step": 23550
    },
    {
      "epoch": 12.34538784067086,
      "grad_norm": 0.4194706678390503,
      "learning_rate": 1.9185660220966097e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 15410360,
      "step": 23555
    },
    {
      "epoch": 12.348008385744235,
      "grad_norm": 0.37089672684669495,
      "learning_rate": 1.9174540086368135e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 15413560,
      "step": 23560
    },
    {
      "epoch": 12.35062893081761,
      "grad_norm": 0.24448587000370026,
      "learning_rate": 1.9163421170389918e-05,
      "loss": 0.4305,
      "num_input_tokens_seen": 15417400,
      "step": 23565
    },
    {
      "epoch": 12.353249475890985,
      "grad_norm": 0.5986145734786987,
      "learning_rate": 1.9152303475357402e-05,
      "loss": 0.4786,
      "num_input_tokens_seen": 15420248,
      "step": 23570
    },
    {
      "epoch": 12.35587002096436,
      "grad_norm": 0.6640640497207642,
      "learning_rate": 1.9141187003596295e-05,
      "loss": 0.5187,
      "num_input_tokens_seen": 15423320,
      "step": 23575
    },
    {
      "epoch": 12.358490566037736,
      "grad_norm": 0.43955662846565247,
      "learning_rate": 1.913007175743202e-05,
      "loss": 0.4339,
      "num_input_tokens_seen": 15426840,
      "step": 23580
    },
    {
      "epoch": 12.36111111111111,
      "grad_norm": 0.36930051445961,
      "learning_rate": 1.9118957739189773e-05,
      "loss": 0.4107,
      "num_input_tokens_seen": 15430488,
      "step": 23585
    },
    {
      "epoch": 12.363731656184486,
      "grad_norm": 0.5185964703559875,
      "learning_rate": 1.910784495119447e-05,
      "loss": 0.5554,
      "num_input_tokens_seen": 15433432,
      "step": 23590
    },
    {
      "epoch": 12.366352201257861,
      "grad_norm": 0.42884349822998047,
      "learning_rate": 1.9096733395770778e-05,
      "loss": 0.3549,
      "num_input_tokens_seen": 15437240,
      "step": 23595
    },
    {
      "epoch": 12.368972746331236,
      "grad_norm": 0.5915472507476807,
      "learning_rate": 1.9085623075243114e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 15440792,
      "step": 23600
    },
    {
      "epoch": 12.371593291404611,
      "grad_norm": 0.501839816570282,
      "learning_rate": 1.9074513991935623e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 15443928,
      "step": 23605
    },
    {
      "epoch": 12.374213836477987,
      "grad_norm": 0.3456398844718933,
      "learning_rate": 1.906340614817221e-05,
      "loss": 0.5116,
      "num_input_tokens_seen": 15446936,
      "step": 23610
    },
    {
      "epoch": 12.376834381551364,
      "grad_norm": 0.6111155152320862,
      "learning_rate": 1.9052299546276506e-05,
      "loss": 0.4263,
      "num_input_tokens_seen": 15450168,
      "step": 23615
    },
    {
      "epoch": 12.379454926624739,
      "grad_norm": 0.3955148458480835,
      "learning_rate": 1.904119418857187e-05,
      "loss": 0.4458,
      "num_input_tokens_seen": 15453592,
      "step": 23620
    },
    {
      "epoch": 12.382075471698114,
      "grad_norm": 0.444327175617218,
      "learning_rate": 1.903009007738144e-05,
      "loss": 0.517,
      "num_input_tokens_seen": 15460056,
      "step": 23625
    },
    {
      "epoch": 12.384696016771489,
      "grad_norm": 0.2961818277835846,
      "learning_rate": 1.9018987215028045e-05,
      "loss": 0.4166,
      "num_input_tokens_seen": 15463160,
      "step": 23630
    },
    {
      "epoch": 12.387316561844864,
      "grad_norm": 0.5477449893951416,
      "learning_rate": 1.9007885603834298e-05,
      "loss": 0.6901,
      "num_input_tokens_seen": 15466360,
      "step": 23635
    },
    {
      "epoch": 12.38993710691824,
      "grad_norm": 0.4427703022956848,
      "learning_rate": 1.8996785246122526e-05,
      "loss": 0.5985,
      "num_input_tokens_seen": 15469112,
      "step": 23640
    },
    {
      "epoch": 12.392557651991615,
      "grad_norm": 0.5735595226287842,
      "learning_rate": 1.898568614421478e-05,
      "loss": 0.4184,
      "num_input_tokens_seen": 15472088,
      "step": 23645
    },
    {
      "epoch": 12.39517819706499,
      "grad_norm": 0.3773793578147888,
      "learning_rate": 1.897458830043288e-05,
      "loss": 0.3433,
      "num_input_tokens_seen": 15475384,
      "step": 23650
    },
    {
      "epoch": 12.397798742138365,
      "grad_norm": 0.33264803886413574,
      "learning_rate": 1.8963491717098372e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 15477528,
      "step": 23655
    },
    {
      "epoch": 12.40041928721174,
      "grad_norm": 0.8594293594360352,
      "learning_rate": 1.8952396396532522e-05,
      "loss": 0.4182,
      "num_input_tokens_seen": 15479768,
      "step": 23660
    },
    {
      "epoch": 12.403039832285115,
      "grad_norm": 0.3852856755256653,
      "learning_rate": 1.8941302341056356e-05,
      "loss": 0.8211,
      "num_input_tokens_seen": 15482264,
      "step": 23665
    },
    {
      "epoch": 12.40566037735849,
      "grad_norm": 0.3162825107574463,
      "learning_rate": 1.8930209552990625e-05,
      "loss": 0.4825,
      "num_input_tokens_seen": 15485144,
      "step": 23670
    },
    {
      "epoch": 12.408280922431866,
      "grad_norm": 0.5763822793960571,
      "learning_rate": 1.8919118034655796e-05,
      "loss": 0.3349,
      "num_input_tokens_seen": 15487480,
      "step": 23675
    },
    {
      "epoch": 12.41090146750524,
      "grad_norm": 0.3289066553115845,
      "learning_rate": 1.8908027788372107e-05,
      "loss": 0.4739,
      "num_input_tokens_seen": 15491576,
      "step": 23680
    },
    {
      "epoch": 12.413522012578616,
      "grad_norm": 0.5699968934059143,
      "learning_rate": 1.88969388164595e-05,
      "loss": 0.5285,
      "num_input_tokens_seen": 15494328,
      "step": 23685
    },
    {
      "epoch": 12.416142557651991,
      "grad_norm": 0.4204627573490143,
      "learning_rate": 1.8885851121237666e-05,
      "loss": 0.5764,
      "num_input_tokens_seen": 15498360,
      "step": 23690
    },
    {
      "epoch": 12.418763102725366,
      "grad_norm": 0.546731173992157,
      "learning_rate": 1.8874764705026023e-05,
      "loss": 0.3787,
      "num_input_tokens_seen": 15501176,
      "step": 23695
    },
    {
      "epoch": 12.421383647798741,
      "grad_norm": 0.537167489528656,
      "learning_rate": 1.8863679570143728e-05,
      "loss": 0.3963,
      "num_input_tokens_seen": 15504440,
      "step": 23700
    },
    {
      "epoch": 12.424004192872117,
      "grad_norm": 0.5191128849983215,
      "learning_rate": 1.885259571890966e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 15507512,
      "step": 23705
    },
    {
      "epoch": 12.426624737945493,
      "grad_norm": 0.6981663107872009,
      "learning_rate": 1.8841513153642427e-05,
      "loss": 0.501,
      "num_input_tokens_seen": 15510840,
      "step": 23710
    },
    {
      "epoch": 12.429245283018869,
      "grad_norm": 0.7521884441375732,
      "learning_rate": 1.8830431876660388e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 15513656,
      "step": 23715
    },
    {
      "epoch": 12.431865828092244,
      "grad_norm": 0.4553252160549164,
      "learning_rate": 1.8819351890281607e-05,
      "loss": 0.3757,
      "num_input_tokens_seen": 15516952,
      "step": 23720
    },
    {
      "epoch": 12.434486373165619,
      "grad_norm": 0.7024857997894287,
      "learning_rate": 1.880827319682391e-05,
      "loss": 0.4227,
      "num_input_tokens_seen": 15519832,
      "step": 23725
    },
    {
      "epoch": 12.437106918238994,
      "grad_norm": 0.3847634196281433,
      "learning_rate": 1.879719579860481e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 15523480,
      "step": 23730
    },
    {
      "epoch": 12.43972746331237,
      "grad_norm": 0.8537208437919617,
      "learning_rate": 1.878611969794159e-05,
      "loss": 0.4402,
      "num_input_tokens_seen": 15526264,
      "step": 23735
    },
    {
      "epoch": 12.442348008385745,
      "grad_norm": 0.50062495470047,
      "learning_rate": 1.8775044897151232e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 15529848,
      "step": 23740
    },
    {
      "epoch": 12.44496855345912,
      "grad_norm": 0.40491729974746704,
      "learning_rate": 1.876397139855047e-05,
      "loss": 0.3109,
      "num_input_tokens_seen": 15533368,
      "step": 23745
    },
    {
      "epoch": 12.447589098532495,
      "grad_norm": 0.2938928008079529,
      "learning_rate": 1.875289920445574e-05,
      "loss": 0.3853,
      "num_input_tokens_seen": 15537080,
      "step": 23750
    },
    {
      "epoch": 12.45020964360587,
      "grad_norm": 0.5191361904144287,
      "learning_rate": 1.8741828317183234e-05,
      "loss": 0.3578,
      "num_input_tokens_seen": 15540248,
      "step": 23755
    },
    {
      "epoch": 12.452830188679245,
      "grad_norm": 0.4055995047092438,
      "learning_rate": 1.873075873904885e-05,
      "loss": 0.4697,
      "num_input_tokens_seen": 15543064,
      "step": 23760
    },
    {
      "epoch": 12.45545073375262,
      "grad_norm": 0.354387104511261,
      "learning_rate": 1.87196904723682e-05,
      "loss": 0.3294,
      "num_input_tokens_seen": 15546968,
      "step": 23765
    },
    {
      "epoch": 12.458071278825996,
      "grad_norm": 0.4775179624557495,
      "learning_rate": 1.8708623519456665e-05,
      "loss": 0.7048,
      "num_input_tokens_seen": 15550040,
      "step": 23770
    },
    {
      "epoch": 12.46069182389937,
      "grad_norm": 0.6436004638671875,
      "learning_rate": 1.8697557882629308e-05,
      "loss": 0.3831,
      "num_input_tokens_seen": 15552824,
      "step": 23775
    },
    {
      "epoch": 12.463312368972746,
      "grad_norm": 0.5842520594596863,
      "learning_rate": 1.868649356420094e-05,
      "loss": 0.4067,
      "num_input_tokens_seen": 15555448,
      "step": 23780
    },
    {
      "epoch": 12.465932914046121,
      "grad_norm": 0.45339733362197876,
      "learning_rate": 1.86754305664861e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 15559096,
      "step": 23785
    },
    {
      "epoch": 12.468553459119496,
      "grad_norm": 0.7564254403114319,
      "learning_rate": 1.866436889179902e-05,
      "loss": 0.6339,
      "num_input_tokens_seen": 15562136,
      "step": 23790
    },
    {
      "epoch": 12.471174004192871,
      "grad_norm": 0.49825039505958557,
      "learning_rate": 1.865330854245369e-05,
      "loss": 0.3588,
      "num_input_tokens_seen": 15565528,
      "step": 23795
    },
    {
      "epoch": 12.473794549266247,
      "grad_norm": 0.5486805438995361,
      "learning_rate": 1.8642249520763804e-05,
      "loss": 0.5917,
      "num_input_tokens_seen": 15569048,
      "step": 23800
    },
    {
      "epoch": 12.476415094339623,
      "grad_norm": 0.32410943508148193,
      "learning_rate": 1.863119182904278e-05,
      "loss": 0.4402,
      "num_input_tokens_seen": 15572088,
      "step": 23805
    },
    {
      "epoch": 12.479035639412999,
      "grad_norm": 0.5372464656829834,
      "learning_rate": 1.8620135469603777e-05,
      "loss": 0.4155,
      "num_input_tokens_seen": 15574488,
      "step": 23810
    },
    {
      "epoch": 12.481656184486374,
      "grad_norm": 0.43506625294685364,
      "learning_rate": 1.8609080444759653e-05,
      "loss": 0.3608,
      "num_input_tokens_seen": 15577176,
      "step": 23815
    },
    {
      "epoch": 12.484276729559749,
      "grad_norm": 0.6328781247138977,
      "learning_rate": 1.8598026756822978e-05,
      "loss": 0.4172,
      "num_input_tokens_seen": 15582168,
      "step": 23820
    },
    {
      "epoch": 12.486897274633124,
      "grad_norm": 0.499764084815979,
      "learning_rate": 1.8586974408106074e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 15585880,
      "step": 23825
    },
    {
      "epoch": 12.4895178197065,
      "grad_norm": 0.49965423345565796,
      "learning_rate": 1.8575923400920954e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 15588920,
      "step": 23830
    },
    {
      "epoch": 12.492138364779874,
      "grad_norm": 0.279822438955307,
      "learning_rate": 1.856487373757937e-05,
      "loss": 0.4403,
      "num_input_tokens_seen": 15592312,
      "step": 23835
    },
    {
      "epoch": 12.49475890985325,
      "grad_norm": 0.4247068166732788,
      "learning_rate": 1.8553825420392792e-05,
      "loss": 0.5115,
      "num_input_tokens_seen": 15595512,
      "step": 23840
    },
    {
      "epoch": 12.497379454926625,
      "grad_norm": 0.7119770050048828,
      "learning_rate": 1.8542778451672385e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 15598040,
      "step": 23845
    },
    {
      "epoch": 12.5,
      "grad_norm": 0.5053831338882446,
      "learning_rate": 1.8531732833729063e-05,
      "loss": 0.3301,
      "num_input_tokens_seen": 15600344,
      "step": 23850
    },
    {
      "epoch": 12.502620545073375,
      "grad_norm": 0.5299876928329468,
      "learning_rate": 1.852068856887343e-05,
      "loss": 0.543,
      "num_input_tokens_seen": 15603960,
      "step": 23855
    },
    {
      "epoch": 12.50524109014675,
      "grad_norm": 0.5258225202560425,
      "learning_rate": 1.8509645659415837e-05,
      "loss": 0.3783,
      "num_input_tokens_seen": 15606328,
      "step": 23860
    },
    {
      "epoch": 12.507861635220126,
      "grad_norm": 0.4846625328063965,
      "learning_rate": 1.849860410766632e-05,
      "loss": 0.3886,
      "num_input_tokens_seen": 15608760,
      "step": 23865
    },
    {
      "epoch": 12.5104821802935,
      "grad_norm": 0.6147931218147278,
      "learning_rate": 1.8487563915934664e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 15612088,
      "step": 23870
    },
    {
      "epoch": 12.513102725366876,
      "grad_norm": 0.9175284504890442,
      "learning_rate": 1.847652508653033e-05,
      "loss": 0.5285,
      "num_input_tokens_seen": 15615640,
      "step": 23875
    },
    {
      "epoch": 12.515723270440251,
      "grad_norm": 0.30393165349960327,
      "learning_rate": 1.846548762176252e-05,
      "loss": 0.3454,
      "num_input_tokens_seen": 15619000,
      "step": 23880
    },
    {
      "epoch": 12.518343815513626,
      "grad_norm": 0.9392755031585693,
      "learning_rate": 1.845445152394015e-05,
      "loss": 0.519,
      "num_input_tokens_seen": 15621720,
      "step": 23885
    },
    {
      "epoch": 12.520964360587001,
      "grad_norm": 0.3618812561035156,
      "learning_rate": 1.8443416795371853e-05,
      "loss": 0.3393,
      "num_input_tokens_seen": 15624920,
      "step": 23890
    },
    {
      "epoch": 12.523584905660378,
      "grad_norm": 0.37801554799079895,
      "learning_rate": 1.8432383438365957e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 15628056,
      "step": 23895
    },
    {
      "epoch": 12.526205450733752,
      "grad_norm": 0.47719261050224304,
      "learning_rate": 1.8421351455230528e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 15630936,
      "step": 23900
    },
    {
      "epoch": 12.528825995807129,
      "grad_norm": 0.6759941577911377,
      "learning_rate": 1.8410320848273315e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 15633816,
      "step": 23905
    },
    {
      "epoch": 12.531446540880504,
      "grad_norm": 0.5305081009864807,
      "learning_rate": 1.8399291619801794e-05,
      "loss": 0.2858,
      "num_input_tokens_seen": 15636312,
      "step": 23910
    },
    {
      "epoch": 12.534067085953879,
      "grad_norm": 0.5746659636497498,
      "learning_rate": 1.838826377212317e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 15638616,
      "step": 23915
    },
    {
      "epoch": 12.536687631027254,
      "grad_norm": 0.41877639293670654,
      "learning_rate": 1.8377237307544326e-05,
      "loss": 0.488,
      "num_input_tokens_seen": 15642488,
      "step": 23920
    },
    {
      "epoch": 12.53930817610063,
      "grad_norm": 0.4998376965522766,
      "learning_rate": 1.8366212228371894e-05,
      "loss": 0.3827,
      "num_input_tokens_seen": 15645304,
      "step": 23925
    },
    {
      "epoch": 12.541928721174004,
      "grad_norm": 0.2523074150085449,
      "learning_rate": 1.8355188536912173e-05,
      "loss": 0.5204,
      "num_input_tokens_seen": 15649208,
      "step": 23930
    },
    {
      "epoch": 12.54454926624738,
      "grad_norm": 1.1067720651626587,
      "learning_rate": 1.83441662354712e-05,
      "loss": 0.4679,
      "num_input_tokens_seen": 15652120,
      "step": 23935
    },
    {
      "epoch": 12.547169811320755,
      "grad_norm": 1.1746948957443237,
      "learning_rate": 1.8333145326354722e-05,
      "loss": 0.6456,
      "num_input_tokens_seen": 15654968,
      "step": 23940
    },
    {
      "epoch": 12.54979035639413,
      "grad_norm": 0.5848717093467712,
      "learning_rate": 1.8322125811868175e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 15658456,
      "step": 23945
    },
    {
      "epoch": 12.552410901467505,
      "grad_norm": 0.8466747999191284,
      "learning_rate": 1.831110769431673e-05,
      "loss": 0.5962,
      "num_input_tokens_seen": 15661816,
      "step": 23950
    },
    {
      "epoch": 12.55503144654088,
      "grad_norm": 0.4679403603076935,
      "learning_rate": 1.8300090976005247e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 15664664,
      "step": 23955
    },
    {
      "epoch": 12.557651991614255,
      "grad_norm": 0.7389859557151794,
      "learning_rate": 1.8289075659238285e-05,
      "loss": 0.4809,
      "num_input_tokens_seen": 15667448,
      "step": 23960
    },
    {
      "epoch": 12.56027253668763,
      "grad_norm": 0.463100790977478,
      "learning_rate": 1.8278061746320134e-05,
      "loss": 0.4354,
      "num_input_tokens_seen": 15671512,
      "step": 23965
    },
    {
      "epoch": 12.562893081761006,
      "grad_norm": 0.3380843698978424,
      "learning_rate": 1.8267049239554786e-05,
      "loss": 0.3602,
      "num_input_tokens_seen": 15675832,
      "step": 23970
    },
    {
      "epoch": 12.565513626834381,
      "grad_norm": 0.4968021512031555,
      "learning_rate": 1.8256038141245912e-05,
      "loss": 0.3728,
      "num_input_tokens_seen": 15679096,
      "step": 23975
    },
    {
      "epoch": 12.568134171907756,
      "grad_norm": 0.5040256381034851,
      "learning_rate": 1.824502845369693e-05,
      "loss": 0.3777,
      "num_input_tokens_seen": 15682200,
      "step": 23980
    },
    {
      "epoch": 12.570754716981131,
      "grad_norm": 0.5714308619499207,
      "learning_rate": 1.8234020179210938e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 15684952,
      "step": 23985
    },
    {
      "epoch": 12.573375262054507,
      "grad_norm": 0.4587361812591553,
      "learning_rate": 1.8223013320090715e-05,
      "loss": 0.3544,
      "num_input_tokens_seen": 15687896,
      "step": 23990
    },
    {
      "epoch": 12.575995807127882,
      "grad_norm": 0.5549058318138123,
      "learning_rate": 1.82120078786388e-05,
      "loss": 0.5651,
      "num_input_tokens_seen": 15690552,
      "step": 23995
    },
    {
      "epoch": 12.578616352201259,
      "grad_norm": 0.9735578894615173,
      "learning_rate": 1.820100385715739e-05,
      "loss": 0.3537,
      "num_input_tokens_seen": 15693528,
      "step": 24000
    },
    {
      "epoch": 12.581236897274634,
      "grad_norm": 0.5307020545005798,
      "learning_rate": 1.8190001257948414e-05,
      "loss": 0.3178,
      "num_input_tokens_seen": 15697464,
      "step": 24005
    },
    {
      "epoch": 12.583857442348009,
      "grad_norm": 0.41364413499832153,
      "learning_rate": 1.8179000083313483e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 15700792,
      "step": 24010
    },
    {
      "epoch": 12.586477987421384,
      "grad_norm": 0.6364686489105225,
      "learning_rate": 1.8168000335553907e-05,
      "loss": 0.5138,
      "num_input_tokens_seen": 15704600,
      "step": 24015
    },
    {
      "epoch": 12.58909853249476,
      "grad_norm": 0.4379199743270874,
      "learning_rate": 1.8157002016970723e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 15707800,
      "step": 24020
    },
    {
      "epoch": 12.591719077568134,
      "grad_norm": 0.3740918040275574,
      "learning_rate": 1.8146005129864644e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 15710616,
      "step": 24025
    },
    {
      "epoch": 12.59433962264151,
      "grad_norm": 0.2806866765022278,
      "learning_rate": 1.81350096765361e-05,
      "loss": 0.3568,
      "num_input_tokens_seen": 15714264,
      "step": 24030
    },
    {
      "epoch": 12.596960167714885,
      "grad_norm": 0.43254533410072327,
      "learning_rate": 1.812401565928521e-05,
      "loss": 0.3702,
      "num_input_tokens_seen": 15717400,
      "step": 24035
    },
    {
      "epoch": 12.59958071278826,
      "grad_norm": 0.6550483703613281,
      "learning_rate": 1.8113023080411806e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 15719864,
      "step": 24040
    },
    {
      "epoch": 12.602201257861635,
      "grad_norm": 0.2491789311170578,
      "learning_rate": 1.810203194221539e-05,
      "loss": 0.5114,
      "num_input_tokens_seen": 15724088,
      "step": 24045
    },
    {
      "epoch": 12.60482180293501,
      "grad_norm": 0.4829045534133911,
      "learning_rate": 1.8091042246995206e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 15727128,
      "step": 24050
    },
    {
      "epoch": 12.607442348008385,
      "grad_norm": 0.6787915229797363,
      "learning_rate": 1.8080053997050154e-05,
      "loss": 0.5921,
      "num_input_tokens_seen": 15729688,
      "step": 24055
    },
    {
      "epoch": 12.61006289308176,
      "grad_norm": 0.6032477021217346,
      "learning_rate": 1.8069067194678867e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 15732696,
      "step": 24060
    },
    {
      "epoch": 12.612683438155136,
      "grad_norm": 0.26575222611427307,
      "learning_rate": 1.8058081842179642e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 15735608,
      "step": 24065
    },
    {
      "epoch": 12.615303983228511,
      "grad_norm": 0.34483107924461365,
      "learning_rate": 1.804709794185051e-05,
      "loss": 0.4334,
      "num_input_tokens_seen": 15739256,
      "step": 24070
    },
    {
      "epoch": 12.617924528301886,
      "grad_norm": 0.3311642110347748,
      "learning_rate": 1.8036115495989164e-05,
      "loss": 0.3639,
      "num_input_tokens_seen": 15741944,
      "step": 24075
    },
    {
      "epoch": 12.620545073375261,
      "grad_norm": 0.39941301941871643,
      "learning_rate": 1.8025134506892997e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 15744888,
      "step": 24080
    },
    {
      "epoch": 12.623165618448636,
      "grad_norm": 0.5646733641624451,
      "learning_rate": 1.8014154976859126e-05,
      "loss": 0.4033,
      "num_input_tokens_seen": 15747736,
      "step": 24085
    },
    {
      "epoch": 12.625786163522012,
      "grad_norm": 0.4824785590171814,
      "learning_rate": 1.800317690818433e-05,
      "loss": 0.4588,
      "num_input_tokens_seen": 15750616,
      "step": 24090
    },
    {
      "epoch": 12.628406708595389,
      "grad_norm": 0.5025907158851624,
      "learning_rate": 1.799220030316511e-05,
      "loss": 0.3379,
      "num_input_tokens_seen": 15754488,
      "step": 24095
    },
    {
      "epoch": 12.631027253668764,
      "grad_norm": 0.5177378058433533,
      "learning_rate": 1.7981225164097633e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 15757432,
      "step": 24100
    },
    {
      "epoch": 12.633647798742139,
      "grad_norm": 0.4987398087978363,
      "learning_rate": 1.7970251493277774e-05,
      "loss": 0.3933,
      "num_input_tokens_seen": 15761208,
      "step": 24105
    },
    {
      "epoch": 12.636268343815514,
      "grad_norm": 0.40111202001571655,
      "learning_rate": 1.7959279293001102e-05,
      "loss": 0.3744,
      "num_input_tokens_seen": 15764152,
      "step": 24110
    },
    {
      "epoch": 12.63888888888889,
      "grad_norm": 0.35785189270973206,
      "learning_rate": 1.7948308565562876e-05,
      "loss": 0.5657,
      "num_input_tokens_seen": 15767064,
      "step": 24115
    },
    {
      "epoch": 12.641509433962264,
      "grad_norm": 0.09670732170343399,
      "learning_rate": 1.793733931325805e-05,
      "loss": 0.4054,
      "num_input_tokens_seen": 15773560,
      "step": 24120
    },
    {
      "epoch": 12.64412997903564,
      "grad_norm": 0.49489426612854004,
      "learning_rate": 1.792637153838127e-05,
      "loss": 0.3784,
      "num_input_tokens_seen": 15777784,
      "step": 24125
    },
    {
      "epoch": 12.646750524109015,
      "grad_norm": 0.3924732804298401,
      "learning_rate": 1.7915405243226864e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 15780760,
      "step": 24130
    },
    {
      "epoch": 12.64937106918239,
      "grad_norm": 0.5427995324134827,
      "learning_rate": 1.7904440430088847e-05,
      "loss": 0.4427,
      "num_input_tokens_seen": 15784312,
      "step": 24135
    },
    {
      "epoch": 12.651991614255765,
      "grad_norm": 0.6058077216148376,
      "learning_rate": 1.7893477101260947e-05,
      "loss": 0.3845,
      "num_input_tokens_seen": 15787000,
      "step": 24140
    },
    {
      "epoch": 12.65461215932914,
      "grad_norm": 0.29575154185295105,
      "learning_rate": 1.7882515259036558e-05,
      "loss": 0.4778,
      "num_input_tokens_seen": 15790104,
      "step": 24145
    },
    {
      "epoch": 12.657232704402515,
      "grad_norm": 0.4513229727745056,
      "learning_rate": 1.787155490570878e-05,
      "loss": 0.4566,
      "num_input_tokens_seen": 15793528,
      "step": 24150
    },
    {
      "epoch": 12.65985324947589,
      "grad_norm": 0.6003416180610657,
      "learning_rate": 1.7860596043570398e-05,
      "loss": 0.489,
      "num_input_tokens_seen": 15795992,
      "step": 24155
    },
    {
      "epoch": 12.662473794549266,
      "grad_norm": 0.7843303680419922,
      "learning_rate": 1.7849638674913863e-05,
      "loss": 0.4284,
      "num_input_tokens_seen": 15798744,
      "step": 24160
    },
    {
      "epoch": 12.665094339622641,
      "grad_norm": 0.4274981915950775,
      "learning_rate": 1.7838682802031344e-05,
      "loss": 0.4714,
      "num_input_tokens_seen": 15801208,
      "step": 24165
    },
    {
      "epoch": 12.667714884696016,
      "grad_norm": 0.2723272740840912,
      "learning_rate": 1.782772842721468e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 15805176,
      "step": 24170
    },
    {
      "epoch": 12.670335429769391,
      "grad_norm": 1.1058030128479004,
      "learning_rate": 1.7816775552755406e-05,
      "loss": 0.4008,
      "num_input_tokens_seen": 15807704,
      "step": 24175
    },
    {
      "epoch": 12.672955974842766,
      "grad_norm": 0.5019263625144958,
      "learning_rate": 1.780582418094473e-05,
      "loss": 0.3985,
      "num_input_tokens_seen": 15810680,
      "step": 24180
    },
    {
      "epoch": 12.675576519916142,
      "grad_norm": 0.767641007900238,
      "learning_rate": 1.7794874314073573e-05,
      "loss": 0.4084,
      "num_input_tokens_seen": 15813688,
      "step": 24185
    },
    {
      "epoch": 12.678197064989519,
      "grad_norm": 0.3158681392669678,
      "learning_rate": 1.7783925954432503e-05,
      "loss": 0.3616,
      "num_input_tokens_seen": 15817752,
      "step": 24190
    },
    {
      "epoch": 12.680817610062894,
      "grad_norm": 0.6270015835762024,
      "learning_rate": 1.777297910431179e-05,
      "loss": 0.5951,
      "num_input_tokens_seen": 15820024,
      "step": 24195
    },
    {
      "epoch": 12.683438155136269,
      "grad_norm": 0.30153974890708923,
      "learning_rate": 1.7762033766001402e-05,
      "loss": 0.3642,
      "num_input_tokens_seen": 15825368,
      "step": 24200
    },
    {
      "epoch": 12.686058700209644,
      "grad_norm": 0.29222652316093445,
      "learning_rate": 1.7751089941790977e-05,
      "loss": 0.4516,
      "num_input_tokens_seen": 15829144,
      "step": 24205
    },
    {
      "epoch": 12.68867924528302,
      "grad_norm": 1.0003509521484375,
      "learning_rate": 1.7740147633969844e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 15832248,
      "step": 24210
    },
    {
      "epoch": 12.691299790356394,
      "grad_norm": 0.4426734149456024,
      "learning_rate": 1.772920684482699e-05,
      "loss": 0.4174,
      "num_input_tokens_seen": 15834840,
      "step": 24215
    },
    {
      "epoch": 12.69392033542977,
      "grad_norm": 0.447377473115921,
      "learning_rate": 1.771826757665112e-05,
      "loss": 0.3557,
      "num_input_tokens_seen": 15837464,
      "step": 24220
    },
    {
      "epoch": 12.696540880503145,
      "grad_norm": 0.3790818750858307,
      "learning_rate": 1.770732983173059e-05,
      "loss": 0.3767,
      "num_input_tokens_seen": 15841752,
      "step": 24225
    },
    {
      "epoch": 12.69916142557652,
      "grad_norm": 0.3375270664691925,
      "learning_rate": 1.7696393612353467e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 15844600,
      "step": 24230
    },
    {
      "epoch": 12.701781970649895,
      "grad_norm": 0.5037317276000977,
      "learning_rate": 1.7685458920807474e-05,
      "loss": 0.369,
      "num_input_tokens_seen": 15847640,
      "step": 24235
    },
    {
      "epoch": 12.70440251572327,
      "grad_norm": 0.6930223703384399,
      "learning_rate": 1.7674525759380033e-05,
      "loss": 0.4842,
      "num_input_tokens_seen": 15850904,
      "step": 24240
    },
    {
      "epoch": 12.707023060796645,
      "grad_norm": 0.5837412476539612,
      "learning_rate": 1.766359413035823e-05,
      "loss": 0.5029,
      "num_input_tokens_seen": 15853880,
      "step": 24245
    },
    {
      "epoch": 12.70964360587002,
      "grad_norm": 0.2277584969997406,
      "learning_rate": 1.7652664036028827e-05,
      "loss": 0.4635,
      "num_input_tokens_seen": 15857016,
      "step": 24250
    },
    {
      "epoch": 12.712264150943396,
      "grad_norm": 0.5313913822174072,
      "learning_rate": 1.7641735478678288e-05,
      "loss": 0.5004,
      "num_input_tokens_seen": 15860888,
      "step": 24255
    },
    {
      "epoch": 12.714884696016771,
      "grad_norm": 0.7171298265457153,
      "learning_rate": 1.763080846059274e-05,
      "loss": 0.4646,
      "num_input_tokens_seen": 15863320,
      "step": 24260
    },
    {
      "epoch": 12.717505241090146,
      "grad_norm": 0.4615815281867981,
      "learning_rate": 1.7619882984057987e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 15867032,
      "step": 24265
    },
    {
      "epoch": 12.720125786163521,
      "grad_norm": 0.5345031023025513,
      "learning_rate": 1.760895905135953e-05,
      "loss": 0.3435,
      "num_input_tokens_seen": 15870552,
      "step": 24270
    },
    {
      "epoch": 12.722746331236896,
      "grad_norm": 0.5038386583328247,
      "learning_rate": 1.7598036664782508e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 15873976,
      "step": 24275
    },
    {
      "epoch": 12.725366876310272,
      "grad_norm": 0.838146448135376,
      "learning_rate": 1.7587115826611766e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 15876664,
      "step": 24280
    },
    {
      "epoch": 12.727987421383649,
      "grad_norm": 0.7936538457870483,
      "learning_rate": 1.7576196539131828e-05,
      "loss": 0.5461,
      "num_input_tokens_seen": 15879736,
      "step": 24285
    },
    {
      "epoch": 12.730607966457024,
      "grad_norm": 0.5469485521316528,
      "learning_rate": 1.7565278804626872e-05,
      "loss": 0.4475,
      "num_input_tokens_seen": 15883192,
      "step": 24290
    },
    {
      "epoch": 12.733228511530399,
      "grad_norm": 0.6725106835365295,
      "learning_rate": 1.755436262538078e-05,
      "loss": 0.5054,
      "num_input_tokens_seen": 15886520,
      "step": 24295
    },
    {
      "epoch": 12.735849056603774,
      "grad_norm": 0.5400012135505676,
      "learning_rate": 1.7543448003677082e-05,
      "loss": 0.4936,
      "num_input_tokens_seen": 15890584,
      "step": 24300
    },
    {
      "epoch": 12.73846960167715,
      "grad_norm": 0.4603731036186218,
      "learning_rate": 1.7532534941798985e-05,
      "loss": 0.3832,
      "num_input_tokens_seen": 15894776,
      "step": 24305
    },
    {
      "epoch": 12.741090146750524,
      "grad_norm": 0.29802465438842773,
      "learning_rate": 1.752162344202939e-05,
      "loss": 0.5398,
      "num_input_tokens_seen": 15903000,
      "step": 24310
    },
    {
      "epoch": 12.7437106918239,
      "grad_norm": 0.6495882868766785,
      "learning_rate": 1.751071350665085e-05,
      "loss": 0.4143,
      "num_input_tokens_seen": 15905624,
      "step": 24315
    },
    {
      "epoch": 12.746331236897275,
      "grad_norm": 0.4577171802520752,
      "learning_rate": 1.7499805137945602e-05,
      "loss": 0.4458,
      "num_input_tokens_seen": 15909336,
      "step": 24320
    },
    {
      "epoch": 12.74895178197065,
      "grad_norm": 0.23598502576351166,
      "learning_rate": 1.7488898338195566e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 15912664,
      "step": 24325
    },
    {
      "epoch": 12.751572327044025,
      "grad_norm": 0.562468945980072,
      "learning_rate": 1.7477993109682288e-05,
      "loss": 0.4487,
      "num_input_tokens_seen": 15915928,
      "step": 24330
    },
    {
      "epoch": 12.7541928721174,
      "grad_norm": 0.27548596262931824,
      "learning_rate": 1.7467089454687043e-05,
      "loss": 0.4877,
      "num_input_tokens_seen": 15918264,
      "step": 24335
    },
    {
      "epoch": 12.756813417190775,
      "grad_norm": 0.7605257034301758,
      "learning_rate": 1.745618737549074e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 15921112,
      "step": 24340
    },
    {
      "epoch": 12.75943396226415,
      "grad_norm": 0.650299608707428,
      "learning_rate": 1.7445286874373974e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 15923512,
      "step": 24345
    },
    {
      "epoch": 12.762054507337526,
      "grad_norm": 0.4101924002170563,
      "learning_rate": 1.743438795361701e-05,
      "loss": 0.521,
      "num_input_tokens_seen": 15926296,
      "step": 24350
    },
    {
      "epoch": 12.764675052410901,
      "grad_norm": 0.3599192798137665,
      "learning_rate": 1.742349061549978e-05,
      "loss": 0.3946,
      "num_input_tokens_seen": 15930520,
      "step": 24355
    },
    {
      "epoch": 12.767295597484276,
      "grad_norm": 0.4017990529537201,
      "learning_rate": 1.741259486230186e-05,
      "loss": 0.3862,
      "num_input_tokens_seen": 15933592,
      "step": 24360
    },
    {
      "epoch": 12.769916142557651,
      "grad_norm": 0.294926255941391,
      "learning_rate": 1.7401700696302546e-05,
      "loss": 0.2992,
      "num_input_tokens_seen": 15937016,
      "step": 24365
    },
    {
      "epoch": 12.772536687631026,
      "grad_norm": 0.4367160201072693,
      "learning_rate": 1.739080811978075e-05,
      "loss": 0.6282,
      "num_input_tokens_seen": 15940280,
      "step": 24370
    },
    {
      "epoch": 12.775157232704402,
      "grad_norm": 0.5383965969085693,
      "learning_rate": 1.7379917135015093e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 15943864,
      "step": 24375
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 0.7260422110557556,
      "learning_rate": 1.736902774428384e-05,
      "loss": 0.5949,
      "num_input_tokens_seen": 15946904,
      "step": 24380
    },
    {
      "epoch": 12.780398322851154,
      "grad_norm": 0.5611492395401001,
      "learning_rate": 1.735813994986491e-05,
      "loss": 0.4905,
      "num_input_tokens_seen": 15949688,
      "step": 24385
    },
    {
      "epoch": 12.783018867924529,
      "grad_norm": 0.6918704509735107,
      "learning_rate": 1.7347253754035933e-05,
      "loss": 0.3432,
      "num_input_tokens_seen": 15953016,
      "step": 24390
    },
    {
      "epoch": 12.785639412997904,
      "grad_norm": 0.6598786115646362,
      "learning_rate": 1.7336369159074154e-05,
      "loss": 0.4504,
      "num_input_tokens_seen": 15955608,
      "step": 24395
    },
    {
      "epoch": 12.78825995807128,
      "grad_norm": 0.45453378558158875,
      "learning_rate": 1.7325486167256522e-05,
      "loss": 0.5086,
      "num_input_tokens_seen": 15959608,
      "step": 24400
    },
    {
      "epoch": 12.790880503144654,
      "grad_norm": 0.5975262522697449,
      "learning_rate": 1.7314604780859623e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 15963000,
      "step": 24405
    },
    {
      "epoch": 12.79350104821803,
      "grad_norm": 0.6750268936157227,
      "learning_rate": 1.730372500215973e-05,
      "loss": 0.51,
      "num_input_tokens_seen": 15967928,
      "step": 24410
    },
    {
      "epoch": 12.796121593291405,
      "grad_norm": 0.5522945523262024,
      "learning_rate": 1.7292846833432763e-05,
      "loss": 0.2969,
      "num_input_tokens_seen": 15971416,
      "step": 24415
    },
    {
      "epoch": 12.79874213836478,
      "grad_norm": 0.46359217166900635,
      "learning_rate": 1.7281970276954305e-05,
      "loss": 0.4884,
      "num_input_tokens_seen": 15974104,
      "step": 24420
    },
    {
      "epoch": 12.801362683438155,
      "grad_norm": 0.4000101089477539,
      "learning_rate": 1.727109533499961e-05,
      "loss": 0.3948,
      "num_input_tokens_seen": 15978200,
      "step": 24425
    },
    {
      "epoch": 12.80398322851153,
      "grad_norm": 0.4576147794723511,
      "learning_rate": 1.72602220098436e-05,
      "loss": 0.3806,
      "num_input_tokens_seen": 15981624,
      "step": 24430
    },
    {
      "epoch": 12.806603773584905,
      "grad_norm": 0.6430555582046509,
      "learning_rate": 1.7249350303760846e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 15985016,
      "step": 24435
    },
    {
      "epoch": 12.80922431865828,
      "grad_norm": 0.46514204144477844,
      "learning_rate": 1.7238480219025593e-05,
      "loss": 0.5981,
      "num_input_tokens_seen": 15988856,
      "step": 24440
    },
    {
      "epoch": 12.811844863731656,
      "grad_norm": 0.2845120131969452,
      "learning_rate": 1.7227611757911723e-05,
      "loss": 0.4729,
      "num_input_tokens_seen": 15992568,
      "step": 24445
    },
    {
      "epoch": 12.814465408805031,
      "grad_norm": 0.5613448619842529,
      "learning_rate": 1.7216744922692794e-05,
      "loss": 0.3168,
      "num_input_tokens_seen": 15996216,
      "step": 24450
    },
    {
      "epoch": 12.817085953878406,
      "grad_norm": 0.9373831152915955,
      "learning_rate": 1.7205879715642037e-05,
      "loss": 0.5197,
      "num_input_tokens_seen": 15999224,
      "step": 24455
    },
    {
      "epoch": 12.819706498951781,
      "grad_norm": 0.4013710021972656,
      "learning_rate": 1.719501613903232e-05,
      "loss": 0.4701,
      "num_input_tokens_seen": 16002456,
      "step": 24460
    },
    {
      "epoch": 12.822327044025156,
      "grad_norm": 0.3009418845176697,
      "learning_rate": 1.7184154195136192e-05,
      "loss": 0.4066,
      "num_input_tokens_seen": 16005240,
      "step": 24465
    },
    {
      "epoch": 12.824947589098532,
      "grad_norm": 0.359720915555954,
      "learning_rate": 1.7173293886225837e-05,
      "loss": 0.3549,
      "num_input_tokens_seen": 16009464,
      "step": 24470
    },
    {
      "epoch": 12.827568134171909,
      "grad_norm": 0.3382495045661926,
      "learning_rate": 1.71624352145731e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 16012888,
      "step": 24475
    },
    {
      "epoch": 12.830188679245284,
      "grad_norm": 0.52149897813797,
      "learning_rate": 1.7151578182449508e-05,
      "loss": 0.5931,
      "num_input_tokens_seen": 16015960,
      "step": 24480
    },
    {
      "epoch": 12.832809224318659,
      "grad_norm": 0.6534160375595093,
      "learning_rate": 1.7140722792126212e-05,
      "loss": 0.4143,
      "num_input_tokens_seen": 16019704,
      "step": 24485
    },
    {
      "epoch": 12.835429769392034,
      "grad_norm": 0.3308013379573822,
      "learning_rate": 1.7129869045874053e-05,
      "loss": 0.4606,
      "num_input_tokens_seen": 16022488,
      "step": 24490
    },
    {
      "epoch": 12.83805031446541,
      "grad_norm": 0.3521662652492523,
      "learning_rate": 1.7119016945963506e-05,
      "loss": 0.3048,
      "num_input_tokens_seen": 16025496,
      "step": 24495
    },
    {
      "epoch": 12.840670859538784,
      "grad_norm": 0.4265153408050537,
      "learning_rate": 1.7108166494664694e-05,
      "loss": 0.4973,
      "num_input_tokens_seen": 16028152,
      "step": 24500
    },
    {
      "epoch": 12.84329140461216,
      "grad_norm": 0.5152459740638733,
      "learning_rate": 1.7097317694247417e-05,
      "loss": 0.3288,
      "num_input_tokens_seen": 16031352,
      "step": 24505
    },
    {
      "epoch": 12.845911949685535,
      "grad_norm": 0.7453183531761169,
      "learning_rate": 1.7086470546981122e-05,
      "loss": 0.4216,
      "num_input_tokens_seen": 16035736,
      "step": 24510
    },
    {
      "epoch": 12.84853249475891,
      "grad_norm": 0.5160130262374878,
      "learning_rate": 1.7075625055134904e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 16040536,
      "step": 24515
    },
    {
      "epoch": 12.851153039832285,
      "grad_norm": 0.46350228786468506,
      "learning_rate": 1.706478122097752e-05,
      "loss": 0.5213,
      "num_input_tokens_seen": 16043960,
      "step": 24520
    },
    {
      "epoch": 12.85377358490566,
      "grad_norm": 0.4625379145145416,
      "learning_rate": 1.7053939046777384e-05,
      "loss": 0.4422,
      "num_input_tokens_seen": 16047800,
      "step": 24525
    },
    {
      "epoch": 12.856394129979035,
      "grad_norm": 0.5955475568771362,
      "learning_rate": 1.7043098534802533e-05,
      "loss": 0.4852,
      "num_input_tokens_seen": 16050360,
      "step": 24530
    },
    {
      "epoch": 12.85901467505241,
      "grad_norm": 0.48738282918930054,
      "learning_rate": 1.7032259687320695e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 16053176,
      "step": 24535
    },
    {
      "epoch": 12.861635220125786,
      "grad_norm": 0.44359469413757324,
      "learning_rate": 1.7021422506599228e-05,
      "loss": 0.3821,
      "num_input_tokens_seen": 16055896,
      "step": 24540
    },
    {
      "epoch": 12.864255765199161,
      "grad_norm": 0.5364851355552673,
      "learning_rate": 1.7010586994905155e-05,
      "loss": 0.5018,
      "num_input_tokens_seen": 16058968,
      "step": 24545
    },
    {
      "epoch": 12.866876310272536,
      "grad_norm": 0.4015810191631317,
      "learning_rate": 1.6999753154505123e-05,
      "loss": 0.3458,
      "num_input_tokens_seen": 16061912,
      "step": 24550
    },
    {
      "epoch": 12.869496855345911,
      "grad_norm": 0.3092220425605774,
      "learning_rate": 1.6988920987665476e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 16064504,
      "step": 24555
    },
    {
      "epoch": 12.872117400419286,
      "grad_norm": 0.4193103611469269,
      "learning_rate": 1.6978090496652156e-05,
      "loss": 0.4547,
      "num_input_tokens_seen": 16067320,
      "step": 24560
    },
    {
      "epoch": 12.874737945492662,
      "grad_norm": 0.6648996472358704,
      "learning_rate": 1.6967261683730778e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 16070968,
      "step": 24565
    },
    {
      "epoch": 12.877358490566039,
      "grad_norm": 0.3214508295059204,
      "learning_rate": 1.695643455116662e-05,
      "loss": 0.4861,
      "num_input_tokens_seen": 16074040,
      "step": 24570
    },
    {
      "epoch": 12.879979035639414,
      "grad_norm": 0.5290654301643372,
      "learning_rate": 1.694560910122458e-05,
      "loss": 0.444,
      "num_input_tokens_seen": 16076792,
      "step": 24575
    },
    {
      "epoch": 12.882599580712789,
      "grad_norm": 0.547207772731781,
      "learning_rate": 1.693478533616924e-05,
      "loss": 0.4205,
      "num_input_tokens_seen": 16080376,
      "step": 24580
    },
    {
      "epoch": 12.885220125786164,
      "grad_norm": 0.2589767277240753,
      "learning_rate": 1.692396325826479e-05,
      "loss": 0.3927,
      "num_input_tokens_seen": 16083640,
      "step": 24585
    },
    {
      "epoch": 12.88784067085954,
      "grad_norm": 0.3438797891139984,
      "learning_rate": 1.6913142869775088e-05,
      "loss": 0.3884,
      "num_input_tokens_seen": 16087064,
      "step": 24590
    },
    {
      "epoch": 12.890461215932914,
      "grad_norm": 0.5463738441467285,
      "learning_rate": 1.6902324172963635e-05,
      "loss": 0.5898,
      "num_input_tokens_seen": 16090584,
      "step": 24595
    },
    {
      "epoch": 12.89308176100629,
      "grad_norm": 0.6661970019340515,
      "learning_rate": 1.6891507170093586e-05,
      "loss": 0.5756,
      "num_input_tokens_seen": 16094200,
      "step": 24600
    },
    {
      "epoch": 12.895702306079665,
      "grad_norm": 0.483860582113266,
      "learning_rate": 1.6880691863427725e-05,
      "loss": 0.5801,
      "num_input_tokens_seen": 16096952,
      "step": 24605
    },
    {
      "epoch": 12.89832285115304,
      "grad_norm": 0.4021174907684326,
      "learning_rate": 1.686987825522851e-05,
      "loss": 0.533,
      "num_input_tokens_seen": 16100024,
      "step": 24610
    },
    {
      "epoch": 12.900943396226415,
      "grad_norm": 0.32237112522125244,
      "learning_rate": 1.6859066347758002e-05,
      "loss": 0.5035,
      "num_input_tokens_seen": 16104824,
      "step": 24615
    },
    {
      "epoch": 12.90356394129979,
      "grad_norm": 0.7431371808052063,
      "learning_rate": 1.6848256143277933e-05,
      "loss": 0.3353,
      "num_input_tokens_seen": 16108536,
      "step": 24620
    },
    {
      "epoch": 12.906184486373165,
      "grad_norm": 0.41679465770721436,
      "learning_rate": 1.683744764404968e-05,
      "loss": 0.4339,
      "num_input_tokens_seen": 16111864,
      "step": 24625
    },
    {
      "epoch": 12.90880503144654,
      "grad_norm": 0.31997355818748474,
      "learning_rate": 1.6826640852334257e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 16115640,
      "step": 24630
    },
    {
      "epoch": 12.911425576519916,
      "grad_norm": 0.31430912017822266,
      "learning_rate": 1.6815835770392324e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 16119032,
      "step": 24635
    },
    {
      "epoch": 12.914046121593291,
      "grad_norm": 0.3671650290489197,
      "learning_rate": 1.6805032400484184e-05,
      "loss": 0.3606,
      "num_input_tokens_seen": 16122200,
      "step": 24640
    },
    {
      "epoch": 12.916666666666666,
      "grad_norm": 0.6354057788848877,
      "learning_rate": 1.6794230744869767e-05,
      "loss": 0.4581,
      "num_input_tokens_seen": 16125080,
      "step": 24645
    },
    {
      "epoch": 12.919287211740041,
      "grad_norm": 0.5310061573982239,
      "learning_rate": 1.6783430805808658e-05,
      "loss": 0.4851,
      "num_input_tokens_seen": 16129304,
      "step": 24650
    },
    {
      "epoch": 12.921907756813416,
      "grad_norm": 0.39865249395370483,
      "learning_rate": 1.6772632585560084e-05,
      "loss": 0.6219,
      "num_input_tokens_seen": 16132664,
      "step": 24655
    },
    {
      "epoch": 12.924528301886792,
      "grad_norm": 0.3910446763038635,
      "learning_rate": 1.6761836086382908e-05,
      "loss": 0.3875,
      "num_input_tokens_seen": 16135608,
      "step": 24660
    },
    {
      "epoch": 12.927148846960169,
      "grad_norm": 0.6638662219047546,
      "learning_rate": 1.6751041310535653e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 16138584,
      "step": 24665
    },
    {
      "epoch": 12.929769392033544,
      "grad_norm": 0.8691232204437256,
      "learning_rate": 1.6740248260276437e-05,
      "loss": 0.4447,
      "num_input_tokens_seen": 16141176,
      "step": 24670
    },
    {
      "epoch": 12.932389937106919,
      "grad_norm": 0.7980852127075195,
      "learning_rate": 1.6729456937863047e-05,
      "loss": 0.3593,
      "num_input_tokens_seen": 16144024,
      "step": 24675
    },
    {
      "epoch": 12.935010482180294,
      "grad_norm": 0.5776178240776062,
      "learning_rate": 1.6718667345552918e-05,
      "loss": 0.3674,
      "num_input_tokens_seen": 16148120,
      "step": 24680
    },
    {
      "epoch": 12.93763102725367,
      "grad_norm": 0.6546623110771179,
      "learning_rate": 1.6707879485603097e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 16151416,
      "step": 24685
    },
    {
      "epoch": 12.940251572327044,
      "grad_norm": 0.5909249782562256,
      "learning_rate": 1.669709336027029e-05,
      "loss": 0.6214,
      "num_input_tokens_seen": 16154392,
      "step": 24690
    },
    {
      "epoch": 12.94287211740042,
      "grad_norm": 0.4205284118652344,
      "learning_rate": 1.6686308971810833e-05,
      "loss": 0.513,
      "num_input_tokens_seen": 16157048,
      "step": 24695
    },
    {
      "epoch": 12.945492662473795,
      "grad_norm": 0.5281216502189636,
      "learning_rate": 1.6675526322480677e-05,
      "loss": 0.5576,
      "num_input_tokens_seen": 16160152,
      "step": 24700
    },
    {
      "epoch": 12.94811320754717,
      "grad_norm": 0.5625064373016357,
      "learning_rate": 1.6664745414535454e-05,
      "loss": 0.4553,
      "num_input_tokens_seen": 16163352,
      "step": 24705
    },
    {
      "epoch": 12.950733752620545,
      "grad_norm": 0.37172868847846985,
      "learning_rate": 1.665396625023039e-05,
      "loss": 0.5602,
      "num_input_tokens_seen": 16166488,
      "step": 24710
    },
    {
      "epoch": 12.95335429769392,
      "grad_norm": 0.8432822823524475,
      "learning_rate": 1.6643188831820375e-05,
      "loss": 0.4471,
      "num_input_tokens_seen": 16169560,
      "step": 24715
    },
    {
      "epoch": 12.955974842767295,
      "grad_norm": 0.3290444016456604,
      "learning_rate": 1.6632413161559912e-05,
      "loss": 0.3458,
      "num_input_tokens_seen": 16172376,
      "step": 24720
    },
    {
      "epoch": 12.95859538784067,
      "grad_norm": 0.3916812837123871,
      "learning_rate": 1.6621639241703162e-05,
      "loss": 0.4767,
      "num_input_tokens_seen": 16175512,
      "step": 24725
    },
    {
      "epoch": 12.961215932914046,
      "grad_norm": 0.5656834244728088,
      "learning_rate": 1.6610867074503893e-05,
      "loss": 0.4184,
      "num_input_tokens_seen": 16178744,
      "step": 24730
    },
    {
      "epoch": 12.963836477987421,
      "grad_norm": 0.7358859181404114,
      "learning_rate": 1.660009666221552e-05,
      "loss": 0.5185,
      "num_input_tokens_seen": 16182936,
      "step": 24735
    },
    {
      "epoch": 12.966457023060796,
      "grad_norm": 0.6086152195930481,
      "learning_rate": 1.6589328007091093e-05,
      "loss": 0.447,
      "num_input_tokens_seen": 16185880,
      "step": 24740
    },
    {
      "epoch": 12.969077568134171,
      "grad_norm": 0.6061739325523376,
      "learning_rate": 1.65785611113833e-05,
      "loss": 0.4195,
      "num_input_tokens_seen": 16188792,
      "step": 24745
    },
    {
      "epoch": 12.971698113207546,
      "grad_norm": 0.44415396451950073,
      "learning_rate": 1.6567795977344447e-05,
      "loss": 0.3476,
      "num_input_tokens_seen": 16191480,
      "step": 24750
    },
    {
      "epoch": 12.974318658280922,
      "grad_norm": 0.5356634259223938,
      "learning_rate": 1.6557032607226466e-05,
      "loss": 0.3368,
      "num_input_tokens_seen": 16193784,
      "step": 24755
    },
    {
      "epoch": 12.976939203354299,
      "grad_norm": 0.2598143219947815,
      "learning_rate": 1.654627100328095e-05,
      "loss": 0.3796,
      "num_input_tokens_seen": 16196312,
      "step": 24760
    },
    {
      "epoch": 12.979559748427674,
      "grad_norm": 0.42165589332580566,
      "learning_rate": 1.6535511167759085e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 16199352,
      "step": 24765
    },
    {
      "epoch": 12.982180293501049,
      "grad_norm": 0.5819838643074036,
      "learning_rate": 1.652475310291172e-05,
      "loss": 0.431,
      "num_input_tokens_seen": 16202872,
      "step": 24770
    },
    {
      "epoch": 12.984800838574424,
      "grad_norm": 0.9299513697624207,
      "learning_rate": 1.6513996810989317e-05,
      "loss": 0.36,
      "num_input_tokens_seen": 16205944,
      "step": 24775
    },
    {
      "epoch": 12.9874213836478,
      "grad_norm": 0.32830771803855896,
      "learning_rate": 1.6503242294241973e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 16208984,
      "step": 24780
    },
    {
      "epoch": 12.990041928721174,
      "grad_norm": 0.30899128317832947,
      "learning_rate": 1.6492489554919404e-05,
      "loss": 0.5142,
      "num_input_tokens_seen": 16212152,
      "step": 24785
    },
    {
      "epoch": 12.99266247379455,
      "grad_norm": 0.5788451433181763,
      "learning_rate": 1.6481738595270953e-05,
      "loss": 0.3704,
      "num_input_tokens_seen": 16216312,
      "step": 24790
    },
    {
      "epoch": 12.995283018867925,
      "grad_norm": 0.37150660157203674,
      "learning_rate": 1.6470989417545608e-05,
      "loss": 0.3635,
      "num_input_tokens_seen": 16220120,
      "step": 24795
    },
    {
      "epoch": 12.9979035639413,
      "grad_norm": 0.689288854598999,
      "learning_rate": 1.646024202399197e-05,
      "loss": 0.4162,
      "num_input_tokens_seen": 16223384,
      "step": 24800
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.4473709166049957,
      "eval_runtime": 9.6473,
      "eval_samples_per_second": 87.9,
      "eval_steps_per_second": 21.975,
      "num_input_tokens_seen": 16225896,
      "step": 24804
    },
    {
      "epoch": 13.000524109014675,
      "grad_norm": 0.5314232110977173,
      "learning_rate": 1.6449496416858284e-05,
      "loss": 0.5426,
      "num_input_tokens_seen": 16226312,
      "step": 24805
    },
    {
      "epoch": 13.00314465408805,
      "grad_norm": 0.5610303282737732,
      "learning_rate": 1.64387525983924e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 16230504,
      "step": 24810
    },
    {
      "epoch": 13.005765199161425,
      "grad_norm": 0.5895457863807678,
      "learning_rate": 1.6428010570841788e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 16234088,
      "step": 24815
    },
    {
      "epoch": 13.0083857442348,
      "grad_norm": 0.5429096817970276,
      "learning_rate": 1.6417270336453573e-05,
      "loss": 0.3186,
      "num_input_tokens_seen": 16237288,
      "step": 24820
    },
    {
      "epoch": 13.011006289308176,
      "grad_norm": 0.43463990092277527,
      "learning_rate": 1.6406531897474492e-05,
      "loss": 0.3761,
      "num_input_tokens_seen": 16240904,
      "step": 24825
    },
    {
      "epoch": 13.01362683438155,
      "grad_norm": 0.5005583167076111,
      "learning_rate": 1.639579525615089e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 16243752,
      "step": 24830
    },
    {
      "epoch": 13.016247379454926,
      "grad_norm": 0.6071816682815552,
      "learning_rate": 1.638506041472877e-05,
      "loss": 0.3486,
      "num_input_tokens_seen": 16247368,
      "step": 24835
    },
    {
      "epoch": 13.018867924528301,
      "grad_norm": 0.7529378533363342,
      "learning_rate": 1.6374327375453724e-05,
      "loss": 0.4208,
      "num_input_tokens_seen": 16250088,
      "step": 24840
    },
    {
      "epoch": 13.021488469601676,
      "grad_norm": 0.3255559504032135,
      "learning_rate": 1.6363596140570982e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 16254024,
      "step": 24845
    },
    {
      "epoch": 13.024109014675052,
      "grad_norm": 0.5138863325119019,
      "learning_rate": 1.6352866712325395e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 16256040,
      "step": 24850
    },
    {
      "epoch": 13.026729559748428,
      "grad_norm": 0.572303056716919,
      "learning_rate": 1.634213909296144e-05,
      "loss": 0.4333,
      "num_input_tokens_seen": 16259016,
      "step": 24855
    },
    {
      "epoch": 13.029350104821804,
      "grad_norm": 0.26479828357696533,
      "learning_rate": 1.6331413284723217e-05,
      "loss": 0.4561,
      "num_input_tokens_seen": 16261576,
      "step": 24860
    },
    {
      "epoch": 13.031970649895179,
      "grad_norm": 0.4392484426498413,
      "learning_rate": 1.6320689289854442e-05,
      "loss": 0.4199,
      "num_input_tokens_seen": 16265544,
      "step": 24865
    },
    {
      "epoch": 13.034591194968554,
      "grad_norm": 0.4400542676448822,
      "learning_rate": 1.6309967110598438e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 16267944,
      "step": 24870
    },
    {
      "epoch": 13.03721174004193,
      "grad_norm": 0.9211270213127136,
      "learning_rate": 1.6299246749198183e-05,
      "loss": 0.6086,
      "num_input_tokens_seen": 16270856,
      "step": 24875
    },
    {
      "epoch": 13.039832285115304,
      "grad_norm": 0.4871114194393158,
      "learning_rate": 1.6288528207896236e-05,
      "loss": 0.2904,
      "num_input_tokens_seen": 16273480,
      "step": 24880
    },
    {
      "epoch": 13.04245283018868,
      "grad_norm": 0.33489009737968445,
      "learning_rate": 1.627781148893481e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 16276712,
      "step": 24885
    },
    {
      "epoch": 13.045073375262055,
      "grad_norm": 0.37567245960235596,
      "learning_rate": 1.6267096594555703e-05,
      "loss": 0.4405,
      "num_input_tokens_seen": 16279720,
      "step": 24890
    },
    {
      "epoch": 13.04769392033543,
      "grad_norm": 0.5526001453399658,
      "learning_rate": 1.6256383527000372e-05,
      "loss": 0.3625,
      "num_input_tokens_seen": 16285128,
      "step": 24895
    },
    {
      "epoch": 13.050314465408805,
      "grad_norm": 0.4071284830570221,
      "learning_rate": 1.6245672288509843e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 16287880,
      "step": 24900
    },
    {
      "epoch": 13.05293501048218,
      "grad_norm": 0.5570069551467896,
      "learning_rate": 1.62349628813248e-05,
      "loss": 0.447,
      "num_input_tokens_seen": 16290856,
      "step": 24905
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 0.530636191368103,
      "learning_rate": 1.6224255307685525e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 16293800,
      "step": 24910
    },
    {
      "epoch": 13.05817610062893,
      "grad_norm": 0.6376869082450867,
      "learning_rate": 1.6213549569831922e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 16296680,
      "step": 24915
    },
    {
      "epoch": 13.060796645702306,
      "grad_norm": 0.5503666996955872,
      "learning_rate": 1.62028456700035e-05,
      "loss": 0.4422,
      "num_input_tokens_seen": 16299432,
      "step": 24920
    },
    {
      "epoch": 13.06341719077568,
      "grad_norm": 0.31877535581588745,
      "learning_rate": 1.619214361043942e-05,
      "loss": 0.4626,
      "num_input_tokens_seen": 16302984,
      "step": 24925
    },
    {
      "epoch": 13.066037735849056,
      "grad_norm": 0.3309158980846405,
      "learning_rate": 1.61814433933784e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 16306280,
      "step": 24930
    },
    {
      "epoch": 13.068658280922431,
      "grad_norm": 0.40034544467926025,
      "learning_rate": 1.617074502105881e-05,
      "loss": 0.3852,
      "num_input_tokens_seen": 16309736,
      "step": 24935
    },
    {
      "epoch": 13.071278825995806,
      "grad_norm": 0.36238205432891846,
      "learning_rate": 1.616004849571864e-05,
      "loss": 0.3176,
      "num_input_tokens_seen": 16313288,
      "step": 24940
    },
    {
      "epoch": 13.073899371069182,
      "grad_norm": 0.39280667901039124,
      "learning_rate": 1.6149353819595463e-05,
      "loss": 0.4346,
      "num_input_tokens_seen": 16315816,
      "step": 24945
    },
    {
      "epoch": 13.076519916142558,
      "grad_norm": 0.9511245489120483,
      "learning_rate": 1.6138660994926508e-05,
      "loss": 0.3853,
      "num_input_tokens_seen": 16319144,
      "step": 24950
    },
    {
      "epoch": 13.079140461215934,
      "grad_norm": 0.32260510325431824,
      "learning_rate": 1.612797002394857e-05,
      "loss": 0.4775,
      "num_input_tokens_seen": 16323464,
      "step": 24955
    },
    {
      "epoch": 13.081761006289309,
      "grad_norm": 0.9424399137496948,
      "learning_rate": 1.611728090889808e-05,
      "loss": 0.6877,
      "num_input_tokens_seen": 16330280,
      "step": 24960
    },
    {
      "epoch": 13.084381551362684,
      "grad_norm": 0.3841564655303955,
      "learning_rate": 1.610659365201109e-05,
      "loss": 0.5941,
      "num_input_tokens_seen": 16333512,
      "step": 24965
    },
    {
      "epoch": 13.08700209643606,
      "grad_norm": 0.6381775736808777,
      "learning_rate": 1.609590825552324e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 16335880,
      "step": 24970
    },
    {
      "epoch": 13.089622641509434,
      "grad_norm": 0.378644198179245,
      "learning_rate": 1.6085224721669807e-05,
      "loss": 0.3616,
      "num_input_tokens_seen": 16338888,
      "step": 24975
    },
    {
      "epoch": 13.09224318658281,
      "grad_norm": 0.6851485967636108,
      "learning_rate": 1.607454305268566e-05,
      "loss": 0.3442,
      "num_input_tokens_seen": 16341928,
      "step": 24980
    },
    {
      "epoch": 13.094863731656185,
      "grad_norm": 0.2693236172199249,
      "learning_rate": 1.606386325080528e-05,
      "loss": 0.3288,
      "num_input_tokens_seen": 16346088,
      "step": 24985
    },
    {
      "epoch": 13.09748427672956,
      "grad_norm": 0.492654949426651,
      "learning_rate": 1.6053185318262746e-05,
      "loss": 0.5124,
      "num_input_tokens_seen": 16349224,
      "step": 24990
    },
    {
      "epoch": 13.100104821802935,
      "grad_norm": 0.3616012930870056,
      "learning_rate": 1.6042509257291783e-05,
      "loss": 0.4456,
      "num_input_tokens_seen": 16351656,
      "step": 24995
    },
    {
      "epoch": 13.10272536687631,
      "grad_norm": 0.6199916005134583,
      "learning_rate": 1.6031835070125685e-05,
      "loss": 0.3232,
      "num_input_tokens_seen": 16354952,
      "step": 25000
    },
    {
      "epoch": 13.105345911949685,
      "grad_norm": 0.26774492859840393,
      "learning_rate": 1.6021162758997382e-05,
      "loss": 0.3712,
      "num_input_tokens_seen": 16358504,
      "step": 25005
    },
    {
      "epoch": 13.10796645702306,
      "grad_norm": 0.29380080103874207,
      "learning_rate": 1.6010492326139398e-05,
      "loss": 0.4047,
      "num_input_tokens_seen": 16361544,
      "step": 25010
    },
    {
      "epoch": 13.110587002096436,
      "grad_norm": 0.5260003209114075,
      "learning_rate": 1.599982377378385e-05,
      "loss": 0.7408,
      "num_input_tokens_seen": 16364904,
      "step": 25015
    },
    {
      "epoch": 13.11320754716981,
      "grad_norm": 0.5689731240272522,
      "learning_rate": 1.5989157104162496e-05,
      "loss": 0.4364,
      "num_input_tokens_seen": 16368104,
      "step": 25020
    },
    {
      "epoch": 13.115828092243186,
      "grad_norm": 0.37698742747306824,
      "learning_rate": 1.5978492319506673e-05,
      "loss": 0.3681,
      "num_input_tokens_seen": 16370600,
      "step": 25025
    },
    {
      "epoch": 13.118448637316561,
      "grad_norm": 0.835131049156189,
      "learning_rate": 1.596782942204733e-05,
      "loss": 0.4824,
      "num_input_tokens_seen": 16373672,
      "step": 25030
    },
    {
      "epoch": 13.121069182389936,
      "grad_norm": 0.495157927274704,
      "learning_rate": 1.595716841401504e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 16376904,
      "step": 25035
    },
    {
      "epoch": 13.123689727463312,
      "grad_norm": 0.3020474314689636,
      "learning_rate": 1.594650929763994e-05,
      "loss": 0.5221,
      "num_input_tokens_seen": 16379400,
      "step": 25040
    },
    {
      "epoch": 13.126310272536688,
      "grad_norm": 0.35801276564598083,
      "learning_rate": 1.5935852075151812e-05,
      "loss": 0.6128,
      "num_input_tokens_seen": 16382696,
      "step": 25045
    },
    {
      "epoch": 13.128930817610064,
      "grad_norm": 0.3687781095504761,
      "learning_rate": 1.5925196748780016e-05,
      "loss": 0.3758,
      "num_input_tokens_seen": 16385960,
      "step": 25050
    },
    {
      "epoch": 13.131551362683439,
      "grad_norm": 0.3949945569038391,
      "learning_rate": 1.5914543320753528e-05,
      "loss": 0.4679,
      "num_input_tokens_seen": 16389896,
      "step": 25055
    },
    {
      "epoch": 13.134171907756814,
      "grad_norm": 0.4738439917564392,
      "learning_rate": 1.5903891793300932e-05,
      "loss": 0.3614,
      "num_input_tokens_seen": 16393608,
      "step": 25060
    },
    {
      "epoch": 13.13679245283019,
      "grad_norm": 0.4102904200553894,
      "learning_rate": 1.5893242168650403e-05,
      "loss": 0.3936,
      "num_input_tokens_seen": 16397704,
      "step": 25065
    },
    {
      "epoch": 13.139412997903564,
      "grad_norm": 0.48330074548721313,
      "learning_rate": 1.5882594449029705e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 16402888,
      "step": 25070
    },
    {
      "epoch": 13.14203354297694,
      "grad_norm": 0.3240622282028198,
      "learning_rate": 1.587194863666624e-05,
      "loss": 0.3681,
      "num_input_tokens_seen": 16405480,
      "step": 25075
    },
    {
      "epoch": 13.144654088050315,
      "grad_norm": 0.8243143558502197,
      "learning_rate": 1.5861304733786974e-05,
      "loss": 0.4815,
      "num_input_tokens_seen": 16408648,
      "step": 25080
    },
    {
      "epoch": 13.14727463312369,
      "grad_norm": 0.3291229009628296,
      "learning_rate": 1.5850662742618505e-05,
      "loss": 0.4186,
      "num_input_tokens_seen": 16411688,
      "step": 25085
    },
    {
      "epoch": 13.149895178197065,
      "grad_norm": 0.7103000283241272,
      "learning_rate": 1.5840022665387004e-05,
      "loss": 0.4727,
      "num_input_tokens_seen": 16414664,
      "step": 25090
    },
    {
      "epoch": 13.15251572327044,
      "grad_norm": 0.4064734876155853,
      "learning_rate": 1.5829384504318268e-05,
      "loss": 0.3865,
      "num_input_tokens_seen": 16417832,
      "step": 25095
    },
    {
      "epoch": 13.155136268343815,
      "grad_norm": 0.27653491497039795,
      "learning_rate": 1.5818748261637662e-05,
      "loss": 0.423,
      "num_input_tokens_seen": 16421672,
      "step": 25100
    },
    {
      "epoch": 13.15775681341719,
      "grad_norm": 0.8973031640052795,
      "learning_rate": 1.5808113939570173e-05,
      "loss": 0.4808,
      "num_input_tokens_seen": 16424520,
      "step": 25105
    },
    {
      "epoch": 13.160377358490566,
      "grad_norm": 0.7072997689247131,
      "learning_rate": 1.5797481540340386e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 16426760,
      "step": 25110
    },
    {
      "epoch": 13.16299790356394,
      "grad_norm": 0.718841552734375,
      "learning_rate": 1.5786851066172466e-05,
      "loss": 0.4366,
      "num_input_tokens_seen": 16430536,
      "step": 25115
    },
    {
      "epoch": 13.165618448637316,
      "grad_norm": 0.5225247144699097,
      "learning_rate": 1.5776222519290208e-05,
      "loss": 0.4204,
      "num_input_tokens_seen": 16433384,
      "step": 25120
    },
    {
      "epoch": 13.168238993710691,
      "grad_norm": 0.692002534866333,
      "learning_rate": 1.576559590191696e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 16436168,
      "step": 25125
    },
    {
      "epoch": 13.170859538784066,
      "grad_norm": 0.5065256953239441,
      "learning_rate": 1.5754971216275692e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 16438760,
      "step": 25130
    },
    {
      "epoch": 13.173480083857442,
      "grad_norm": 0.443899542093277,
      "learning_rate": 1.5744348464588975e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 16441864,
      "step": 25135
    },
    {
      "epoch": 13.176100628930818,
      "grad_norm": 0.49920907616615295,
      "learning_rate": 1.5733727649078968e-05,
      "loss": 0.4598,
      "num_input_tokens_seen": 16445512,
      "step": 25140
    },
    {
      "epoch": 13.178721174004194,
      "grad_norm": 0.4044252336025238,
      "learning_rate": 1.5723108771967414e-05,
      "loss": 0.4743,
      "num_input_tokens_seen": 16448264,
      "step": 25145
    },
    {
      "epoch": 13.181341719077569,
      "grad_norm": 0.7102372646331787,
      "learning_rate": 1.5712491835475685e-05,
      "loss": 0.509,
      "num_input_tokens_seen": 16451880,
      "step": 25150
    },
    {
      "epoch": 13.183962264150944,
      "grad_norm": 0.6045408248901367,
      "learning_rate": 1.5701876841824698e-05,
      "loss": 0.476,
      "num_input_tokens_seen": 16454728,
      "step": 25155
    },
    {
      "epoch": 13.18658280922432,
      "grad_norm": 0.4514174461364746,
      "learning_rate": 1.5691263793234993e-05,
      "loss": 0.3554,
      "num_input_tokens_seen": 16457480,
      "step": 25160
    },
    {
      "epoch": 13.189203354297694,
      "grad_norm": 0.41809961199760437,
      "learning_rate": 1.5680652691926706e-05,
      "loss": 0.5081,
      "num_input_tokens_seen": 16460232,
      "step": 25165
    },
    {
      "epoch": 13.19182389937107,
      "grad_norm": 0.5633764266967773,
      "learning_rate": 1.5670043540119553e-05,
      "loss": 0.3623,
      "num_input_tokens_seen": 16462856,
      "step": 25170
    },
    {
      "epoch": 13.194444444444445,
      "grad_norm": 0.6967121362686157,
      "learning_rate": 1.5659436340032855e-05,
      "loss": 0.3608,
      "num_input_tokens_seen": 16465640,
      "step": 25175
    },
    {
      "epoch": 13.19706498951782,
      "grad_norm": 0.37218406796455383,
      "learning_rate": 1.5648831093885515e-05,
      "loss": 0.3688,
      "num_input_tokens_seen": 16468552,
      "step": 25180
    },
    {
      "epoch": 13.199685534591195,
      "grad_norm": 0.5736125707626343,
      "learning_rate": 1.5638227803896015e-05,
      "loss": 0.4055,
      "num_input_tokens_seen": 16471176,
      "step": 25185
    },
    {
      "epoch": 13.20230607966457,
      "grad_norm": 0.3488158583641052,
      "learning_rate": 1.562762647228246e-05,
      "loss": 0.3988,
      "num_input_tokens_seen": 16473896,
      "step": 25190
    },
    {
      "epoch": 13.204926624737945,
      "grad_norm": 0.18218496441841125,
      "learning_rate": 1.5617027101262515e-05,
      "loss": 0.3863,
      "num_input_tokens_seen": 16478600,
      "step": 25195
    },
    {
      "epoch": 13.20754716981132,
      "grad_norm": 0.8144711852073669,
      "learning_rate": 1.5606429693053463e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 16482152,
      "step": 25200
    },
    {
      "epoch": 13.210167714884696,
      "grad_norm": 0.4478156268596649,
      "learning_rate": 1.5595834249872145e-05,
      "loss": 0.4874,
      "num_input_tokens_seen": 16485128,
      "step": 25205
    },
    {
      "epoch": 13.21278825995807,
      "grad_norm": 0.7092603445053101,
      "learning_rate": 1.558524077393502e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 16488552,
      "step": 25210
    },
    {
      "epoch": 13.215408805031446,
      "grad_norm": 0.5605754256248474,
      "learning_rate": 1.557464926745811e-05,
      "loss": 0.3748,
      "num_input_tokens_seen": 16490984,
      "step": 25215
    },
    {
      "epoch": 13.218029350104821,
      "grad_norm": 0.744796872138977,
      "learning_rate": 1.5564059732657045e-05,
      "loss": 0.493,
      "num_input_tokens_seen": 16493928,
      "step": 25220
    },
    {
      "epoch": 13.220649895178196,
      "grad_norm": 0.4871434271335602,
      "learning_rate": 1.555347217174703e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 16496904,
      "step": 25225
    },
    {
      "epoch": 13.223270440251572,
      "grad_norm": 0.5630735754966736,
      "learning_rate": 1.554288658694287e-05,
      "loss": 0.3628,
      "num_input_tokens_seen": 16499464,
      "step": 25230
    },
    {
      "epoch": 13.225890985324948,
      "grad_norm": 0.43511763215065,
      "learning_rate": 1.553230298045895e-05,
      "loss": 0.4072,
      "num_input_tokens_seen": 16502376,
      "step": 25235
    },
    {
      "epoch": 13.228511530398324,
      "grad_norm": 0.39565184712409973,
      "learning_rate": 1.5521721354509223e-05,
      "loss": 0.4254,
      "num_input_tokens_seen": 16505288,
      "step": 25240
    },
    {
      "epoch": 13.231132075471699,
      "grad_norm": 0.6013386845588684,
      "learning_rate": 1.5511141711307258e-05,
      "loss": 0.5884,
      "num_input_tokens_seen": 16507912,
      "step": 25245
    },
    {
      "epoch": 13.233752620545074,
      "grad_norm": 0.44299110770225525,
      "learning_rate": 1.5500564053066184e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 16510824,
      "step": 25250
    },
    {
      "epoch": 13.23637316561845,
      "grad_norm": 0.7518419027328491,
      "learning_rate": 1.5489988381998743e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 16514408,
      "step": 25255
    },
    {
      "epoch": 13.238993710691824,
      "grad_norm": 0.6211830973625183,
      "learning_rate": 1.5479414700317237e-05,
      "loss": 0.3739,
      "num_input_tokens_seen": 16517480,
      "step": 25260
    },
    {
      "epoch": 13.2416142557652,
      "grad_norm": 0.4251166880130768,
      "learning_rate": 1.546884301023356e-05,
      "loss": 0.3669,
      "num_input_tokens_seen": 16520968,
      "step": 25265
    },
    {
      "epoch": 13.244234800838575,
      "grad_norm": 0.39407211542129517,
      "learning_rate": 1.5458273313959186e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 16523976,
      "step": 25270
    },
    {
      "epoch": 13.24685534591195,
      "grad_norm": 0.5496869683265686,
      "learning_rate": 1.5447705613705177e-05,
      "loss": 0.4168,
      "num_input_tokens_seen": 16527336,
      "step": 25275
    },
    {
      "epoch": 13.249475890985325,
      "grad_norm": 0.36426201462745667,
      "learning_rate": 1.5437139911682176e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 16531368,
      "step": 25280
    },
    {
      "epoch": 13.2520964360587,
      "grad_norm": 0.2793086767196655,
      "learning_rate": 1.54265762101004e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 16535016,
      "step": 25285
    },
    {
      "epoch": 13.254716981132075,
      "grad_norm": 0.5176423192024231,
      "learning_rate": 1.5416014511169663e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 16539528,
      "step": 25290
    },
    {
      "epoch": 13.25733752620545,
      "grad_norm": 0.7343547940254211,
      "learning_rate": 1.5405454817099363e-05,
      "loss": 0.4552,
      "num_input_tokens_seen": 16543752,
      "step": 25295
    },
    {
      "epoch": 13.259958071278826,
      "grad_norm": 0.3515257239341736,
      "learning_rate": 1.5394897130098446e-05,
      "loss": 0.5687,
      "num_input_tokens_seen": 16548040,
      "step": 25300
    },
    {
      "epoch": 13.2625786163522,
      "grad_norm": 0.6173111796379089,
      "learning_rate": 1.5384341452375465e-05,
      "loss": 0.3449,
      "num_input_tokens_seen": 16551496,
      "step": 25305
    },
    {
      "epoch": 13.265199161425576,
      "grad_norm": 0.9434686899185181,
      "learning_rate": 1.5373787786138556e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 16554984,
      "step": 25310
    },
    {
      "epoch": 13.267819706498951,
      "grad_norm": 0.419027179479599,
      "learning_rate": 1.5363236133595416e-05,
      "loss": 0.4403,
      "num_input_tokens_seen": 16558888,
      "step": 25315
    },
    {
      "epoch": 13.270440251572326,
      "grad_norm": 0.4161315858364105,
      "learning_rate": 1.5352686496953344e-05,
      "loss": 0.3598,
      "num_input_tokens_seen": 16562216,
      "step": 25320
    },
    {
      "epoch": 13.273060796645701,
      "grad_norm": 0.48528364300727844,
      "learning_rate": 1.534213887841919e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 16565320,
      "step": 25325
    },
    {
      "epoch": 13.275681341719078,
      "grad_norm": 0.46942269802093506,
      "learning_rate": 1.5331593280199395e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 16568072,
      "step": 25330
    },
    {
      "epoch": 13.278301886792454,
      "grad_norm": 0.40109187364578247,
      "learning_rate": 1.532104970449999e-05,
      "loss": 0.4017,
      "num_input_tokens_seen": 16571912,
      "step": 25335
    },
    {
      "epoch": 13.280922431865829,
      "grad_norm": 0.5707410573959351,
      "learning_rate": 1.5310508153526555e-05,
      "loss": 0.3464,
      "num_input_tokens_seen": 16577384,
      "step": 25340
    },
    {
      "epoch": 13.283542976939204,
      "grad_norm": 0.6310285925865173,
      "learning_rate": 1.5299968629484276e-05,
      "loss": 0.5748,
      "num_input_tokens_seen": 16581416,
      "step": 25345
    },
    {
      "epoch": 13.286163522012579,
      "grad_norm": 0.38296496868133545,
      "learning_rate": 1.5289431134577904e-05,
      "loss": 0.341,
      "num_input_tokens_seen": 16583944,
      "step": 25350
    },
    {
      "epoch": 13.288784067085954,
      "grad_norm": 0.3152932822704315,
      "learning_rate": 1.527889567101174e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 16586888,
      "step": 25355
    },
    {
      "epoch": 13.29140461215933,
      "grad_norm": 0.3513353168964386,
      "learning_rate": 1.5268362240989707e-05,
      "loss": 0.4161,
      "num_input_tokens_seen": 16590376,
      "step": 25360
    },
    {
      "epoch": 13.294025157232705,
      "grad_norm": 0.39526671171188354,
      "learning_rate": 1.525783084671526e-05,
      "loss": 0.3684,
      "num_input_tokens_seen": 16593672,
      "step": 25365
    },
    {
      "epoch": 13.29664570230608,
      "grad_norm": 0.7495171427726746,
      "learning_rate": 1.5247301490391458e-05,
      "loss": 0.5191,
      "num_input_tokens_seen": 16596840,
      "step": 25370
    },
    {
      "epoch": 13.299266247379455,
      "grad_norm": 0.7817217111587524,
      "learning_rate": 1.5236774174220922e-05,
      "loss": 0.5188,
      "num_input_tokens_seen": 16599880,
      "step": 25375
    },
    {
      "epoch": 13.30188679245283,
      "grad_norm": 0.4958135485649109,
      "learning_rate": 1.5226248900405852e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 16602856,
      "step": 25380
    },
    {
      "epoch": 13.304507337526205,
      "grad_norm": 0.511715829372406,
      "learning_rate": 1.5215725671147996e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 16606440,
      "step": 25385
    },
    {
      "epoch": 13.30712788259958,
      "grad_norm": 0.40997618436813354,
      "learning_rate": 1.5205204488648709e-05,
      "loss": 0.4226,
      "num_input_tokens_seen": 16610248,
      "step": 25390
    },
    {
      "epoch": 13.309748427672956,
      "grad_norm": 0.6308817267417908,
      "learning_rate": 1.5194685355108893e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 16612584,
      "step": 25395
    },
    {
      "epoch": 13.31236897274633,
      "grad_norm": 0.2707509398460388,
      "learning_rate": 1.518416827272904e-05,
      "loss": 0.4188,
      "num_input_tokens_seen": 16616168,
      "step": 25400
    },
    {
      "epoch": 13.314989517819706,
      "grad_norm": 0.4929789900779724,
      "learning_rate": 1.5173653243709207e-05,
      "loss": 0.4733,
      "num_input_tokens_seen": 16618984,
      "step": 25405
    },
    {
      "epoch": 13.317610062893081,
      "grad_norm": 0.455685555934906,
      "learning_rate": 1.5163140270249e-05,
      "loss": 0.4916,
      "num_input_tokens_seen": 16623208,
      "step": 25410
    },
    {
      "epoch": 13.320230607966456,
      "grad_norm": 0.2839718759059906,
      "learning_rate": 1.5152629354547626e-05,
      "loss": 0.3962,
      "num_input_tokens_seen": 16627656,
      "step": 25415
    },
    {
      "epoch": 13.322851153039831,
      "grad_norm": 0.5915322303771973,
      "learning_rate": 1.5142120498803842e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 16630280,
      "step": 25420
    },
    {
      "epoch": 13.325471698113208,
      "grad_norm": 0.4386894106864929,
      "learning_rate": 1.5131613705215991e-05,
      "loss": 0.4114,
      "num_input_tokens_seen": 16632904,
      "step": 25425
    },
    {
      "epoch": 13.328092243186584,
      "grad_norm": 0.3954704999923706,
      "learning_rate": 1.5121108975981965e-05,
      "loss": 0.5029,
      "num_input_tokens_seen": 16635624,
      "step": 25430
    },
    {
      "epoch": 13.330712788259959,
      "grad_norm": 0.6894662976264954,
      "learning_rate": 1.5110606313299248e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 16638760,
      "step": 25435
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 0.6752972602844238,
      "learning_rate": 1.5100105719364855e-05,
      "loss": 0.5349,
      "num_input_tokens_seen": 16641448,
      "step": 25440
    },
    {
      "epoch": 13.335953878406709,
      "grad_norm": 0.5632578730583191,
      "learning_rate": 1.50896071963754e-05,
      "loss": 0.3764,
      "num_input_tokens_seen": 16644520,
      "step": 25445
    },
    {
      "epoch": 13.338574423480084,
      "grad_norm": 0.30723994970321655,
      "learning_rate": 1.5079110746527055e-05,
      "loss": 0.448,
      "num_input_tokens_seen": 16648584,
      "step": 25450
    },
    {
      "epoch": 13.34119496855346,
      "grad_norm": 0.4245336949825287,
      "learning_rate": 1.5068616372015565e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 16651784,
      "step": 25455
    },
    {
      "epoch": 13.343815513626835,
      "grad_norm": 0.46763667464256287,
      "learning_rate": 1.5058124075036223e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 16654024,
      "step": 25460
    },
    {
      "epoch": 13.34643605870021,
      "grad_norm": 0.4407747983932495,
      "learning_rate": 1.5047633857783908e-05,
      "loss": 0.4486,
      "num_input_tokens_seen": 16657480,
      "step": 25465
    },
    {
      "epoch": 13.349056603773585,
      "grad_norm": 0.7259566783905029,
      "learning_rate": 1.5037145722453044e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 16661096,
      "step": 25470
    },
    {
      "epoch": 13.35167714884696,
      "grad_norm": 0.43012893199920654,
      "learning_rate": 1.5026659671237624e-05,
      "loss": 0.3531,
      "num_input_tokens_seen": 16663688,
      "step": 25475
    },
    {
      "epoch": 13.354297693920335,
      "grad_norm": 0.6830061674118042,
      "learning_rate": 1.501617570633123e-05,
      "loss": 0.4579,
      "num_input_tokens_seen": 16666024,
      "step": 25480
    },
    {
      "epoch": 13.35691823899371,
      "grad_norm": 0.33607491850852966,
      "learning_rate": 1.500569382992697e-05,
      "loss": 0.4364,
      "num_input_tokens_seen": 16669608,
      "step": 25485
    },
    {
      "epoch": 13.359538784067086,
      "grad_norm": 0.6671941876411438,
      "learning_rate": 1.4995214044217547e-05,
      "loss": 0.5022,
      "num_input_tokens_seen": 16673480,
      "step": 25490
    },
    {
      "epoch": 13.36215932914046,
      "grad_norm": 0.43270793557167053,
      "learning_rate": 1.49847363513952e-05,
      "loss": 0.3138,
      "num_input_tokens_seen": 16677704,
      "step": 25495
    },
    {
      "epoch": 13.364779874213836,
      "grad_norm": 0.5961599349975586,
      "learning_rate": 1.4974260753651748e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 16680744,
      "step": 25500
    },
    {
      "epoch": 13.367400419287211,
      "grad_norm": 1.1248514652252197,
      "learning_rate": 1.4963787253178568e-05,
      "loss": 0.496,
      "num_input_tokens_seen": 16683048,
      "step": 25505
    },
    {
      "epoch": 13.370020964360586,
      "grad_norm": 0.5965104699134827,
      "learning_rate": 1.4953315852166594e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 16685832,
      "step": 25510
    },
    {
      "epoch": 13.372641509433961,
      "grad_norm": 0.3725369870662689,
      "learning_rate": 1.4942846552806322e-05,
      "loss": 0.4855,
      "num_input_tokens_seen": 16688168,
      "step": 25515
    },
    {
      "epoch": 13.375262054507338,
      "grad_norm": 0.4111960828304291,
      "learning_rate": 1.4932379357287829e-05,
      "loss": 0.4274,
      "num_input_tokens_seen": 16690888,
      "step": 25520
    },
    {
      "epoch": 13.377882599580714,
      "grad_norm": 0.4152529537677765,
      "learning_rate": 1.49219142678007e-05,
      "loss": 0.4401,
      "num_input_tokens_seen": 16694280,
      "step": 25525
    },
    {
      "epoch": 13.380503144654089,
      "grad_norm": 0.45011186599731445,
      "learning_rate": 1.4911451286534128e-05,
      "loss": 0.4799,
      "num_input_tokens_seen": 16697128,
      "step": 25530
    },
    {
      "epoch": 13.383123689727464,
      "grad_norm": 0.32562172412872314,
      "learning_rate": 1.490099041567686e-05,
      "loss": 0.3923,
      "num_input_tokens_seen": 16701192,
      "step": 25535
    },
    {
      "epoch": 13.385744234800839,
      "grad_norm": 0.4718361794948578,
      "learning_rate": 1.4890531657417176e-05,
      "loss": 0.4029,
      "num_input_tokens_seen": 16706280,
      "step": 25540
    },
    {
      "epoch": 13.388364779874214,
      "grad_norm": 0.698053240776062,
      "learning_rate": 1.4880075013942938e-05,
      "loss": 0.3469,
      "num_input_tokens_seen": 16709448,
      "step": 25545
    },
    {
      "epoch": 13.39098532494759,
      "grad_norm": 0.2945379912853241,
      "learning_rate": 1.4869620487441558e-05,
      "loss": 0.5784,
      "num_input_tokens_seen": 16713256,
      "step": 25550
    },
    {
      "epoch": 13.393605870020965,
      "grad_norm": 0.5597636103630066,
      "learning_rate": 1.4859168080099988e-05,
      "loss": 0.4864,
      "num_input_tokens_seen": 16716296,
      "step": 25555
    },
    {
      "epoch": 13.39622641509434,
      "grad_norm": 0.590401828289032,
      "learning_rate": 1.484871779410477e-05,
      "loss": 0.4466,
      "num_input_tokens_seen": 16719432,
      "step": 25560
    },
    {
      "epoch": 13.398846960167715,
      "grad_norm": 0.48308804631233215,
      "learning_rate": 1.4838269631641971e-05,
      "loss": 0.3457,
      "num_input_tokens_seen": 16722536,
      "step": 25565
    },
    {
      "epoch": 13.40146750524109,
      "grad_norm": 0.322635680437088,
      "learning_rate": 1.482782359489724e-05,
      "loss": 0.5566,
      "num_input_tokens_seen": 16725704,
      "step": 25570
    },
    {
      "epoch": 13.404088050314465,
      "grad_norm": 0.5920621156692505,
      "learning_rate": 1.4817379686055754e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 16729192,
      "step": 25575
    },
    {
      "epoch": 13.40670859538784,
      "grad_norm": 0.39749830961227417,
      "learning_rate": 1.4806937907302282e-05,
      "loss": 0.4104,
      "num_input_tokens_seen": 16732168,
      "step": 25580
    },
    {
      "epoch": 13.409329140461216,
      "grad_norm": 0.5202305316925049,
      "learning_rate": 1.4796498260821103e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 16735496,
      "step": 25585
    },
    {
      "epoch": 13.41194968553459,
      "grad_norm": 0.4424927234649658,
      "learning_rate": 1.4786060748796076e-05,
      "loss": 0.5142,
      "num_input_tokens_seen": 16739176,
      "step": 25590
    },
    {
      "epoch": 13.414570230607966,
      "grad_norm": 0.4735943675041199,
      "learning_rate": 1.4775625373410617e-05,
      "loss": 0.3842,
      "num_input_tokens_seen": 16742344,
      "step": 25595
    },
    {
      "epoch": 13.417190775681341,
      "grad_norm": 0.3730877935886383,
      "learning_rate": 1.4765192136847685e-05,
      "loss": 0.3596,
      "num_input_tokens_seen": 16745512,
      "step": 25600
    },
    {
      "epoch": 13.419811320754716,
      "grad_norm": 0.4831167161464691,
      "learning_rate": 1.47547610412898e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 16748584,
      "step": 25605
    },
    {
      "epoch": 13.422431865828091,
      "grad_norm": 0.26919540762901306,
      "learning_rate": 1.4744332088919011e-05,
      "loss": 0.4605,
      "num_input_tokens_seen": 16752168,
      "step": 25610
    },
    {
      "epoch": 13.425052410901468,
      "grad_norm": 0.2786962687969208,
      "learning_rate": 1.4733905281916954e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 16755688,
      "step": 25615
    },
    {
      "epoch": 13.427672955974844,
      "grad_norm": 0.31974172592163086,
      "learning_rate": 1.472348062246478e-05,
      "loss": 0.3555,
      "num_input_tokens_seen": 16759016,
      "step": 25620
    },
    {
      "epoch": 13.430293501048219,
      "grad_norm": 0.7029229402542114,
      "learning_rate": 1.4713058112743228e-05,
      "loss": 0.5341,
      "num_input_tokens_seen": 16762600,
      "step": 25625
    },
    {
      "epoch": 13.432914046121594,
      "grad_norm": 0.7426270842552185,
      "learning_rate": 1.4702637754932553e-05,
      "loss": 0.4674,
      "num_input_tokens_seen": 16765768,
      "step": 25630
    },
    {
      "epoch": 13.435534591194969,
      "grad_norm": 0.6165300607681274,
      "learning_rate": 1.4692219551212587e-05,
      "loss": 0.5094,
      "num_input_tokens_seen": 16768584,
      "step": 25635
    },
    {
      "epoch": 13.438155136268344,
      "grad_norm": 0.6584640741348267,
      "learning_rate": 1.4681803503762695e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 16771240,
      "step": 25640
    },
    {
      "epoch": 13.44077568134172,
      "grad_norm": 0.5009941458702087,
      "learning_rate": 1.4671389614761788e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 16774792,
      "step": 25645
    },
    {
      "epoch": 13.443396226415095,
      "grad_norm": 0.40374910831451416,
      "learning_rate": 1.4660977886388339e-05,
      "loss": 0.5691,
      "num_input_tokens_seen": 16777864,
      "step": 25650
    },
    {
      "epoch": 13.44601677148847,
      "grad_norm": 0.2790077030658722,
      "learning_rate": 1.4650568320820362e-05,
      "loss": 0.3571,
      "num_input_tokens_seen": 16781768,
      "step": 25655
    },
    {
      "epoch": 13.448637316561845,
      "grad_norm": 0.5264816880226135,
      "learning_rate": 1.4640160920235423e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 16784648,
      "step": 25660
    },
    {
      "epoch": 13.45125786163522,
      "grad_norm": 0.4052306115627289,
      "learning_rate": 1.4629755686810631e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 16787368,
      "step": 25665
    },
    {
      "epoch": 13.453878406708595,
      "grad_norm": 0.4506368637084961,
      "learning_rate": 1.4619352622722632e-05,
      "loss": 0.454,
      "num_input_tokens_seen": 16791048,
      "step": 25670
    },
    {
      "epoch": 13.45649895178197,
      "grad_norm": 0.5400939583778381,
      "learning_rate": 1.4608951730147635e-05,
      "loss": 0.4835,
      "num_input_tokens_seen": 16794504,
      "step": 25675
    },
    {
      "epoch": 13.459119496855346,
      "grad_norm": 0.3793764114379883,
      "learning_rate": 1.4598553011261395e-05,
      "loss": 0.3378,
      "num_input_tokens_seen": 16797960,
      "step": 25680
    },
    {
      "epoch": 13.46174004192872,
      "grad_norm": 0.48441988229751587,
      "learning_rate": 1.4588156468239192e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 16800456,
      "step": 25685
    },
    {
      "epoch": 13.464360587002096,
      "grad_norm": 0.8067144155502319,
      "learning_rate": 1.4577762103255882e-05,
      "loss": 0.4384,
      "num_input_tokens_seen": 16803048,
      "step": 25690
    },
    {
      "epoch": 13.466981132075471,
      "grad_norm": 0.568499743938446,
      "learning_rate": 1.4567369918485828e-05,
      "loss": 0.5684,
      "num_input_tokens_seen": 16807272,
      "step": 25695
    },
    {
      "epoch": 13.469601677148846,
      "grad_norm": 0.3483860194683075,
      "learning_rate": 1.4556979916102976e-05,
      "loss": 0.3116,
      "num_input_tokens_seen": 16810792,
      "step": 25700
    },
    {
      "epoch": 13.472222222222221,
      "grad_norm": 0.4807281196117401,
      "learning_rate": 1.454659209828077e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 16814024,
      "step": 25705
    },
    {
      "epoch": 13.474842767295598,
      "grad_norm": 0.2803876996040344,
      "learning_rate": 1.4536206467192243e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 16817928,
      "step": 25710
    },
    {
      "epoch": 13.477463312368974,
      "grad_norm": 0.3596050441265106,
      "learning_rate": 1.4525823025009944e-05,
      "loss": 0.3574,
      "num_input_tokens_seen": 16820680,
      "step": 25715
    },
    {
      "epoch": 13.480083857442349,
      "grad_norm": 0.318645179271698,
      "learning_rate": 1.4515441773905983e-05,
      "loss": 0.3964,
      "num_input_tokens_seen": 16824328,
      "step": 25720
    },
    {
      "epoch": 13.482704402515724,
      "grad_norm": 0.8907736539840698,
      "learning_rate": 1.4505062716051988e-05,
      "loss": 0.4447,
      "num_input_tokens_seen": 16827880,
      "step": 25725
    },
    {
      "epoch": 13.485324947589099,
      "grad_norm": 0.6601956486701965,
      "learning_rate": 1.4494685853619133e-05,
      "loss": 0.6404,
      "num_input_tokens_seen": 16830984,
      "step": 25730
    },
    {
      "epoch": 13.487945492662474,
      "grad_norm": 0.4729340970516205,
      "learning_rate": 1.4484311188778143e-05,
      "loss": 0.5829,
      "num_input_tokens_seen": 16835144,
      "step": 25735
    },
    {
      "epoch": 13.49056603773585,
      "grad_norm": 0.6348549127578735,
      "learning_rate": 1.4473938723699285e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 16838152,
      "step": 25740
    },
    {
      "epoch": 13.493186582809225,
      "grad_norm": 0.4768712520599365,
      "learning_rate": 1.4463568460552368e-05,
      "loss": 0.3422,
      "num_input_tokens_seen": 16841480,
      "step": 25745
    },
    {
      "epoch": 13.4958071278826,
      "grad_norm": 0.31471654772758484,
      "learning_rate": 1.4453200401506723e-05,
      "loss": 0.4006,
      "num_input_tokens_seen": 16845288,
      "step": 25750
    },
    {
      "epoch": 13.498427672955975,
      "grad_norm": 0.5696283578872681,
      "learning_rate": 1.4442834548731218e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 16849096,
      "step": 25755
    },
    {
      "epoch": 13.50104821802935,
      "grad_norm": 0.4850475788116455,
      "learning_rate": 1.4432470904394285e-05,
      "loss": 0.4985,
      "num_input_tokens_seen": 16852232,
      "step": 25760
    },
    {
      "epoch": 13.503668763102725,
      "grad_norm": 0.7063138484954834,
      "learning_rate": 1.4422109470663878e-05,
      "loss": 0.5915,
      "num_input_tokens_seen": 16855176,
      "step": 25765
    },
    {
      "epoch": 13.5062893081761,
      "grad_norm": 0.40975895524024963,
      "learning_rate": 1.441175024970749e-05,
      "loss": 0.3231,
      "num_input_tokens_seen": 16858344,
      "step": 25770
    },
    {
      "epoch": 13.508909853249476,
      "grad_norm": 0.46197712421417236,
      "learning_rate": 1.4401393243692163e-05,
      "loss": 0.5451,
      "num_input_tokens_seen": 16861896,
      "step": 25775
    },
    {
      "epoch": 13.51153039832285,
      "grad_norm": 0.4416011869907379,
      "learning_rate": 1.4391038454784455e-05,
      "loss": 0.4489,
      "num_input_tokens_seen": 16865704,
      "step": 25780
    },
    {
      "epoch": 13.514150943396226,
      "grad_norm": 0.5414552092552185,
      "learning_rate": 1.4380685885150458e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 16868616,
      "step": 25785
    },
    {
      "epoch": 13.516771488469601,
      "grad_norm": 0.2718217074871063,
      "learning_rate": 1.437033553695582e-05,
      "loss": 0.3789,
      "num_input_tokens_seen": 16871816,
      "step": 25790
    },
    {
      "epoch": 13.519392033542976,
      "grad_norm": 0.6995325684547424,
      "learning_rate": 1.4359987412365721e-05,
      "loss": 0.4193,
      "num_input_tokens_seen": 16875016,
      "step": 25795
    },
    {
      "epoch": 13.522012578616351,
      "grad_norm": 0.564261794090271,
      "learning_rate": 1.4349641513544876e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 16877928,
      "step": 25800
    },
    {
      "epoch": 13.524633123689728,
      "grad_norm": 0.42205026745796204,
      "learning_rate": 1.4339297842657523e-05,
      "loss": 0.3765,
      "num_input_tokens_seen": 16881896,
      "step": 25805
    },
    {
      "epoch": 13.527253668763104,
      "grad_norm": 0.7453575134277344,
      "learning_rate": 1.4328956401867427e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 16884744,
      "step": 25810
    },
    {
      "epoch": 13.529874213836479,
      "grad_norm": 0.506356954574585,
      "learning_rate": 1.4318617193337908e-05,
      "loss": 0.2378,
      "num_input_tokens_seen": 16890376,
      "step": 25815
    },
    {
      "epoch": 13.532494758909854,
      "grad_norm": 0.41748499870300293,
      "learning_rate": 1.4308280219231812e-05,
      "loss": 0.3751,
      "num_input_tokens_seen": 16892840,
      "step": 25820
    },
    {
      "epoch": 13.535115303983229,
      "grad_norm": 0.583160400390625,
      "learning_rate": 1.4297945481711527e-05,
      "loss": 0.5634,
      "num_input_tokens_seen": 16896552,
      "step": 25825
    },
    {
      "epoch": 13.537735849056604,
      "grad_norm": 1.1989024877548218,
      "learning_rate": 1.4287612982938946e-05,
      "loss": 0.4275,
      "num_input_tokens_seen": 16899112,
      "step": 25830
    },
    {
      "epoch": 13.54035639412998,
      "grad_norm": 1.0241411924362183,
      "learning_rate": 1.4277282725075523e-05,
      "loss": 0.2662,
      "num_input_tokens_seen": 16902536,
      "step": 25835
    },
    {
      "epoch": 13.542976939203355,
      "grad_norm": 0.41933637857437134,
      "learning_rate": 1.426695471028221e-05,
      "loss": 0.4955,
      "num_input_tokens_seen": 16905576,
      "step": 25840
    },
    {
      "epoch": 13.54559748427673,
      "grad_norm": 0.4146784842014313,
      "learning_rate": 1.4256628940719524e-05,
      "loss": 0.5813,
      "num_input_tokens_seen": 16909128,
      "step": 25845
    },
    {
      "epoch": 13.548218029350105,
      "grad_norm": 0.6665096879005432,
      "learning_rate": 1.42463054185475e-05,
      "loss": 0.4105,
      "num_input_tokens_seen": 16911592,
      "step": 25850
    },
    {
      "epoch": 13.55083857442348,
      "grad_norm": 1.0300588607788086,
      "learning_rate": 1.4235984145925705e-05,
      "loss": 0.5276,
      "num_input_tokens_seen": 16914536,
      "step": 25855
    },
    {
      "epoch": 13.553459119496855,
      "grad_norm": 0.44758138060569763,
      "learning_rate": 1.4225665125013224e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 16920936,
      "step": 25860
    },
    {
      "epoch": 13.55607966457023,
      "grad_norm": 0.637872576713562,
      "learning_rate": 1.4215348357968669e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 16923464,
      "step": 25865
    },
    {
      "epoch": 13.558700209643606,
      "grad_norm": 0.38643112778663635,
      "learning_rate": 1.42050338469502e-05,
      "loss": 0.4518,
      "num_input_tokens_seen": 16926152,
      "step": 25870
    },
    {
      "epoch": 13.56132075471698,
      "grad_norm": 0.461417555809021,
      "learning_rate": 1.4194721594115495e-05,
      "loss": 0.3996,
      "num_input_tokens_seen": 16929928,
      "step": 25875
    },
    {
      "epoch": 13.563941299790356,
      "grad_norm": 0.39947813749313354,
      "learning_rate": 1.4184411601621769e-05,
      "loss": 0.5106,
      "num_input_tokens_seen": 16933864,
      "step": 25880
    },
    {
      "epoch": 13.566561844863731,
      "grad_norm": 0.8011016249656677,
      "learning_rate": 1.4174103871625738e-05,
      "loss": 0.3728,
      "num_input_tokens_seen": 16936392,
      "step": 25885
    },
    {
      "epoch": 13.569182389937106,
      "grad_norm": 0.2678191363811493,
      "learning_rate": 1.4163798406283674e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 16940008,
      "step": 25890
    },
    {
      "epoch": 13.571802935010481,
      "grad_norm": 0.34594252705574036,
      "learning_rate": 1.415349520775135e-05,
      "loss": 0.4319,
      "num_input_tokens_seen": 16943176,
      "step": 25895
    },
    {
      "epoch": 13.574423480083858,
      "grad_norm": 0.6508281230926514,
      "learning_rate": 1.4143194278184086e-05,
      "loss": 0.5434,
      "num_input_tokens_seen": 16945992,
      "step": 25900
    },
    {
      "epoch": 13.577044025157234,
      "grad_norm": 0.36637821793556213,
      "learning_rate": 1.4132895619736724e-05,
      "loss": 0.6014,
      "num_input_tokens_seen": 16949352,
      "step": 25905
    },
    {
      "epoch": 13.579664570230609,
      "grad_norm": 0.36943280696868896,
      "learning_rate": 1.4122599234563613e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 16953000,
      "step": 25910
    },
    {
      "epoch": 13.582285115303984,
      "grad_norm": 0.3115813732147217,
      "learning_rate": 1.4112305124818647e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 16956232,
      "step": 25915
    },
    {
      "epoch": 13.584905660377359,
      "grad_norm": 0.44875314831733704,
      "learning_rate": 1.4102013292655242e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 16959464,
      "step": 25920
    },
    {
      "epoch": 13.587526205450734,
      "grad_norm": 0.6245035529136658,
      "learning_rate": 1.409172374022632e-05,
      "loss": 0.4739,
      "num_input_tokens_seen": 16962440,
      "step": 25925
    },
    {
      "epoch": 13.59014675052411,
      "grad_norm": 0.36621004343032837,
      "learning_rate": 1.4081436469684339e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 16965608,
      "step": 25930
    },
    {
      "epoch": 13.592767295597485,
      "grad_norm": 0.5098950862884521,
      "learning_rate": 1.4071151483181294e-05,
      "loss": 0.4461,
      "num_input_tokens_seen": 16969864,
      "step": 25935
    },
    {
      "epoch": 13.59538784067086,
      "grad_norm": 0.6810804605484009,
      "learning_rate": 1.4060868782868669e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 16972840,
      "step": 25940
    },
    {
      "epoch": 13.598008385744235,
      "grad_norm": 0.5004249811172485,
      "learning_rate": 1.405058837089749e-05,
      "loss": 0.5163,
      "num_input_tokens_seen": 16975816,
      "step": 25945
    },
    {
      "epoch": 13.60062893081761,
      "grad_norm": 0.42281097173690796,
      "learning_rate": 1.404031024941832e-05,
      "loss": 0.3788,
      "num_input_tokens_seen": 16979016,
      "step": 25950
    },
    {
      "epoch": 13.603249475890985,
      "grad_norm": 0.39761415123939514,
      "learning_rate": 1.40300344205812e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 16982056,
      "step": 25955
    },
    {
      "epoch": 13.60587002096436,
      "grad_norm": 0.6483640074729919,
      "learning_rate": 1.4019760886535738e-05,
      "loss": 0.3566,
      "num_input_tokens_seen": 16985032,
      "step": 25960
    },
    {
      "epoch": 13.608490566037736,
      "grad_norm": 0.44602230191230774,
      "learning_rate": 1.4009489649431018e-05,
      "loss": 0.4964,
      "num_input_tokens_seen": 16988008,
      "step": 25965
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 0.40641582012176514,
      "learning_rate": 1.3999220711415675e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 16990696,
      "step": 25970
    },
    {
      "epoch": 13.613731656184486,
      "grad_norm": 0.31530067324638367,
      "learning_rate": 1.3988954074637867e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 16994088,
      "step": 25975
    },
    {
      "epoch": 13.616352201257861,
      "grad_norm": 0.4338202476501465,
      "learning_rate": 1.3978689741245238e-05,
      "loss": 0.4763,
      "num_input_tokens_seen": 16997000,
      "step": 25980
    },
    {
      "epoch": 13.618972746331236,
      "grad_norm": 0.47827741503715515,
      "learning_rate": 1.3968427713384986e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 16999912,
      "step": 25985
    },
    {
      "epoch": 13.621593291404611,
      "grad_norm": 0.5899587869644165,
      "learning_rate": 1.3958167993203791e-05,
      "loss": 0.4798,
      "num_input_tokens_seen": 17003368,
      "step": 25990
    },
    {
      "epoch": 13.624213836477988,
      "grad_norm": 0.2847682535648346,
      "learning_rate": 1.3947910582847879e-05,
      "loss": 0.4219,
      "num_input_tokens_seen": 17007336,
      "step": 25995
    },
    {
      "epoch": 13.626834381551364,
      "grad_norm": 0.6497418880462646,
      "learning_rate": 1.3937655484462985e-05,
      "loss": 0.3852,
      "num_input_tokens_seen": 17010056,
      "step": 26000
    },
    {
      "epoch": 13.629454926624739,
      "grad_norm": 0.5200483798980713,
      "learning_rate": 1.392740270019437e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 17012776,
      "step": 26005
    },
    {
      "epoch": 13.632075471698114,
      "grad_norm": 0.426412433385849,
      "learning_rate": 1.3917152232186776e-05,
      "loss": 0.398,
      "num_input_tokens_seen": 17016872,
      "step": 26010
    },
    {
      "epoch": 13.634696016771489,
      "grad_norm": 0.5402864217758179,
      "learning_rate": 1.3906904082584504e-05,
      "loss": 0.5723,
      "num_input_tokens_seen": 17020680,
      "step": 26015
    },
    {
      "epoch": 13.637316561844864,
      "grad_norm": 0.3932470977306366,
      "learning_rate": 1.3896658253531334e-05,
      "loss": 0.4336,
      "num_input_tokens_seen": 17024168,
      "step": 26020
    },
    {
      "epoch": 13.63993710691824,
      "grad_norm": 0.4103757441043854,
      "learning_rate": 1.3886414747170581e-05,
      "loss": 0.503,
      "num_input_tokens_seen": 17027304,
      "step": 26025
    },
    {
      "epoch": 13.642557651991615,
      "grad_norm": 0.3391498625278473,
      "learning_rate": 1.3876173565645072e-05,
      "loss": 0.3769,
      "num_input_tokens_seen": 17030344,
      "step": 26030
    },
    {
      "epoch": 13.64517819706499,
      "grad_norm": 0.31194567680358887,
      "learning_rate": 1.3865934711097156e-05,
      "loss": 0.4468,
      "num_input_tokens_seen": 17034504,
      "step": 26035
    },
    {
      "epoch": 13.647798742138365,
      "grad_norm": 0.692677915096283,
      "learning_rate": 1.3855698185668675e-05,
      "loss": 0.4856,
      "num_input_tokens_seen": 17038120,
      "step": 26040
    },
    {
      "epoch": 13.65041928721174,
      "grad_norm": 0.5903641581535339,
      "learning_rate": 1.3845463991500985e-05,
      "loss": 0.4213,
      "num_input_tokens_seen": 17041128,
      "step": 26045
    },
    {
      "epoch": 13.653039832285115,
      "grad_norm": 0.3882724642753601,
      "learning_rate": 1.3835232130734968e-05,
      "loss": 0.8009,
      "num_input_tokens_seen": 17043464,
      "step": 26050
    },
    {
      "epoch": 13.65566037735849,
      "grad_norm": 0.4086623191833496,
      "learning_rate": 1.3825002605511017e-05,
      "loss": 0.5844,
      "num_input_tokens_seen": 17045960,
      "step": 26055
    },
    {
      "epoch": 13.658280922431866,
      "grad_norm": 0.6380544304847717,
      "learning_rate": 1.3814775417969037e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 17050280,
      "step": 26060
    },
    {
      "epoch": 13.66090146750524,
      "grad_norm": 0.41390690207481384,
      "learning_rate": 1.3804550570248431e-05,
      "loss": 0.552,
      "num_input_tokens_seen": 17053352,
      "step": 26065
    },
    {
      "epoch": 13.663522012578616,
      "grad_norm": 0.4632161855697632,
      "learning_rate": 1.3794328064488116e-05,
      "loss": 0.5717,
      "num_input_tokens_seen": 17055880,
      "step": 26070
    },
    {
      "epoch": 13.666142557651991,
      "grad_norm": 1.064009428024292,
      "learning_rate": 1.3784107902826523e-05,
      "loss": 0.4102,
      "num_input_tokens_seen": 17059368,
      "step": 26075
    },
    {
      "epoch": 13.668763102725366,
      "grad_norm": 0.40027689933776855,
      "learning_rate": 1.3773890087401603e-05,
      "loss": 0.4566,
      "num_input_tokens_seen": 17062184,
      "step": 26080
    },
    {
      "epoch": 13.671383647798741,
      "grad_norm": 0.5288861989974976,
      "learning_rate": 1.3763674620350802e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 17066504,
      "step": 26085
    },
    {
      "epoch": 13.674004192872118,
      "grad_norm": 0.40821418166160583,
      "learning_rate": 1.3753461503811089e-05,
      "loss": 0.4705,
      "num_input_tokens_seen": 17069512,
      "step": 26090
    },
    {
      "epoch": 13.676624737945493,
      "grad_norm": 0.4397789239883423,
      "learning_rate": 1.3743250739918922e-05,
      "loss": 0.4951,
      "num_input_tokens_seen": 17072040,
      "step": 26095
    },
    {
      "epoch": 13.679245283018869,
      "grad_norm": 0.5430120825767517,
      "learning_rate": 1.3733042330810265e-05,
      "loss": 0.3729,
      "num_input_tokens_seen": 17074536,
      "step": 26100
    },
    {
      "epoch": 13.681865828092244,
      "grad_norm": 0.3859555125236511,
      "learning_rate": 1.3722836278620616e-05,
      "loss": 0.3263,
      "num_input_tokens_seen": 17080168,
      "step": 26105
    },
    {
      "epoch": 13.684486373165619,
      "grad_norm": 0.8400462865829468,
      "learning_rate": 1.3712632585484957e-05,
      "loss": 0.4362,
      "num_input_tokens_seen": 17083208,
      "step": 26110
    },
    {
      "epoch": 13.687106918238994,
      "grad_norm": 0.4611712396144867,
      "learning_rate": 1.37024312535378e-05,
      "loss": 0.357,
      "num_input_tokens_seen": 17086632,
      "step": 26115
    },
    {
      "epoch": 13.68972746331237,
      "grad_norm": 0.6091726422309875,
      "learning_rate": 1.3692232284913133e-05,
      "loss": 0.494,
      "num_input_tokens_seen": 17090952,
      "step": 26120
    },
    {
      "epoch": 13.692348008385745,
      "grad_norm": 0.4152428209781647,
      "learning_rate": 1.3682035681744459e-05,
      "loss": 0.4292,
      "num_input_tokens_seen": 17093416,
      "step": 26125
    },
    {
      "epoch": 13.69496855345912,
      "grad_norm": 0.5434824824333191,
      "learning_rate": 1.3671841446164795e-05,
      "loss": 0.4539,
      "num_input_tokens_seen": 17097160,
      "step": 26130
    },
    {
      "epoch": 13.697589098532495,
      "grad_norm": 0.3032866418361664,
      "learning_rate": 1.3661649580306669e-05,
      "loss": 0.329,
      "num_input_tokens_seen": 17100680,
      "step": 26135
    },
    {
      "epoch": 13.70020964360587,
      "grad_norm": 0.4013344347476959,
      "learning_rate": 1.3651460086302087e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 17104456,
      "step": 26140
    },
    {
      "epoch": 13.702830188679245,
      "grad_norm": 0.4789547324180603,
      "learning_rate": 1.3641272966282609e-05,
      "loss": 0.3457,
      "num_input_tokens_seen": 17107080,
      "step": 26145
    },
    {
      "epoch": 13.70545073375262,
      "grad_norm": 1.1787939071655273,
      "learning_rate": 1.3631088222379218e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 17113064,
      "step": 26150
    },
    {
      "epoch": 13.708071278825996,
      "grad_norm": 0.4246126115322113,
      "learning_rate": 1.3620905856722468e-05,
      "loss": 0.4677,
      "num_input_tokens_seen": 17116072,
      "step": 26155
    },
    {
      "epoch": 13.71069182389937,
      "grad_norm": 0.4471943974494934,
      "learning_rate": 1.3610725871442392e-05,
      "loss": 0.3656,
      "num_input_tokens_seen": 17119720,
      "step": 26160
    },
    {
      "epoch": 13.713312368972746,
      "grad_norm": 1.5707124471664429,
      "learning_rate": 1.3600548268668532e-05,
      "loss": 0.4705,
      "num_input_tokens_seen": 17122312,
      "step": 26165
    },
    {
      "epoch": 13.715932914046121,
      "grad_norm": 0.7628295421600342,
      "learning_rate": 1.3590373050529927e-05,
      "loss": 0.4874,
      "num_input_tokens_seen": 17125544,
      "step": 26170
    },
    {
      "epoch": 13.718553459119496,
      "grad_norm": 0.3822871148586273,
      "learning_rate": 1.3580200219155116e-05,
      "loss": 0.4695,
      "num_input_tokens_seen": 17128968,
      "step": 26175
    },
    {
      "epoch": 13.721174004192871,
      "grad_norm": 0.2713906466960907,
      "learning_rate": 1.3570029776672122e-05,
      "loss": 0.4891,
      "num_input_tokens_seen": 17132712,
      "step": 26180
    },
    {
      "epoch": 13.723794549266248,
      "grad_norm": 0.47957348823547363,
      "learning_rate": 1.35598617252085e-05,
      "loss": 0.4031,
      "num_input_tokens_seen": 17136104,
      "step": 26185
    },
    {
      "epoch": 13.726415094339622,
      "grad_norm": 0.35722628235816956,
      "learning_rate": 1.354969606689129e-05,
      "loss": 0.3033,
      "num_input_tokens_seen": 17139240,
      "step": 26190
    },
    {
      "epoch": 13.729035639412999,
      "grad_norm": 0.5187379121780396,
      "learning_rate": 1.353953280384704e-05,
      "loss": 0.4492,
      "num_input_tokens_seen": 17142344,
      "step": 26195
    },
    {
      "epoch": 13.731656184486374,
      "grad_norm": 0.5048112869262695,
      "learning_rate": 1.352937193820177e-05,
      "loss": 0.5278,
      "num_input_tokens_seen": 17145992,
      "step": 26200
    },
    {
      "epoch": 13.734276729559749,
      "grad_norm": 0.33844467997550964,
      "learning_rate": 1.3519213472081038e-05,
      "loss": 0.4802,
      "num_input_tokens_seen": 17149544,
      "step": 26205
    },
    {
      "epoch": 13.736897274633124,
      "grad_norm": 0.4478645622730255,
      "learning_rate": 1.3509057407609854e-05,
      "loss": 0.4484,
      "num_input_tokens_seen": 17152616,
      "step": 26210
    },
    {
      "epoch": 13.7395178197065,
      "grad_norm": 0.5984859466552734,
      "learning_rate": 1.3498903746912769e-05,
      "loss": 0.578,
      "num_input_tokens_seen": 17154888,
      "step": 26215
    },
    {
      "epoch": 13.742138364779874,
      "grad_norm": 0.4677006006240845,
      "learning_rate": 1.3488752492113807e-05,
      "loss": 0.3739,
      "num_input_tokens_seen": 17157960,
      "step": 26220
    },
    {
      "epoch": 13.74475890985325,
      "grad_norm": 0.5180816054344177,
      "learning_rate": 1.3478603645336508e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 17160936,
      "step": 26225
    },
    {
      "epoch": 13.747379454926625,
      "grad_norm": 0.6264020204544067,
      "learning_rate": 1.3468457208703877e-05,
      "loss": 0.3562,
      "num_input_tokens_seen": 17164136,
      "step": 26230
    },
    {
      "epoch": 13.75,
      "grad_norm": 0.665576696395874,
      "learning_rate": 1.3458313184338437e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 17167176,
      "step": 26235
    },
    {
      "epoch": 13.752620545073375,
      "grad_norm": 0.39523106813430786,
      "learning_rate": 1.3448171574362201e-05,
      "loss": 0.4982,
      "num_input_tokens_seen": 17170504,
      "step": 26240
    },
    {
      "epoch": 13.75524109014675,
      "grad_norm": 0.42594990134239197,
      "learning_rate": 1.3438032380896681e-05,
      "loss": 0.3275,
      "num_input_tokens_seen": 17173896,
      "step": 26245
    },
    {
      "epoch": 13.757861635220126,
      "grad_norm": 0.5061226487159729,
      "learning_rate": 1.342789560606289e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 17176904,
      "step": 26250
    },
    {
      "epoch": 13.7604821802935,
      "grad_norm": 0.5126939415931702,
      "learning_rate": 1.341776125198131e-05,
      "loss": 0.3716,
      "num_input_tokens_seen": 17179784,
      "step": 26255
    },
    {
      "epoch": 13.763102725366876,
      "grad_norm": 0.7243406176567078,
      "learning_rate": 1.3407629320771948e-05,
      "loss": 0.4098,
      "num_input_tokens_seen": 17182984,
      "step": 26260
    },
    {
      "epoch": 13.765723270440251,
      "grad_norm": 0.3124243915081024,
      "learning_rate": 1.3397499814554266e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 17185608,
      "step": 26265
    },
    {
      "epoch": 13.768343815513626,
      "grad_norm": 0.5349884033203125,
      "learning_rate": 1.3387372735447256e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 17189512,
      "step": 26270
    },
    {
      "epoch": 13.770964360587001,
      "grad_norm": 0.4784131944179535,
      "learning_rate": 1.3377248085569399e-05,
      "loss": 0.4809,
      "num_input_tokens_seen": 17192360,
      "step": 26275
    },
    {
      "epoch": 13.773584905660378,
      "grad_norm": 0.5719372034072876,
      "learning_rate": 1.336712586703863e-05,
      "loss": 0.4727,
      "num_input_tokens_seen": 17196296,
      "step": 26280
    },
    {
      "epoch": 13.776205450733752,
      "grad_norm": 0.6674726605415344,
      "learning_rate": 1.3357006081972414e-05,
      "loss": 0.4661,
      "num_input_tokens_seen": 17199176,
      "step": 26285
    },
    {
      "epoch": 13.778825995807129,
      "grad_norm": 0.2601095736026764,
      "learning_rate": 1.3346888732487706e-05,
      "loss": 0.3647,
      "num_input_tokens_seen": 17203688,
      "step": 26290
    },
    {
      "epoch": 13.781446540880504,
      "grad_norm": 0.8032667636871338,
      "learning_rate": 1.3336773820700921e-05,
      "loss": 0.5014,
      "num_input_tokens_seen": 17206632,
      "step": 26295
    },
    {
      "epoch": 13.784067085953879,
      "grad_norm": 0.3607727289199829,
      "learning_rate": 1.3326661348727992e-05,
      "loss": 0.4766,
      "num_input_tokens_seen": 17209992,
      "step": 26300
    },
    {
      "epoch": 13.786687631027254,
      "grad_norm": 0.7493746876716614,
      "learning_rate": 1.3316551318684342e-05,
      "loss": 0.366,
      "num_input_tokens_seen": 17212968,
      "step": 26305
    },
    {
      "epoch": 13.78930817610063,
      "grad_norm": 1.9072339534759521,
      "learning_rate": 1.3306443732684854e-05,
      "loss": 0.3594,
      "num_input_tokens_seen": 17216200,
      "step": 26310
    },
    {
      "epoch": 13.791928721174004,
      "grad_norm": 0.5610479116439819,
      "learning_rate": 1.3296338592843929e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 17219144,
      "step": 26315
    },
    {
      "epoch": 13.79454926624738,
      "grad_norm": 0.3670852780342102,
      "learning_rate": 1.3286235901275457e-05,
      "loss": 0.3559,
      "num_input_tokens_seen": 17222792,
      "step": 26320
    },
    {
      "epoch": 13.797169811320755,
      "grad_norm": 0.42340439558029175,
      "learning_rate": 1.327613566009279e-05,
      "loss": 0.5107,
      "num_input_tokens_seen": 17226152,
      "step": 26325
    },
    {
      "epoch": 13.79979035639413,
      "grad_norm": 0.6447741389274597,
      "learning_rate": 1.3266037871408798e-05,
      "loss": 0.5109,
      "num_input_tokens_seen": 17228456,
      "step": 26330
    },
    {
      "epoch": 13.802410901467505,
      "grad_norm": 0.6415557861328125,
      "learning_rate": 1.3255942537335805e-05,
      "loss": 0.3556,
      "num_input_tokens_seen": 17231400,
      "step": 26335
    },
    {
      "epoch": 13.80503144654088,
      "grad_norm": 0.3029257655143738,
      "learning_rate": 1.3245849659985652e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 17235112,
      "step": 26340
    },
    {
      "epoch": 13.807651991614255,
      "grad_norm": 0.5221166610717773,
      "learning_rate": 1.3235759241469658e-05,
      "loss": 0.4375,
      "num_input_tokens_seen": 17238280,
      "step": 26345
    },
    {
      "epoch": 13.81027253668763,
      "grad_norm": 0.47522178292274475,
      "learning_rate": 1.322567128389861e-05,
      "loss": 0.4798,
      "num_input_tokens_seen": 17241192,
      "step": 26350
    },
    {
      "epoch": 13.812893081761006,
      "grad_norm": 0.9822267889976501,
      "learning_rate": 1.3215585789382812e-05,
      "loss": 0.6512,
      "num_input_tokens_seen": 17244712,
      "step": 26355
    },
    {
      "epoch": 13.815513626834381,
      "grad_norm": 0.4170745015144348,
      "learning_rate": 1.320550276003201e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 17248648,
      "step": 26360
    },
    {
      "epoch": 13.818134171907756,
      "grad_norm": 0.48817965388298035,
      "learning_rate": 1.3195422197955473e-05,
      "loss": 0.5146,
      "num_input_tokens_seen": 17252136,
      "step": 26365
    },
    {
      "epoch": 13.820754716981131,
      "grad_norm": 0.4670957326889038,
      "learning_rate": 1.3185344105261937e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 17254952,
      "step": 26370
    },
    {
      "epoch": 13.823375262054507,
      "grad_norm": 0.4984295666217804,
      "learning_rate": 1.3175268484059639e-05,
      "loss": 0.4054,
      "num_input_tokens_seen": 17258280,
      "step": 26375
    },
    {
      "epoch": 13.825995807127882,
      "grad_norm": 0.7058253288269043,
      "learning_rate": 1.316519533645626e-05,
      "loss": 0.4476,
      "num_input_tokens_seen": 17260840,
      "step": 26380
    },
    {
      "epoch": 13.828616352201259,
      "grad_norm": 0.5136238932609558,
      "learning_rate": 1.3155124664559004e-05,
      "loss": 0.4998,
      "num_input_tokens_seen": 17264488,
      "step": 26385
    },
    {
      "epoch": 13.831236897274634,
      "grad_norm": 0.7414189577102661,
      "learning_rate": 1.3145056470474525e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 17267592,
      "step": 26390
    },
    {
      "epoch": 13.833857442348009,
      "grad_norm": 0.5270785093307495,
      "learning_rate": 1.313499075630899e-05,
      "loss": 0.3041,
      "num_input_tokens_seen": 17270888,
      "step": 26395
    },
    {
      "epoch": 13.836477987421384,
      "grad_norm": 0.45916542410850525,
      "learning_rate": 1.312492752416802e-05,
      "loss": 0.4057,
      "num_input_tokens_seen": 17274152,
      "step": 26400
    },
    {
      "epoch": 13.83909853249476,
      "grad_norm": 0.4640324115753174,
      "learning_rate": 1.3114866776156748e-05,
      "loss": 0.6772,
      "num_input_tokens_seen": 17276808,
      "step": 26405
    },
    {
      "epoch": 13.841719077568134,
      "grad_norm": 0.3352217674255371,
      "learning_rate": 1.3104808514379748e-05,
      "loss": 0.3176,
      "num_input_tokens_seen": 17281192,
      "step": 26410
    },
    {
      "epoch": 13.84433962264151,
      "grad_norm": 0.7939228415489197,
      "learning_rate": 1.3094752740941094e-05,
      "loss": 0.3591,
      "num_input_tokens_seen": 17284936,
      "step": 26415
    },
    {
      "epoch": 13.846960167714885,
      "grad_norm": 0.6004365086555481,
      "learning_rate": 1.3084699457944344e-05,
      "loss": 0.5336,
      "num_input_tokens_seen": 17287624,
      "step": 26420
    },
    {
      "epoch": 13.84958071278826,
      "grad_norm": 1.0015422105789185,
      "learning_rate": 1.3074648667492528e-05,
      "loss": 0.4781,
      "num_input_tokens_seen": 17290600,
      "step": 26425
    },
    {
      "epoch": 13.852201257861635,
      "grad_norm": 0.4808834195137024,
      "learning_rate": 1.3064600371688171e-05,
      "loss": 0.4527,
      "num_input_tokens_seen": 17292904,
      "step": 26430
    },
    {
      "epoch": 13.85482180293501,
      "grad_norm": 0.40252891182899475,
      "learning_rate": 1.3054554572633249e-05,
      "loss": 0.4008,
      "num_input_tokens_seen": 17296232,
      "step": 26435
    },
    {
      "epoch": 13.857442348008385,
      "grad_norm": 0.46496230363845825,
      "learning_rate": 1.3044511272429222e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 17300360,
      "step": 26440
    },
    {
      "epoch": 13.86006289308176,
      "grad_norm": 0.46681028604507446,
      "learning_rate": 1.3034470473177038e-05,
      "loss": 0.4787,
      "num_input_tokens_seen": 17303720,
      "step": 26445
    },
    {
      "epoch": 13.862683438155136,
      "grad_norm": 0.9760419726371765,
      "learning_rate": 1.3024432176977125e-05,
      "loss": 0.4667,
      "num_input_tokens_seen": 17306568,
      "step": 26450
    },
    {
      "epoch": 13.865303983228511,
      "grad_norm": 0.5464662313461304,
      "learning_rate": 1.3014396385929372e-05,
      "loss": 0.417,
      "num_input_tokens_seen": 17310280,
      "step": 26455
    },
    {
      "epoch": 13.867924528301886,
      "grad_norm": 0.6402568817138672,
      "learning_rate": 1.3004363102133165e-05,
      "loss": 0.4021,
      "num_input_tokens_seen": 17315272,
      "step": 26460
    },
    {
      "epoch": 13.870545073375261,
      "grad_norm": 0.4072776138782501,
      "learning_rate": 1.2994332327687348e-05,
      "loss": 0.3368,
      "num_input_tokens_seen": 17319016,
      "step": 26465
    },
    {
      "epoch": 13.873165618448636,
      "grad_norm": 0.6125025153160095,
      "learning_rate": 1.2984304064690228e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 17323496,
      "step": 26470
    },
    {
      "epoch": 13.875786163522012,
      "grad_norm": 0.9071197509765625,
      "learning_rate": 1.2974278315239619e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 17325992,
      "step": 26475
    },
    {
      "epoch": 13.878406708595389,
      "grad_norm": 0.5130497217178345,
      "learning_rate": 1.2964255081432786e-05,
      "loss": 0.5006,
      "num_input_tokens_seen": 17328488,
      "step": 26480
    },
    {
      "epoch": 13.881027253668764,
      "grad_norm": 0.4379744231700897,
      "learning_rate": 1.2954234365366488e-05,
      "loss": 0.5632,
      "num_input_tokens_seen": 17331144,
      "step": 26485
    },
    {
      "epoch": 13.883647798742139,
      "grad_norm": 0.23898902535438538,
      "learning_rate": 1.2944216169136936e-05,
      "loss": 0.3292,
      "num_input_tokens_seen": 17335592,
      "step": 26490
    },
    {
      "epoch": 13.886268343815514,
      "grad_norm": 0.498931884765625,
      "learning_rate": 1.2934200494839815e-05,
      "loss": 0.4882,
      "num_input_tokens_seen": 17338856,
      "step": 26495
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 0.5148570537567139,
      "learning_rate": 1.2924187344570293e-05,
      "loss": 0.4809,
      "num_input_tokens_seen": 17342664,
      "step": 26500
    },
    {
      "epoch": 13.891509433962264,
      "grad_norm": 0.5318724513053894,
      "learning_rate": 1.2914176720423011e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 17346120,
      "step": 26505
    },
    {
      "epoch": 13.89412997903564,
      "grad_norm": 1.2166575193405151,
      "learning_rate": 1.2904168624492088e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 17349864,
      "step": 26510
    },
    {
      "epoch": 13.896750524109015,
      "grad_norm": 0.5594207048416138,
      "learning_rate": 1.2894163058871092e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 17352808,
      "step": 26515
    },
    {
      "epoch": 13.89937106918239,
      "grad_norm": 1.1223642826080322,
      "learning_rate": 1.2884160025653064e-05,
      "loss": 0.659,
      "num_input_tokens_seen": 17356232,
      "step": 26520
    },
    {
      "epoch": 13.901991614255765,
      "grad_norm": 0.44132497906684875,
      "learning_rate": 1.2874159526930532e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 17359272,
      "step": 26525
    },
    {
      "epoch": 13.90461215932914,
      "grad_norm": 0.3391801714897156,
      "learning_rate": 1.286416156479549e-05,
      "loss": 0.4541,
      "num_input_tokens_seen": 17363880,
      "step": 26530
    },
    {
      "epoch": 13.907232704402515,
      "grad_norm": 0.638221025466919,
      "learning_rate": 1.2854166141339394e-05,
      "loss": 0.4098,
      "num_input_tokens_seen": 17366600,
      "step": 26535
    },
    {
      "epoch": 13.90985324947589,
      "grad_norm": 0.45605048537254333,
      "learning_rate": 1.2844173258653183e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 17369544,
      "step": 26540
    },
    {
      "epoch": 13.912473794549266,
      "grad_norm": 0.4369024932384491,
      "learning_rate": 1.2834182918827247e-05,
      "loss": 0.4331,
      "num_input_tokens_seen": 17372360,
      "step": 26545
    },
    {
      "epoch": 13.915094339622641,
      "grad_norm": 1.4913533926010132,
      "learning_rate": 1.282419512395144e-05,
      "loss": 0.4811,
      "num_input_tokens_seen": 17375112,
      "step": 26550
    },
    {
      "epoch": 13.917714884696016,
      "grad_norm": 0.421392023563385,
      "learning_rate": 1.2814209876115102e-05,
      "loss": 0.3404,
      "num_input_tokens_seen": 17378024,
      "step": 26555
    },
    {
      "epoch": 13.920335429769391,
      "grad_norm": 0.39323267340660095,
      "learning_rate": 1.2804227177407038e-05,
      "loss": 0.3454,
      "num_input_tokens_seen": 17380744,
      "step": 26560
    },
    {
      "epoch": 13.922955974842766,
      "grad_norm": 0.31837165355682373,
      "learning_rate": 1.279424702991552e-05,
      "loss": 0.4406,
      "num_input_tokens_seen": 17383560,
      "step": 26565
    },
    {
      "epoch": 13.925576519916142,
      "grad_norm": 0.4154159724712372,
      "learning_rate": 1.2784269435728264e-05,
      "loss": 0.3812,
      "num_input_tokens_seen": 17386344,
      "step": 26570
    },
    {
      "epoch": 13.928197064989519,
      "grad_norm": 0.5349096059799194,
      "learning_rate": 1.2774294396932488e-05,
      "loss": 0.5168,
      "num_input_tokens_seen": 17389416,
      "step": 26575
    },
    {
      "epoch": 13.930817610062894,
      "grad_norm": 0.5026395320892334,
      "learning_rate": 1.2764321915614835e-05,
      "loss": 0.3744,
      "num_input_tokens_seen": 17392264,
      "step": 26580
    },
    {
      "epoch": 13.933438155136269,
      "grad_norm": 0.6340610980987549,
      "learning_rate": 1.2754351993861444e-05,
      "loss": 0.4193,
      "num_input_tokens_seen": 17395720,
      "step": 26585
    },
    {
      "epoch": 13.936058700209644,
      "grad_norm": 0.3085429072380066,
      "learning_rate": 1.2744384633757922e-05,
      "loss": 0.3808,
      "num_input_tokens_seen": 17398568,
      "step": 26590
    },
    {
      "epoch": 13.93867924528302,
      "grad_norm": 0.7945199608802795,
      "learning_rate": 1.2734419837389306e-05,
      "loss": 0.3564,
      "num_input_tokens_seen": 17401160,
      "step": 26595
    },
    {
      "epoch": 13.941299790356394,
      "grad_norm": 0.6212934255599976,
      "learning_rate": 1.2724457606840123e-05,
      "loss": 0.4897,
      "num_input_tokens_seen": 17404232,
      "step": 26600
    },
    {
      "epoch": 13.94392033542977,
      "grad_norm": 0.44725364446640015,
      "learning_rate": 1.2714497944194376e-05,
      "loss": 0.3032,
      "num_input_tokens_seen": 17407368,
      "step": 26605
    },
    {
      "epoch": 13.946540880503145,
      "grad_norm": 0.31779569387435913,
      "learning_rate": 1.2704540851535484e-05,
      "loss": 0.4247,
      "num_input_tokens_seen": 17410888,
      "step": 26610
    },
    {
      "epoch": 13.94916142557652,
      "grad_norm": 0.3520193099975586,
      "learning_rate": 1.2694586330946376e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 17413896,
      "step": 26615
    },
    {
      "epoch": 13.951781970649895,
      "grad_norm": 0.37957853078842163,
      "learning_rate": 1.2684634384509427e-05,
      "loss": 0.5542,
      "num_input_tokens_seen": 17417416,
      "step": 26620
    },
    {
      "epoch": 13.95440251572327,
      "grad_norm": 0.5030068755149841,
      "learning_rate": 1.2674685014306456e-05,
      "loss": 0.3792,
      "num_input_tokens_seen": 17420616,
      "step": 26625
    },
    {
      "epoch": 13.957023060796645,
      "grad_norm": 0.3207189738750458,
      "learning_rate": 1.2664738222418776e-05,
      "loss": 0.4541,
      "num_input_tokens_seen": 17424008,
      "step": 26630
    },
    {
      "epoch": 13.95964360587002,
      "grad_norm": 1.042925477027893,
      "learning_rate": 1.2654794010927118e-05,
      "loss": 0.5805,
      "num_input_tokens_seen": 17427272,
      "step": 26635
    },
    {
      "epoch": 13.962264150943396,
      "grad_norm": 0.5563085675239563,
      "learning_rate": 1.2644852381911715e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 17430408,
      "step": 26640
    },
    {
      "epoch": 13.964884696016771,
      "grad_norm": 0.45384523272514343,
      "learning_rate": 1.2634913337452248e-05,
      "loss": 0.4353,
      "num_input_tokens_seen": 17433192,
      "step": 26645
    },
    {
      "epoch": 13.967505241090146,
      "grad_norm": 0.5767471194267273,
      "learning_rate": 1.2624976879627832e-05,
      "loss": 0.3437,
      "num_input_tokens_seen": 17436232,
      "step": 26650
    },
    {
      "epoch": 13.970125786163521,
      "grad_norm": 0.5217151045799255,
      "learning_rate": 1.2615043010517069e-05,
      "loss": 0.5106,
      "num_input_tokens_seen": 17439688,
      "step": 26655
    },
    {
      "epoch": 13.972746331236896,
      "grad_norm": 0.6183866262435913,
      "learning_rate": 1.2605111732198027e-05,
      "loss": 0.4468,
      "num_input_tokens_seen": 17442088,
      "step": 26660
    },
    {
      "epoch": 13.975366876310272,
      "grad_norm": 0.2720390260219574,
      "learning_rate": 1.2595183046748188e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 17445960,
      "step": 26665
    },
    {
      "epoch": 13.977987421383649,
      "grad_norm": 0.3123055696487427,
      "learning_rate": 1.2585256956244545e-05,
      "loss": 0.4677,
      "num_input_tokens_seen": 17449096,
      "step": 26670
    },
    {
      "epoch": 13.980607966457024,
      "grad_norm": 0.29400885105133057,
      "learning_rate": 1.2575333462763506e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 17452744,
      "step": 26675
    },
    {
      "epoch": 13.983228511530399,
      "grad_norm": 0.6343994140625,
      "learning_rate": 1.2565412568380957e-05,
      "loss": 0.4898,
      "num_input_tokens_seen": 17456872,
      "step": 26680
    },
    {
      "epoch": 13.985849056603774,
      "grad_norm": 0.4523339569568634,
      "learning_rate": 1.2555494275172237e-05,
      "loss": 0.4175,
      "num_input_tokens_seen": 17460008,
      "step": 26685
    },
    {
      "epoch": 13.98846960167715,
      "grad_norm": 0.5532884001731873,
      "learning_rate": 1.2545578585212148e-05,
      "loss": 0.3615,
      "num_input_tokens_seen": 17463336,
      "step": 26690
    },
    {
      "epoch": 13.991090146750524,
      "grad_norm": 0.7443028688430786,
      "learning_rate": 1.2535665500574922e-05,
      "loss": 0.4461,
      "num_input_tokens_seen": 17465960,
      "step": 26695
    },
    {
      "epoch": 13.9937106918239,
      "grad_norm": 0.34603410959243774,
      "learning_rate": 1.2525755023334285e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 17469448,
      "step": 26700
    },
    {
      "epoch": 13.996331236897275,
      "grad_norm": 0.6717372536659241,
      "learning_rate": 1.251584715556337e-05,
      "loss": 0.4385,
      "num_input_tokens_seen": 17473128,
      "step": 26705
    },
    {
      "epoch": 13.99895178197065,
      "grad_norm": 0.5834943652153015,
      "learning_rate": 1.2505941899334805e-05,
      "loss": 0.4816,
      "num_input_tokens_seen": 17476616,
      "step": 26710
    },
    {
      "epoch": 14.0,
      "eval_loss": 0.4472961723804474,
      "eval_runtime": 9.7213,
      "eval_samples_per_second": 87.232,
      "eval_steps_per_second": 21.808,
      "num_input_tokens_seen": 17477224,
      "step": 26712
    },
    {
      "epoch": 14.001572327044025,
      "grad_norm": 0.4915034770965576,
      "learning_rate": 1.2496039256720665e-05,
      "loss": 0.421,
      "num_input_tokens_seen": 17479176,
      "step": 26715
    },
    {
      "epoch": 14.0041928721174,
      "grad_norm": 0.375532865524292,
      "learning_rate": 1.2486139229792452e-05,
      "loss": 0.4243,
      "num_input_tokens_seen": 17482760,
      "step": 26720
    },
    {
      "epoch": 14.006813417190775,
      "grad_norm": 0.35893845558166504,
      "learning_rate": 1.2476241820621152e-05,
      "loss": 0.5195,
      "num_input_tokens_seen": 17485992,
      "step": 26725
    },
    {
      "epoch": 14.00943396226415,
      "grad_norm": 0.6087861061096191,
      "learning_rate": 1.2466347031277176e-05,
      "loss": 0.3814,
      "num_input_tokens_seen": 17488712,
      "step": 26730
    },
    {
      "epoch": 14.012054507337526,
      "grad_norm": 0.5098602771759033,
      "learning_rate": 1.2456454863830411e-05,
      "loss": 0.6716,
      "num_input_tokens_seen": 17491784,
      "step": 26735
    },
    {
      "epoch": 14.014675052410901,
      "grad_norm": 0.6417111158370972,
      "learning_rate": 1.2446565320350185e-05,
      "loss": 0.4048,
      "num_input_tokens_seen": 17495368,
      "step": 26740
    },
    {
      "epoch": 14.017295597484276,
      "grad_norm": 0.4953629672527313,
      "learning_rate": 1.2436678402905283e-05,
      "loss": 0.3358,
      "num_input_tokens_seen": 17499272,
      "step": 26745
    },
    {
      "epoch": 14.019916142557651,
      "grad_norm": 0.6264583468437195,
      "learning_rate": 1.2426794113563933e-05,
      "loss": 0.5166,
      "num_input_tokens_seen": 17502728,
      "step": 26750
    },
    {
      "epoch": 14.022536687631026,
      "grad_norm": 0.7738600969314575,
      "learning_rate": 1.24169124543938e-05,
      "loss": 0.564,
      "num_input_tokens_seen": 17506984,
      "step": 26755
    },
    {
      "epoch": 14.025157232704403,
      "grad_norm": 0.3189222812652588,
      "learning_rate": 1.2407033427462026e-05,
      "loss": 0.5861,
      "num_input_tokens_seen": 17511848,
      "step": 26760
    },
    {
      "epoch": 14.027777777777779,
      "grad_norm": 0.5654281973838806,
      "learning_rate": 1.2397157034835192e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 17514696,
      "step": 26765
    },
    {
      "epoch": 14.030398322851154,
      "grad_norm": 0.37871912121772766,
      "learning_rate": 1.2387283278579325e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 17517352,
      "step": 26770
    },
    {
      "epoch": 14.033018867924529,
      "grad_norm": 0.3400212824344635,
      "learning_rate": 1.237741216075991e-05,
      "loss": 0.4342,
      "num_input_tokens_seen": 17520872,
      "step": 26775
    },
    {
      "epoch": 14.035639412997904,
      "grad_norm": 0.47039610147476196,
      "learning_rate": 1.2367543683441868e-05,
      "loss": 0.3498,
      "num_input_tokens_seen": 17523880,
      "step": 26780
    },
    {
      "epoch": 14.03825995807128,
      "grad_norm": 0.22602173686027527,
      "learning_rate": 1.2357677848689558e-05,
      "loss": 0.3582,
      "num_input_tokens_seen": 17527016,
      "step": 26785
    },
    {
      "epoch": 14.040880503144654,
      "grad_norm": 1.2434004545211792,
      "learning_rate": 1.2347814658566811e-05,
      "loss": 0.5165,
      "num_input_tokens_seen": 17531688,
      "step": 26790
    },
    {
      "epoch": 14.04350104821803,
      "grad_norm": 0.4852937161922455,
      "learning_rate": 1.2337954115136894e-05,
      "loss": 0.3331,
      "num_input_tokens_seen": 17534760,
      "step": 26795
    },
    {
      "epoch": 14.046121593291405,
      "grad_norm": 0.7110370397567749,
      "learning_rate": 1.2328096220462529e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 17537672,
      "step": 26800
    },
    {
      "epoch": 14.04874213836478,
      "grad_norm": 0.5983697175979614,
      "learning_rate": 1.2318240976605865e-05,
      "loss": 0.4041,
      "num_input_tokens_seen": 17541480,
      "step": 26805
    },
    {
      "epoch": 14.051362683438155,
      "grad_norm": 0.38434720039367676,
      "learning_rate": 1.23083883856285e-05,
      "loss": 0.3107,
      "num_input_tokens_seen": 17544040,
      "step": 26810
    },
    {
      "epoch": 14.05398322851153,
      "grad_norm": 0.4247094690799713,
      "learning_rate": 1.229853844959149e-05,
      "loss": 0.3335,
      "num_input_tokens_seen": 17546376,
      "step": 26815
    },
    {
      "epoch": 14.056603773584905,
      "grad_norm": 0.49659326672554016,
      "learning_rate": 1.228869117055533e-05,
      "loss": 0.3698,
      "num_input_tokens_seen": 17549576,
      "step": 26820
    },
    {
      "epoch": 14.05922431865828,
      "grad_norm": 0.3498929440975189,
      "learning_rate": 1.227884655057997e-05,
      "loss": 0.3894,
      "num_input_tokens_seen": 17552776,
      "step": 26825
    },
    {
      "epoch": 14.061844863731656,
      "grad_norm": 0.5091464519500732,
      "learning_rate": 1.2269004591724777e-05,
      "loss": 0.4641,
      "num_input_tokens_seen": 17556328,
      "step": 26830
    },
    {
      "epoch": 14.064465408805031,
      "grad_norm": 0.6783746480941772,
      "learning_rate": 1.2259165296048575e-05,
      "loss": 0.7436,
      "num_input_tokens_seen": 17559144,
      "step": 26835
    },
    {
      "epoch": 14.067085953878406,
      "grad_norm": 0.6659204363822937,
      "learning_rate": 1.2249328665609638e-05,
      "loss": 0.54,
      "num_input_tokens_seen": 17561960,
      "step": 26840
    },
    {
      "epoch": 14.069706498951781,
      "grad_norm": 0.3166906535625458,
      "learning_rate": 1.2239494702465676e-05,
      "loss": 0.5065,
      "num_input_tokens_seen": 17565480,
      "step": 26845
    },
    {
      "epoch": 14.072327044025156,
      "grad_norm": 0.5611873269081116,
      "learning_rate": 1.2229663408673842e-05,
      "loss": 0.4659,
      "num_input_tokens_seen": 17568456,
      "step": 26850
    },
    {
      "epoch": 14.074947589098532,
      "grad_norm": 0.5028332471847534,
      "learning_rate": 1.221983478629074e-05,
      "loss": 0.38,
      "num_input_tokens_seen": 17571336,
      "step": 26855
    },
    {
      "epoch": 14.077568134171909,
      "grad_norm": 0.7052468657493591,
      "learning_rate": 1.22100088373724e-05,
      "loss": 0.6108,
      "num_input_tokens_seen": 17577576,
      "step": 26860
    },
    {
      "epoch": 14.080188679245284,
      "grad_norm": 0.411699503660202,
      "learning_rate": 1.2200185563974283e-05,
      "loss": 0.3674,
      "num_input_tokens_seen": 17581352,
      "step": 26865
    },
    {
      "epoch": 14.082809224318659,
      "grad_norm": 0.38662785291671753,
      "learning_rate": 1.219036496815132e-05,
      "loss": 0.5144,
      "num_input_tokens_seen": 17584424,
      "step": 26870
    },
    {
      "epoch": 14.085429769392034,
      "grad_norm": 0.29541879892349243,
      "learning_rate": 1.2180547051957863e-05,
      "loss": 0.4265,
      "num_input_tokens_seen": 17587656,
      "step": 26875
    },
    {
      "epoch": 14.08805031446541,
      "grad_norm": 0.46663644909858704,
      "learning_rate": 1.2170731817447723e-05,
      "loss": 0.4241,
      "num_input_tokens_seen": 17591528,
      "step": 26880
    },
    {
      "epoch": 14.090670859538784,
      "grad_norm": 0.5227928757667542,
      "learning_rate": 1.216091926667412e-05,
      "loss": 0.4705,
      "num_input_tokens_seen": 17594536,
      "step": 26885
    },
    {
      "epoch": 14.09329140461216,
      "grad_norm": 0.4479154646396637,
      "learning_rate": 1.2151109401689723e-05,
      "loss": 0.4975,
      "num_input_tokens_seen": 17597640,
      "step": 26890
    },
    {
      "epoch": 14.095911949685535,
      "grad_norm": 0.6911330223083496,
      "learning_rate": 1.214130222454665e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 17600136,
      "step": 26895
    },
    {
      "epoch": 14.09853249475891,
      "grad_norm": 0.4510404169559479,
      "learning_rate": 1.2131497737296452e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 17602856,
      "step": 26900
    },
    {
      "epoch": 14.101153039832285,
      "grad_norm": 0.5121756792068481,
      "learning_rate": 1.2121695941990124e-05,
      "loss": 0.4022,
      "num_input_tokens_seen": 17606344,
      "step": 26905
    },
    {
      "epoch": 14.10377358490566,
      "grad_norm": 0.6010337471961975,
      "learning_rate": 1.2111896840678075e-05,
      "loss": 0.4307,
      "num_input_tokens_seen": 17609064,
      "step": 26910
    },
    {
      "epoch": 14.106394129979035,
      "grad_norm": 0.28199464082717896,
      "learning_rate": 1.210210043541018e-05,
      "loss": 0.4625,
      "num_input_tokens_seen": 17612456,
      "step": 26915
    },
    {
      "epoch": 14.10901467505241,
      "grad_norm": 0.7623864412307739,
      "learning_rate": 1.2092306728235717e-05,
      "loss": 0.355,
      "num_input_tokens_seen": 17616424,
      "step": 26920
    },
    {
      "epoch": 14.111635220125786,
      "grad_norm": 0.9373316764831543,
      "learning_rate": 1.2082515721203427e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 17619944,
      "step": 26925
    },
    {
      "epoch": 14.114255765199161,
      "grad_norm": 0.3365289866924286,
      "learning_rate": 1.2072727416361479e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 17623368,
      "step": 26930
    },
    {
      "epoch": 14.116876310272536,
      "grad_norm": 0.6438720226287842,
      "learning_rate": 1.2062941815757484e-05,
      "loss": 0.6187,
      "num_input_tokens_seen": 17626760,
      "step": 26935
    },
    {
      "epoch": 14.119496855345911,
      "grad_norm": 0.7201592922210693,
      "learning_rate": 1.2053158921438454e-05,
      "loss": 0.5328,
      "num_input_tokens_seen": 17629608,
      "step": 26940
    },
    {
      "epoch": 14.122117400419286,
      "grad_norm": 0.42890217900276184,
      "learning_rate": 1.2043378735450885e-05,
      "loss": 0.3863,
      "num_input_tokens_seen": 17632072,
      "step": 26945
    },
    {
      "epoch": 14.124737945492662,
      "grad_norm": 0.3577166795730591,
      "learning_rate": 1.2033601259840657e-05,
      "loss": 0.4229,
      "num_input_tokens_seen": 17635720,
      "step": 26950
    },
    {
      "epoch": 14.127358490566039,
      "grad_norm": 0.43954017758369446,
      "learning_rate": 1.2023826496653118e-05,
      "loss": 0.4053,
      "num_input_tokens_seen": 17638568,
      "step": 26955
    },
    {
      "epoch": 14.129979035639414,
      "grad_norm": 0.5804535746574402,
      "learning_rate": 1.2014054447933044e-05,
      "loss": 0.3621,
      "num_input_tokens_seen": 17643080,
      "step": 26960
    },
    {
      "epoch": 14.132599580712789,
      "grad_norm": 0.9326061010360718,
      "learning_rate": 1.200428511572462e-05,
      "loss": 0.4057,
      "num_input_tokens_seen": 17646024,
      "step": 26965
    },
    {
      "epoch": 14.135220125786164,
      "grad_norm": 0.32560262084007263,
      "learning_rate": 1.1994518502071484e-05,
      "loss": 0.4183,
      "num_input_tokens_seen": 17648936,
      "step": 26970
    },
    {
      "epoch": 14.13784067085954,
      "grad_norm": 0.29273951053619385,
      "learning_rate": 1.1984754609016716e-05,
      "loss": 0.4163,
      "num_input_tokens_seen": 17653160,
      "step": 26975
    },
    {
      "epoch": 14.140461215932914,
      "grad_norm": 0.23035329580307007,
      "learning_rate": 1.1974993438602785e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 17657384,
      "step": 26980
    },
    {
      "epoch": 14.14308176100629,
      "grad_norm": 0.6278455853462219,
      "learning_rate": 1.1965234992871641e-05,
      "loss": 0.4513,
      "num_input_tokens_seen": 17660424,
      "step": 26985
    },
    {
      "epoch": 14.145702306079665,
      "grad_norm": 0.3837626576423645,
      "learning_rate": 1.195547927386462e-05,
      "loss": 0.4583,
      "num_input_tokens_seen": 17664424,
      "step": 26990
    },
    {
      "epoch": 14.14832285115304,
      "grad_norm": 0.8413761854171753,
      "learning_rate": 1.1945726283622514e-05,
      "loss": 0.4667,
      "num_input_tokens_seen": 17667368,
      "step": 26995
    },
    {
      "epoch": 14.150943396226415,
      "grad_norm": 0.5067566633224487,
      "learning_rate": 1.1935976024185547e-05,
      "loss": 0.4169,
      "num_input_tokens_seen": 17670184,
      "step": 27000
    },
    {
      "epoch": 14.15356394129979,
      "grad_norm": 0.5781554579734802,
      "learning_rate": 1.1926228497593341e-05,
      "loss": 0.4975,
      "num_input_tokens_seen": 17673448,
      "step": 27005
    },
    {
      "epoch": 14.156184486373165,
      "grad_norm": 0.6020471453666687,
      "learning_rate": 1.1916483705884985e-05,
      "loss": 0.3982,
      "num_input_tokens_seen": 17676328,
      "step": 27010
    },
    {
      "epoch": 14.15880503144654,
      "grad_norm": 0.43992820382118225,
      "learning_rate": 1.1906741651098983e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 17680232,
      "step": 27015
    },
    {
      "epoch": 14.161425576519916,
      "grad_norm": 0.39959508180618286,
      "learning_rate": 1.1897002335273239e-05,
      "loss": 0.4396,
      "num_input_tokens_seen": 17683880,
      "step": 27020
    },
    {
      "epoch": 14.164046121593291,
      "grad_norm": 0.3955145478248596,
      "learning_rate": 1.1887265760445121e-05,
      "loss": 0.5692,
      "num_input_tokens_seen": 17687080,
      "step": 27025
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 0.3793655037879944,
      "learning_rate": 1.1877531928651417e-05,
      "loss": 0.4415,
      "num_input_tokens_seen": 17690344,
      "step": 27030
    },
    {
      "epoch": 14.169287211740041,
      "grad_norm": 0.6003156900405884,
      "learning_rate": 1.1867800841928313e-05,
      "loss": 0.5059,
      "num_input_tokens_seen": 17695272,
      "step": 27035
    },
    {
      "epoch": 14.171907756813416,
      "grad_norm": 0.4486010670661926,
      "learning_rate": 1.1858072502311465e-05,
      "loss": 0.3128,
      "num_input_tokens_seen": 17698216,
      "step": 27040
    },
    {
      "epoch": 14.174528301886792,
      "grad_norm": 0.39104411005973816,
      "learning_rate": 1.1848346911835908e-05,
      "loss": 0.3889,
      "num_input_tokens_seen": 17700968,
      "step": 27045
    },
    {
      "epoch": 14.177148846960169,
      "grad_norm": 0.5572051405906677,
      "learning_rate": 1.1838624072536137e-05,
      "loss": 0.3389,
      "num_input_tokens_seen": 17703496,
      "step": 27050
    },
    {
      "epoch": 14.179769392033544,
      "grad_norm": 0.2822210490703583,
      "learning_rate": 1.1828903986446055e-05,
      "loss": 0.3605,
      "num_input_tokens_seen": 17706536,
      "step": 27055
    },
    {
      "epoch": 14.182389937106919,
      "grad_norm": 0.4734441637992859,
      "learning_rate": 1.1819186655599007e-05,
      "loss": 0.3854,
      "num_input_tokens_seen": 17710376,
      "step": 27060
    },
    {
      "epoch": 14.185010482180294,
      "grad_norm": 0.3091236650943756,
      "learning_rate": 1.1809472082027734e-05,
      "loss": 0.3782,
      "num_input_tokens_seen": 17713320,
      "step": 27065
    },
    {
      "epoch": 14.18763102725367,
      "grad_norm": 0.3632130026817322,
      "learning_rate": 1.1799760267764407e-05,
      "loss": 0.4842,
      "num_input_tokens_seen": 17716392,
      "step": 27070
    },
    {
      "epoch": 14.190251572327044,
      "grad_norm": 0.5998337268829346,
      "learning_rate": 1.1790051214840634e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 17720488,
      "step": 27075
    },
    {
      "epoch": 14.19287211740042,
      "grad_norm": 0.4334302246570587,
      "learning_rate": 1.1780344925287443e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 17724040,
      "step": 27080
    },
    {
      "epoch": 14.195492662473795,
      "grad_norm": 0.44510504603385925,
      "learning_rate": 1.1770641401135282e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 17727208,
      "step": 27085
    },
    {
      "epoch": 14.19811320754717,
      "grad_norm": 0.368598610162735,
      "learning_rate": 1.1760940644414003e-05,
      "loss": 0.3364,
      "num_input_tokens_seen": 17730504,
      "step": 27090
    },
    {
      "epoch": 14.200733752620545,
      "grad_norm": 0.6717376112937927,
      "learning_rate": 1.1751242657152914e-05,
      "loss": 0.4302,
      "num_input_tokens_seen": 17733800,
      "step": 27095
    },
    {
      "epoch": 14.20335429769392,
      "grad_norm": 0.6497684717178345,
      "learning_rate": 1.1741547441380698e-05,
      "loss": 0.5196,
      "num_input_tokens_seen": 17736072,
      "step": 27100
    },
    {
      "epoch": 14.205974842767295,
      "grad_norm": 0.533877432346344,
      "learning_rate": 1.17318549991255e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 17739336,
      "step": 27105
    },
    {
      "epoch": 14.20859538784067,
      "grad_norm": 0.3395906984806061,
      "learning_rate": 1.1722165332414863e-05,
      "loss": 0.4576,
      "num_input_tokens_seen": 17743528,
      "step": 27110
    },
    {
      "epoch": 14.211215932914046,
      "grad_norm": 0.6028692722320557,
      "learning_rate": 1.1712478443275768e-05,
      "loss": 0.3366,
      "num_input_tokens_seen": 17746024,
      "step": 27115
    },
    {
      "epoch": 14.213836477987421,
      "grad_norm": 0.4819314479827881,
      "learning_rate": 1.170279433373459e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 17750088,
      "step": 27120
    },
    {
      "epoch": 14.216457023060796,
      "grad_norm": 0.7358178496360779,
      "learning_rate": 1.1693113005817125e-05,
      "loss": 0.5746,
      "num_input_tokens_seen": 17753224,
      "step": 27125
    },
    {
      "epoch": 14.219077568134171,
      "grad_norm": 0.4481196701526642,
      "learning_rate": 1.1683434461548606e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 17756488,
      "step": 27130
    },
    {
      "epoch": 14.221698113207546,
      "grad_norm": 0.5074687004089355,
      "learning_rate": 1.1673758702953675e-05,
      "loss": 0.461,
      "num_input_tokens_seen": 17759336,
      "step": 27135
    },
    {
      "epoch": 14.224318658280922,
      "grad_norm": 0.5399202108383179,
      "learning_rate": 1.1664085732056398e-05,
      "loss": 0.547,
      "num_input_tokens_seen": 17761672,
      "step": 27140
    },
    {
      "epoch": 14.226939203354299,
      "grad_norm": 0.5833566188812256,
      "learning_rate": 1.1654415550880243e-05,
      "loss": 0.4635,
      "num_input_tokens_seen": 17764136,
      "step": 27145
    },
    {
      "epoch": 14.229559748427674,
      "grad_norm": 0.3014107346534729,
      "learning_rate": 1.1644748161448088e-05,
      "loss": 0.4164,
      "num_input_tokens_seen": 17767048,
      "step": 27150
    },
    {
      "epoch": 14.232180293501049,
      "grad_norm": 0.27110546827316284,
      "learning_rate": 1.1635083565782257e-05,
      "loss": 0.5118,
      "num_input_tokens_seen": 17770184,
      "step": 27155
    },
    {
      "epoch": 14.234800838574424,
      "grad_norm": 0.6205654144287109,
      "learning_rate": 1.1625421765904465e-05,
      "loss": 0.3892,
      "num_input_tokens_seen": 17772680,
      "step": 27160
    },
    {
      "epoch": 14.2374213836478,
      "grad_norm": 0.5364089608192444,
      "learning_rate": 1.1615762763835858e-05,
      "loss": 0.4561,
      "num_input_tokens_seen": 17775784,
      "step": 27165
    },
    {
      "epoch": 14.240041928721174,
      "grad_norm": 0.3543245494365692,
      "learning_rate": 1.1606106561596993e-05,
      "loss": 0.3253,
      "num_input_tokens_seen": 17779016,
      "step": 27170
    },
    {
      "epoch": 14.24266247379455,
      "grad_norm": 0.33730220794677734,
      "learning_rate": 1.1596453161207829e-05,
      "loss": 0.3641,
      "num_input_tokens_seen": 17782248,
      "step": 27175
    },
    {
      "epoch": 14.245283018867925,
      "grad_norm": 0.47915637493133545,
      "learning_rate": 1.1586802564687741e-05,
      "loss": 0.4752,
      "num_input_tokens_seen": 17785320,
      "step": 27180
    },
    {
      "epoch": 14.2479035639413,
      "grad_norm": 0.8367030024528503,
      "learning_rate": 1.1577154774055528e-05,
      "loss": 0.395,
      "num_input_tokens_seen": 17788520,
      "step": 27185
    },
    {
      "epoch": 14.250524109014675,
      "grad_norm": 0.3427007496356964,
      "learning_rate": 1.1567509791329401e-05,
      "loss": 0.8558,
      "num_input_tokens_seen": 17792200,
      "step": 27190
    },
    {
      "epoch": 14.25314465408805,
      "grad_norm": 0.33579859137535095,
      "learning_rate": 1.155786761852699e-05,
      "loss": 0.3543,
      "num_input_tokens_seen": 17795816,
      "step": 27195
    },
    {
      "epoch": 14.255765199161425,
      "grad_norm": 0.41213443875312805,
      "learning_rate": 1.1548228257665317e-05,
      "loss": 0.4488,
      "num_input_tokens_seen": 17798568,
      "step": 27200
    },
    {
      "epoch": 14.2583857442348,
      "grad_norm": 0.5199078917503357,
      "learning_rate": 1.1538591710760815e-05,
      "loss": 0.3724,
      "num_input_tokens_seen": 17801064,
      "step": 27205
    },
    {
      "epoch": 14.261006289308176,
      "grad_norm": 0.46643051505088806,
      "learning_rate": 1.1528957979829355e-05,
      "loss": 0.4269,
      "num_input_tokens_seen": 17803400,
      "step": 27210
    },
    {
      "epoch": 14.26362683438155,
      "grad_norm": 0.6369353532791138,
      "learning_rate": 1.1519327066886198e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 17806696,
      "step": 27215
    },
    {
      "epoch": 14.266247379454926,
      "grad_norm": 0.3825358748435974,
      "learning_rate": 1.1509698973946032e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 17810216,
      "step": 27220
    },
    {
      "epoch": 14.268867924528301,
      "grad_norm": 0.6231425404548645,
      "learning_rate": 1.1500073703022926e-05,
      "loss": 0.5044,
      "num_input_tokens_seen": 17813544,
      "step": 27225
    },
    {
      "epoch": 14.271488469601676,
      "grad_norm": 0.4869782626628876,
      "learning_rate": 1.1490451256130394e-05,
      "loss": 0.5576,
      "num_input_tokens_seen": 17816712,
      "step": 27230
    },
    {
      "epoch": 14.274109014675052,
      "grad_norm": 0.5245117545127869,
      "learning_rate": 1.1480831635281328e-05,
      "loss": 0.324,
      "num_input_tokens_seen": 17819016,
      "step": 27235
    },
    {
      "epoch": 14.276729559748428,
      "grad_norm": 0.5174924731254578,
      "learning_rate": 1.1471214842488046e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 17822536,
      "step": 27240
    },
    {
      "epoch": 14.279350104821804,
      "grad_norm": 0.3246302604675293,
      "learning_rate": 1.1461600879762277e-05,
      "loss": 0.4078,
      "num_input_tokens_seen": 17826056,
      "step": 27245
    },
    {
      "epoch": 14.281970649895179,
      "grad_norm": 0.45374464988708496,
      "learning_rate": 1.1451989749115157e-05,
      "loss": 0.4349,
      "num_input_tokens_seen": 17828712,
      "step": 27250
    },
    {
      "epoch": 14.284591194968554,
      "grad_norm": 0.5532503724098206,
      "learning_rate": 1.1442381452557219e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 17832072,
      "step": 27255
    },
    {
      "epoch": 14.28721174004193,
      "grad_norm": 0.34059759974479675,
      "learning_rate": 1.1432775992098397e-05,
      "loss": 0.426,
      "num_input_tokens_seen": 17834920,
      "step": 27260
    },
    {
      "epoch": 14.289832285115304,
      "grad_norm": 0.5801016688346863,
      "learning_rate": 1.1423173369748056e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 17838568,
      "step": 27265
    },
    {
      "epoch": 14.29245283018868,
      "grad_norm": 1.3288884162902832,
      "learning_rate": 1.1413573587514952e-05,
      "loss": 0.4347,
      "num_input_tokens_seen": 17842504,
      "step": 27270
    },
    {
      "epoch": 14.295073375262055,
      "grad_norm": 0.5274880528450012,
      "learning_rate": 1.140397664740726e-05,
      "loss": 0.5994,
      "num_input_tokens_seen": 17846312,
      "step": 27275
    },
    {
      "epoch": 14.29769392033543,
      "grad_norm": 0.5432972311973572,
      "learning_rate": 1.1394382551432534e-05,
      "loss": 0.4021,
      "num_input_tokens_seen": 17848424,
      "step": 27280
    },
    {
      "epoch": 14.300314465408805,
      "grad_norm": 0.6280845403671265,
      "learning_rate": 1.1384791301597767e-05,
      "loss": 0.4359,
      "num_input_tokens_seen": 17851176,
      "step": 27285
    },
    {
      "epoch": 14.30293501048218,
      "grad_norm": 0.6825987696647644,
      "learning_rate": 1.1375202899909317e-05,
      "loss": 0.3553,
      "num_input_tokens_seen": 17853800,
      "step": 27290
    },
    {
      "epoch": 14.305555555555555,
      "grad_norm": 0.6424171924591064,
      "learning_rate": 1.1365617348372984e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 17857160,
      "step": 27295
    },
    {
      "epoch": 14.30817610062893,
      "grad_norm": 0.4544006288051605,
      "learning_rate": 1.135603464899396e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 17860968,
      "step": 27300
    },
    {
      "epoch": 14.310796645702306,
      "grad_norm": 0.5449968576431274,
      "learning_rate": 1.1346454803776821e-05,
      "loss": 0.5177,
      "num_input_tokens_seen": 17863848,
      "step": 27305
    },
    {
      "epoch": 14.31341719077568,
      "grad_norm": 0.5421252846717834,
      "learning_rate": 1.1336877814725566e-05,
      "loss": 0.3693,
      "num_input_tokens_seen": 17868616,
      "step": 27310
    },
    {
      "epoch": 14.316037735849056,
      "grad_norm": 0.5447478890419006,
      "learning_rate": 1.1327303683843607e-05,
      "loss": 0.4041,
      "num_input_tokens_seen": 17871720,
      "step": 27315
    },
    {
      "epoch": 14.318658280922431,
      "grad_norm": 0.3568566143512726,
      "learning_rate": 1.1317732413133723e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 17874664,
      "step": 27320
    },
    {
      "epoch": 14.321278825995806,
      "grad_norm": 0.4039428234100342,
      "learning_rate": 1.1308164004598118e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 17878184,
      "step": 27325
    },
    {
      "epoch": 14.323899371069182,
      "grad_norm": 0.7337590456008911,
      "learning_rate": 1.129859846023841e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 17881416,
      "step": 27330
    },
    {
      "epoch": 14.326519916142558,
      "grad_norm": 0.4511480927467346,
      "learning_rate": 1.1289035782055582e-05,
      "loss": 0.5927,
      "num_input_tokens_seen": 17884136,
      "step": 27335
    },
    {
      "epoch": 14.329140461215934,
      "grad_norm": 0.6838142275810242,
      "learning_rate": 1.1279475972050044e-05,
      "loss": 0.5527,
      "num_input_tokens_seen": 17886792,
      "step": 27340
    },
    {
      "epoch": 14.331761006289309,
      "grad_norm": 0.6453275680541992,
      "learning_rate": 1.1269919032221612e-05,
      "loss": 0.4586,
      "num_input_tokens_seen": 17889864,
      "step": 27345
    },
    {
      "epoch": 14.334381551362684,
      "grad_norm": 0.5347544550895691,
      "learning_rate": 1.1260364964569466e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 17893160,
      "step": 27350
    },
    {
      "epoch": 14.33700209643606,
      "grad_norm": 0.3943687975406647,
      "learning_rate": 1.125081377109223e-05,
      "loss": 0.535,
      "num_input_tokens_seen": 17896360,
      "step": 27355
    },
    {
      "epoch": 14.339622641509434,
      "grad_norm": 0.4674961268901825,
      "learning_rate": 1.1241265453787882e-05,
      "loss": 0.5197,
      "num_input_tokens_seen": 17899208,
      "step": 27360
    },
    {
      "epoch": 14.34224318658281,
      "grad_norm": 0.3678845763206482,
      "learning_rate": 1.1231720014653837e-05,
      "loss": 0.407,
      "num_input_tokens_seen": 17901864,
      "step": 27365
    },
    {
      "epoch": 14.344863731656185,
      "grad_norm": 0.4324520230293274,
      "learning_rate": 1.1222177455686896e-05,
      "loss": 0.3941,
      "num_input_tokens_seen": 17905288,
      "step": 27370
    },
    {
      "epoch": 14.34748427672956,
      "grad_norm": 0.45092689990997314,
      "learning_rate": 1.1212637778883237e-05,
      "loss": 0.4162,
      "num_input_tokens_seen": 17908104,
      "step": 27375
    },
    {
      "epoch": 14.350104821802935,
      "grad_norm": 0.3133566677570343,
      "learning_rate": 1.1203100986238468e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 17912008,
      "step": 27380
    },
    {
      "epoch": 14.35272536687631,
      "grad_norm": 0.49680668115615845,
      "learning_rate": 1.1193567079747561e-05,
      "loss": 0.3686,
      "num_input_tokens_seen": 17914568,
      "step": 27385
    },
    {
      "epoch": 14.355345911949685,
      "grad_norm": 0.29068922996520996,
      "learning_rate": 1.1184036061404907e-05,
      "loss": 0.4238,
      "num_input_tokens_seen": 17918536,
      "step": 27390
    },
    {
      "epoch": 14.35796645702306,
      "grad_norm": 0.574885368347168,
      "learning_rate": 1.117450793320429e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 17921320,
      "step": 27395
    },
    {
      "epoch": 14.360587002096436,
      "grad_norm": 0.8617792725563049,
      "learning_rate": 1.1164982697138893e-05,
      "loss": 0.4711,
      "num_input_tokens_seen": 17924552,
      "step": 27400
    },
    {
      "epoch": 14.36320754716981,
      "grad_norm": 0.4641822874546051,
      "learning_rate": 1.1155460355201267e-05,
      "loss": 0.5182,
      "num_input_tokens_seen": 17928072,
      "step": 27405
    },
    {
      "epoch": 14.365828092243186,
      "grad_norm": 0.6805521249771118,
      "learning_rate": 1.1145940909383399e-05,
      "loss": 0.4506,
      "num_input_tokens_seen": 17930408,
      "step": 27410
    },
    {
      "epoch": 14.368448637316561,
      "grad_norm": 0.6020456552505493,
      "learning_rate": 1.1136424361676626e-05,
      "loss": 0.3819,
      "num_input_tokens_seen": 17933416,
      "step": 27415
    },
    {
      "epoch": 14.371069182389936,
      "grad_norm": 0.5715217590332031,
      "learning_rate": 1.1126910714071711e-05,
      "loss": 0.4353,
      "num_input_tokens_seen": 17936232,
      "step": 27420
    },
    {
      "epoch": 14.373689727463312,
      "grad_norm": 0.3562607765197754,
      "learning_rate": 1.1117399968558806e-05,
      "loss": 0.5741,
      "num_input_tokens_seen": 17940648,
      "step": 27425
    },
    {
      "epoch": 14.376310272536688,
      "grad_norm": 0.6655498147010803,
      "learning_rate": 1.1107892127127454e-05,
      "loss": 0.6616,
      "num_input_tokens_seen": 17943304,
      "step": 27430
    },
    {
      "epoch": 14.378930817610064,
      "grad_norm": 0.5632167458534241,
      "learning_rate": 1.109838719176658e-05,
      "loss": 0.4275,
      "num_input_tokens_seen": 17945864,
      "step": 27435
    },
    {
      "epoch": 14.381551362683439,
      "grad_norm": 0.8463792204856873,
      "learning_rate": 1.1088885164464497e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 17950152,
      "step": 27440
    },
    {
      "epoch": 14.384171907756814,
      "grad_norm": 0.3877030909061432,
      "learning_rate": 1.1079386047208932e-05,
      "loss": 0.6012,
      "num_input_tokens_seen": 17954024,
      "step": 27445
    },
    {
      "epoch": 14.38679245283019,
      "grad_norm": 0.32153138518333435,
      "learning_rate": 1.1069889841986991e-05,
      "loss": 0.3518,
      "num_input_tokens_seen": 17957896,
      "step": 27450
    },
    {
      "epoch": 14.389412997903564,
      "grad_norm": 0.3616790175437927,
      "learning_rate": 1.1060396550785182e-05,
      "loss": 0.5084,
      "num_input_tokens_seen": 17960936,
      "step": 27455
    },
    {
      "epoch": 14.39203354297694,
      "grad_norm": 1.1748427152633667,
      "learning_rate": 1.1050906175589382e-05,
      "loss": 0.4074,
      "num_input_tokens_seen": 17963496,
      "step": 27460
    },
    {
      "epoch": 14.394654088050315,
      "grad_norm": 0.71409010887146,
      "learning_rate": 1.1041418718384863e-05,
      "loss": 0.5055,
      "num_input_tokens_seen": 17966344,
      "step": 27465
    },
    {
      "epoch": 14.39727463312369,
      "grad_norm": 0.25202813744544983,
      "learning_rate": 1.1031934181156298e-05,
      "loss": 0.7085,
      "num_input_tokens_seen": 17969640,
      "step": 27470
    },
    {
      "epoch": 14.399895178197065,
      "grad_norm": 0.26887398958206177,
      "learning_rate": 1.1022452565887747e-05,
      "loss": 0.4346,
      "num_input_tokens_seen": 17973064,
      "step": 27475
    },
    {
      "epoch": 14.40251572327044,
      "grad_norm": 0.5055760145187378,
      "learning_rate": 1.1012973874562657e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 17977192,
      "step": 27480
    },
    {
      "epoch": 14.405136268343815,
      "grad_norm": 0.47805893421173096,
      "learning_rate": 1.1003498109163865e-05,
      "loss": 0.3895,
      "num_input_tokens_seen": 17981160,
      "step": 27485
    },
    {
      "epoch": 14.40775681341719,
      "grad_norm": 0.4464866816997528,
      "learning_rate": 1.0994025271673592e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 17984744,
      "step": 27490
    },
    {
      "epoch": 14.410377358490566,
      "grad_norm": 0.3968425989151001,
      "learning_rate": 1.0984555364073431e-05,
      "loss": 0.5137,
      "num_input_tokens_seen": 17988328,
      "step": 27495
    },
    {
      "epoch": 14.41299790356394,
      "grad_norm": 0.648396909236908,
      "learning_rate": 1.0975088388344393e-05,
      "loss": 0.368,
      "num_input_tokens_seen": 17991144,
      "step": 27500
    },
    {
      "epoch": 14.415618448637316,
      "grad_norm": 1.0655772686004639,
      "learning_rate": 1.096562434646686e-05,
      "loss": 0.398,
      "num_input_tokens_seen": 17994344,
      "step": 27505
    },
    {
      "epoch": 14.418238993710691,
      "grad_norm": 0.28983938694000244,
      "learning_rate": 1.0956163240420611e-05,
      "loss": 0.435,
      "num_input_tokens_seen": 17997704,
      "step": 27510
    },
    {
      "epoch": 14.420859538784066,
      "grad_norm": 0.37569740414619446,
      "learning_rate": 1.094670507218479e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 18001896,
      "step": 27515
    },
    {
      "epoch": 14.423480083857442,
      "grad_norm": 0.4730232059955597,
      "learning_rate": 1.093724984373793e-05,
      "loss": 0.5099,
      "num_input_tokens_seen": 18005288,
      "step": 27520
    },
    {
      "epoch": 14.426100628930818,
      "grad_norm": 0.4670782685279846,
      "learning_rate": 1.0927797557057968e-05,
      "loss": 0.5362,
      "num_input_tokens_seen": 18008744,
      "step": 27525
    },
    {
      "epoch": 14.428721174004194,
      "grad_norm": 0.39070427417755127,
      "learning_rate": 1.0918348214122209e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 18012712,
      "step": 27530
    },
    {
      "epoch": 14.431341719077569,
      "grad_norm": 0.4313296675682068,
      "learning_rate": 1.0908901816907366e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 18015656,
      "step": 27535
    },
    {
      "epoch": 14.433962264150944,
      "grad_norm": 0.676031231880188,
      "learning_rate": 1.08994583673895e-05,
      "loss": 0.566,
      "num_input_tokens_seen": 18018120,
      "step": 27540
    },
    {
      "epoch": 14.43658280922432,
      "grad_norm": 0.4003193974494934,
      "learning_rate": 1.0890017867544068e-05,
      "loss": 0.4556,
      "num_input_tokens_seen": 18021736,
      "step": 27545
    },
    {
      "epoch": 14.439203354297694,
      "grad_norm": 1.0428297519683838,
      "learning_rate": 1.088058031934592e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 18024584,
      "step": 27550
    },
    {
      "epoch": 14.44182389937107,
      "grad_norm": 0.7022221088409424,
      "learning_rate": 1.087114572476929e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 18027688,
      "step": 27555
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 0.6910873651504517,
      "learning_rate": 1.0861714085787783e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 18030568,
      "step": 27560
    },
    {
      "epoch": 14.44706498951782,
      "grad_norm": 0.6223630309104919,
      "learning_rate": 1.0852285404374402e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 18033128,
      "step": 27565
    },
    {
      "epoch": 14.449685534591195,
      "grad_norm": 0.5196942687034607,
      "learning_rate": 1.0842859682501505e-05,
      "loss": 0.4623,
      "num_input_tokens_seen": 18036776,
      "step": 27570
    },
    {
      "epoch": 14.45230607966457,
      "grad_norm": 0.563416600227356,
      "learning_rate": 1.0833436922140841e-05,
      "loss": 0.3741,
      "num_input_tokens_seen": 18040392,
      "step": 27575
    },
    {
      "epoch": 14.454926624737945,
      "grad_norm": 0.5103354454040527,
      "learning_rate": 1.0824017125263555e-05,
      "loss": 0.5832,
      "num_input_tokens_seen": 18043464,
      "step": 27580
    },
    {
      "epoch": 14.45754716981132,
      "grad_norm": 0.375645250082016,
      "learning_rate": 1.0814600293840154e-05,
      "loss": 0.445,
      "num_input_tokens_seen": 18046792,
      "step": 27585
    },
    {
      "epoch": 14.460167714884696,
      "grad_norm": 0.4526757001876831,
      "learning_rate": 1.080518642984055e-05,
      "loss": 0.3409,
      "num_input_tokens_seen": 18050024,
      "step": 27590
    },
    {
      "epoch": 14.46278825995807,
      "grad_norm": 0.3897675573825836,
      "learning_rate": 1.0795775535233988e-05,
      "loss": 0.4787,
      "num_input_tokens_seen": 18055624,
      "step": 27595
    },
    {
      "epoch": 14.465408805031446,
      "grad_norm": 0.43798112869262695,
      "learning_rate": 1.0786367611989145e-05,
      "loss": 0.2954,
      "num_input_tokens_seen": 18060712,
      "step": 27600
    },
    {
      "epoch": 14.468029350104821,
      "grad_norm": 0.36692216992378235,
      "learning_rate": 1.0776962662074031e-05,
      "loss": 0.5485,
      "num_input_tokens_seen": 18063176,
      "step": 27605
    },
    {
      "epoch": 14.470649895178196,
      "grad_norm": 0.46939289569854736,
      "learning_rate": 1.0767560687456062e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 18066696,
      "step": 27610
    },
    {
      "epoch": 14.473270440251572,
      "grad_norm": 0.5691983103752136,
      "learning_rate": 1.0758161690102036e-05,
      "loss": 0.423,
      "num_input_tokens_seen": 18069640,
      "step": 27615
    },
    {
      "epoch": 14.475890985324948,
      "grad_norm": 0.560189425945282,
      "learning_rate": 1.0748765671978095e-05,
      "loss": 0.3793,
      "num_input_tokens_seen": 18072232,
      "step": 27620
    },
    {
      "epoch": 14.478511530398324,
      "grad_norm": 0.7333765029907227,
      "learning_rate": 1.0739372635049797e-05,
      "loss": 0.4193,
      "num_input_tokens_seen": 18074760,
      "step": 27625
    },
    {
      "epoch": 14.481132075471699,
      "grad_norm": 0.25286591053009033,
      "learning_rate": 1.0729982581282042e-05,
      "loss": 0.3076,
      "num_input_tokens_seen": 18078056,
      "step": 27630
    },
    {
      "epoch": 14.483752620545074,
      "grad_norm": 0.49415937066078186,
      "learning_rate": 1.0720595512639129e-05,
      "loss": 0.4428,
      "num_input_tokens_seen": 18081672,
      "step": 27635
    },
    {
      "epoch": 14.48637316561845,
      "grad_norm": 0.4229891896247864,
      "learning_rate": 1.0711211431084728e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 18085736,
      "step": 27640
    },
    {
      "epoch": 14.488993710691824,
      "grad_norm": 0.39493778347969055,
      "learning_rate": 1.0701830338581887e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 18089608,
      "step": 27645
    },
    {
      "epoch": 14.4916142557652,
      "grad_norm": 0.7141770720481873,
      "learning_rate": 1.069245223709301e-05,
      "loss": 0.6312,
      "num_input_tokens_seen": 18092424,
      "step": 27650
    },
    {
      "epoch": 14.494234800838575,
      "grad_norm": 0.43734127283096313,
      "learning_rate": 1.0683077128579904e-05,
      "loss": 0.5415,
      "num_input_tokens_seen": 18095848,
      "step": 27655
    },
    {
      "epoch": 14.49685534591195,
      "grad_norm": 0.467922180891037,
      "learning_rate": 1.0673705015003718e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 18099080,
      "step": 27660
    },
    {
      "epoch": 14.499475890985325,
      "grad_norm": 0.5439822673797607,
      "learning_rate": 1.0664335898325002e-05,
      "loss": 0.3996,
      "num_input_tokens_seen": 18101640,
      "step": 27665
    },
    {
      "epoch": 14.5020964360587,
      "grad_norm": 0.6426020860671997,
      "learning_rate": 1.065496978050367e-05,
      "loss": 0.3904,
      "num_input_tokens_seen": 18104840,
      "step": 27670
    },
    {
      "epoch": 14.504716981132075,
      "grad_norm": 0.6960967183113098,
      "learning_rate": 1.0645606663499e-05,
      "loss": 0.426,
      "num_input_tokens_seen": 18107240,
      "step": 27675
    },
    {
      "epoch": 14.50733752620545,
      "grad_norm": 0.6785925626754761,
      "learning_rate": 1.0636246549269649e-05,
      "loss": 0.5529,
      "num_input_tokens_seen": 18110664,
      "step": 27680
    },
    {
      "epoch": 14.509958071278826,
      "grad_norm": 0.456602543592453,
      "learning_rate": 1.062688943977366e-05,
      "loss": 0.488,
      "num_input_tokens_seen": 18114120,
      "step": 27685
    },
    {
      "epoch": 14.5125786163522,
      "grad_norm": 0.7345374822616577,
      "learning_rate": 1.0617535336968413e-05,
      "loss": 0.4914,
      "num_input_tokens_seen": 18117032,
      "step": 27690
    },
    {
      "epoch": 14.515199161425576,
      "grad_norm": 0.4528736174106598,
      "learning_rate": 1.0608184242810698e-05,
      "loss": 0.4448,
      "num_input_tokens_seen": 18120072,
      "step": 27695
    },
    {
      "epoch": 14.517819706498951,
      "grad_norm": 1.047282338142395,
      "learning_rate": 1.0598836159256642e-05,
      "loss": 0.5667,
      "num_input_tokens_seen": 18122824,
      "step": 27700
    },
    {
      "epoch": 14.520440251572326,
      "grad_norm": 0.4494263231754303,
      "learning_rate": 1.0589491088261763e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 18126120,
      "step": 27705
    },
    {
      "epoch": 14.523060796645701,
      "grad_norm": 0.5240842700004578,
      "learning_rate": 1.0580149031780945e-05,
      "loss": 0.4406,
      "num_input_tokens_seen": 18129320,
      "step": 27710
    },
    {
      "epoch": 14.525681341719078,
      "grad_norm": 0.5865836143493652,
      "learning_rate": 1.0570809991768447e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 18132648,
      "step": 27715
    },
    {
      "epoch": 14.528301886792454,
      "grad_norm": 0.4768258333206177,
      "learning_rate": 1.0561473970177873e-05,
      "loss": 0.3973,
      "num_input_tokens_seen": 18135432,
      "step": 27720
    },
    {
      "epoch": 14.530922431865829,
      "grad_norm": 0.6053224205970764,
      "learning_rate": 1.0552140968962232e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 18138536,
      "step": 27725
    },
    {
      "epoch": 14.533542976939204,
      "grad_norm": 0.6735681295394897,
      "learning_rate": 1.054281099007386e-05,
      "loss": 0.3083,
      "num_input_tokens_seen": 18142568,
      "step": 27730
    },
    {
      "epoch": 14.536163522012579,
      "grad_norm": 0.45250797271728516,
      "learning_rate": 1.0533484035464492e-05,
      "loss": 0.4821,
      "num_input_tokens_seen": 18145352,
      "step": 27735
    },
    {
      "epoch": 14.538784067085954,
      "grad_norm": 0.42084282636642456,
      "learning_rate": 1.0524160107085231e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 18152584,
      "step": 27740
    },
    {
      "epoch": 14.54140461215933,
      "grad_norm": 0.6245284080505371,
      "learning_rate": 1.0514839206886518e-05,
      "loss": 0.4425,
      "num_input_tokens_seen": 18155624,
      "step": 27745
    },
    {
      "epoch": 14.544025157232705,
      "grad_norm": 0.6788398027420044,
      "learning_rate": 1.0505521336818195e-05,
      "loss": 0.4027,
      "num_input_tokens_seen": 18158728,
      "step": 27750
    },
    {
      "epoch": 14.54664570230608,
      "grad_norm": 0.7018723487854004,
      "learning_rate": 1.0496206498829438e-05,
      "loss": 0.3816,
      "num_input_tokens_seen": 18161512,
      "step": 27755
    },
    {
      "epoch": 14.549266247379455,
      "grad_norm": 0.4162730872631073,
      "learning_rate": 1.0486894694868814e-05,
      "loss": 0.4802,
      "num_input_tokens_seen": 18164616,
      "step": 27760
    },
    {
      "epoch": 14.55188679245283,
      "grad_norm": 0.7913058996200562,
      "learning_rate": 1.0477585926884246e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 18167080,
      "step": 27765
    },
    {
      "epoch": 14.554507337526205,
      "grad_norm": 0.2537243366241455,
      "learning_rate": 1.046828019682303e-05,
      "loss": 0.3985,
      "num_input_tokens_seen": 18170792,
      "step": 27770
    },
    {
      "epoch": 14.55712788259958,
      "grad_norm": 0.9141443967819214,
      "learning_rate": 1.0458977506631808e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 18173960,
      "step": 27775
    },
    {
      "epoch": 14.559748427672956,
      "grad_norm": 0.3967131972312927,
      "learning_rate": 1.0449677858256593e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 18177864,
      "step": 27780
    },
    {
      "epoch": 14.56236897274633,
      "grad_norm": 0.47724828124046326,
      "learning_rate": 1.0440381253642773e-05,
      "loss": 0.4825,
      "num_input_tokens_seen": 18181672,
      "step": 27785
    },
    {
      "epoch": 14.564989517819706,
      "grad_norm": 0.5370535254478455,
      "learning_rate": 1.0431087694735087e-05,
      "loss": 0.3825,
      "num_input_tokens_seen": 18185000,
      "step": 27790
    },
    {
      "epoch": 14.567610062893081,
      "grad_norm": 0.3822701871395111,
      "learning_rate": 1.0421797183477644e-05,
      "loss": 0.4133,
      "num_input_tokens_seen": 18188104,
      "step": 27795
    },
    {
      "epoch": 14.570230607966456,
      "grad_norm": 0.23731917142868042,
      "learning_rate": 1.041250972181392e-05,
      "loss": 0.4997,
      "num_input_tokens_seen": 18192872,
      "step": 27800
    },
    {
      "epoch": 14.572851153039831,
      "grad_norm": 0.5058792233467102,
      "learning_rate": 1.0403225311686743e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 18196040,
      "step": 27805
    },
    {
      "epoch": 14.575471698113208,
      "grad_norm": 0.34799444675445557,
      "learning_rate": 1.0393943955038293e-05,
      "loss": 0.4027,
      "num_input_tokens_seen": 18199592,
      "step": 27810
    },
    {
      "epoch": 14.578092243186584,
      "grad_norm": 0.3116770386695862,
      "learning_rate": 1.038466565381013e-05,
      "loss": 0.4828,
      "num_input_tokens_seen": 18203016,
      "step": 27815
    },
    {
      "epoch": 14.580712788259959,
      "grad_norm": 0.31079137325286865,
      "learning_rate": 1.0375390409943175e-05,
      "loss": 0.4269,
      "num_input_tokens_seen": 18206152,
      "step": 27820
    },
    {
      "epoch": 14.583333333333334,
      "grad_norm": 0.42206448316574097,
      "learning_rate": 1.0366118225377708e-05,
      "loss": 0.3452,
      "num_input_tokens_seen": 18209096,
      "step": 27825
    },
    {
      "epoch": 14.585953878406709,
      "grad_norm": 0.30931755900382996,
      "learning_rate": 1.0356849102053353e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 18212680,
      "step": 27830
    },
    {
      "epoch": 14.588574423480084,
      "grad_norm": 0.5409465432167053,
      "learning_rate": 1.0347583041909104e-05,
      "loss": 0.3769,
      "num_input_tokens_seen": 18215720,
      "step": 27835
    },
    {
      "epoch": 14.59119496855346,
      "grad_norm": 0.4917721450328827,
      "learning_rate": 1.0338320046883315e-05,
      "loss": 0.426,
      "num_input_tokens_seen": 18218888,
      "step": 27840
    },
    {
      "epoch": 14.593815513626835,
      "grad_norm": 0.4063382148742676,
      "learning_rate": 1.0329060118913705e-05,
      "loss": 0.3712,
      "num_input_tokens_seen": 18221768,
      "step": 27845
    },
    {
      "epoch": 14.59643605870021,
      "grad_norm": 0.761986494064331,
      "learning_rate": 1.0319803259937353e-05,
      "loss": 0.41,
      "num_input_tokens_seen": 18225640,
      "step": 27850
    },
    {
      "epoch": 14.599056603773585,
      "grad_norm": 0.2860598862171173,
      "learning_rate": 1.0310549471890676e-05,
      "loss": 0.4755,
      "num_input_tokens_seen": 18229928,
      "step": 27855
    },
    {
      "epoch": 14.60167714884696,
      "grad_norm": 0.6477767825126648,
      "learning_rate": 1.0301298756709454e-05,
      "loss": 0.557,
      "num_input_tokens_seen": 18232840,
      "step": 27860
    },
    {
      "epoch": 14.604297693920335,
      "grad_norm": 0.5851904153823853,
      "learning_rate": 1.0292051116328844e-05,
      "loss": 0.4288,
      "num_input_tokens_seen": 18236200,
      "step": 27865
    },
    {
      "epoch": 14.60691823899371,
      "grad_norm": 0.573104977607727,
      "learning_rate": 1.028280655268334e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 18240072,
      "step": 27870
    },
    {
      "epoch": 14.609538784067086,
      "grad_norm": 0.4882412254810333,
      "learning_rate": 1.0273565067706806e-05,
      "loss": 0.4593,
      "num_input_tokens_seen": 18243464,
      "step": 27875
    },
    {
      "epoch": 14.61215932914046,
      "grad_norm": 0.36661437153816223,
      "learning_rate": 1.0264326663332457e-05,
      "loss": 0.5216,
      "num_input_tokens_seen": 18246088,
      "step": 27880
    },
    {
      "epoch": 14.614779874213836,
      "grad_norm": 0.5495105981826782,
      "learning_rate": 1.0255091341492859e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 18249512,
      "step": 27885
    },
    {
      "epoch": 14.617400419287211,
      "grad_norm": 0.4379930794239044,
      "learning_rate": 1.0245859104119923e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 18252328,
      "step": 27890
    },
    {
      "epoch": 14.620020964360586,
      "grad_norm": 0.8412379622459412,
      "learning_rate": 1.0236629953144938e-05,
      "loss": 0.3665,
      "num_input_tokens_seen": 18256168,
      "step": 27895
    },
    {
      "epoch": 14.622641509433961,
      "grad_norm": 0.6237373948097229,
      "learning_rate": 1.0227403890498537e-05,
      "loss": 0.5646,
      "num_input_tokens_seen": 18259208,
      "step": 27900
    },
    {
      "epoch": 14.625262054507338,
      "grad_norm": 0.569580078125,
      "learning_rate": 1.021818091811072e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 18261992,
      "step": 27905
    },
    {
      "epoch": 14.627882599580714,
      "grad_norm": 0.44078564643859863,
      "learning_rate": 1.0208961037910813e-05,
      "loss": 0.4638,
      "num_input_tokens_seen": 18265384,
      "step": 27910
    },
    {
      "epoch": 14.630503144654089,
      "grad_norm": 0.37057530879974365,
      "learning_rate": 1.0199744251827503e-05,
      "loss": 0.4603,
      "num_input_tokens_seen": 18268008,
      "step": 27915
    },
    {
      "epoch": 14.633123689727464,
      "grad_norm": 0.5803650617599487,
      "learning_rate": 1.0190530561788845e-05,
      "loss": 0.5542,
      "num_input_tokens_seen": 18270920,
      "step": 27920
    },
    {
      "epoch": 14.635744234800839,
      "grad_norm": 0.6555641889572144,
      "learning_rate": 1.0181319969722236e-05,
      "loss": 0.294,
      "num_input_tokens_seen": 18275112,
      "step": 27925
    },
    {
      "epoch": 14.638364779874214,
      "grad_norm": 0.4058855473995209,
      "learning_rate": 1.0172112477554441e-05,
      "loss": 0.3308,
      "num_input_tokens_seen": 18278888,
      "step": 27930
    },
    {
      "epoch": 14.64098532494759,
      "grad_norm": 0.3591873347759247,
      "learning_rate": 1.016290808721154e-05,
      "loss": 0.3863,
      "num_input_tokens_seen": 18281768,
      "step": 27935
    },
    {
      "epoch": 14.643605870020965,
      "grad_norm": 0.5367789268493652,
      "learning_rate": 1.0153706800619007e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 18284360,
      "step": 27940
    },
    {
      "epoch": 14.64622641509434,
      "grad_norm": 0.6923370957374573,
      "learning_rate": 1.0144508619701623e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 18287048,
      "step": 27945
    },
    {
      "epoch": 14.648846960167715,
      "grad_norm": 0.5847424268722534,
      "learning_rate": 1.013531354638356e-05,
      "loss": 0.5674,
      "num_input_tokens_seen": 18289672,
      "step": 27950
    },
    {
      "epoch": 14.65146750524109,
      "grad_norm": 0.39150500297546387,
      "learning_rate": 1.0126121582588316e-05,
      "loss": 0.5416,
      "num_input_tokens_seen": 18293096,
      "step": 27955
    },
    {
      "epoch": 14.654088050314465,
      "grad_norm": 0.24822433292865753,
      "learning_rate": 1.0116932730238754e-05,
      "loss": 0.346,
      "num_input_tokens_seen": 18297608,
      "step": 27960
    },
    {
      "epoch": 14.65670859538784,
      "grad_norm": 0.41876736283302307,
      "learning_rate": 1.0107746991257063e-05,
      "loss": 0.4132,
      "num_input_tokens_seen": 18301352,
      "step": 27965
    },
    {
      "epoch": 14.659329140461216,
      "grad_norm": 0.7503466010093689,
      "learning_rate": 1.0098564367564811e-05,
      "loss": 0.3708,
      "num_input_tokens_seen": 18304552,
      "step": 27970
    },
    {
      "epoch": 14.66194968553459,
      "grad_norm": 0.33598601818084717,
      "learning_rate": 1.0089384861082876e-05,
      "loss": 0.3368,
      "num_input_tokens_seen": 18307400,
      "step": 27975
    },
    {
      "epoch": 14.664570230607966,
      "grad_norm": 1.1122453212738037,
      "learning_rate": 1.0080208473731523e-05,
      "loss": 0.8121,
      "num_input_tokens_seen": 18311432,
      "step": 27980
    },
    {
      "epoch": 14.667190775681341,
      "grad_norm": 0.5119360089302063,
      "learning_rate": 1.0071035207430352e-05,
      "loss": 0.386,
      "num_input_tokens_seen": 18315720,
      "step": 27985
    },
    {
      "epoch": 14.669811320754716,
      "grad_norm": 0.3395081162452698,
      "learning_rate": 1.0061865064098285e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 18318952,
      "step": 27990
    },
    {
      "epoch": 14.672431865828091,
      "grad_norm": 1.300952672958374,
      "learning_rate": 1.0052698045653625e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 18322856,
      "step": 27995
    },
    {
      "epoch": 14.675052410901468,
      "grad_norm": 0.3368455767631531,
      "learning_rate": 1.0043534154014015e-05,
      "loss": 0.3717,
      "num_input_tokens_seen": 18326248,
      "step": 28000
    },
    {
      "epoch": 14.677672955974844,
      "grad_norm": 0.6262207627296448,
      "learning_rate": 1.003437339109642e-05,
      "loss": 0.5139,
      "num_input_tokens_seen": 18329704,
      "step": 28005
    },
    {
      "epoch": 14.680293501048219,
      "grad_norm": 0.612956166267395,
      "learning_rate": 1.0025215758817183e-05,
      "loss": 0.4821,
      "num_input_tokens_seen": 18333160,
      "step": 28010
    },
    {
      "epoch": 14.682914046121594,
      "grad_norm": 1.30815589427948,
      "learning_rate": 1.0016061259091958e-05,
      "loss": 0.5515,
      "num_input_tokens_seen": 18336232,
      "step": 28015
    },
    {
      "epoch": 14.685534591194969,
      "grad_norm": 0.9016631841659546,
      "learning_rate": 1.000690989383577e-05,
      "loss": 0.3563,
      "num_input_tokens_seen": 18339304,
      "step": 28020
    },
    {
      "epoch": 14.688155136268344,
      "grad_norm": 0.5485749244689941,
      "learning_rate": 9.997761664962993e-06,
      "loss": 0.4267,
      "num_input_tokens_seen": 18342536,
      "step": 28025
    },
    {
      "epoch": 14.69077568134172,
      "grad_norm": 0.7113211154937744,
      "learning_rate": 9.988616574387314e-06,
      "loss": 0.3951,
      "num_input_tokens_seen": 18345448,
      "step": 28030
    },
    {
      "epoch": 14.693396226415095,
      "grad_norm": 0.40612322092056274,
      "learning_rate": 9.979474624021787e-06,
      "loss": 0.3815,
      "num_input_tokens_seen": 18348296,
      "step": 28035
    },
    {
      "epoch": 14.69601677148847,
      "grad_norm": 0.5279413461685181,
      "learning_rate": 9.970335815778815e-06,
      "loss": 0.4811,
      "num_input_tokens_seen": 18351656,
      "step": 28040
    },
    {
      "epoch": 14.698637316561845,
      "grad_norm": 0.467006117105484,
      "learning_rate": 9.96120015157011e-06,
      "loss": 0.4786,
      "num_input_tokens_seen": 18355016,
      "step": 28045
    },
    {
      "epoch": 14.70125786163522,
      "grad_norm": 0.31920379400253296,
      "learning_rate": 9.952067633306767e-06,
      "loss": 0.5349,
      "num_input_tokens_seen": 18357800,
      "step": 28050
    },
    {
      "epoch": 14.703878406708595,
      "grad_norm": 0.6227661967277527,
      "learning_rate": 9.942938262899204e-06,
      "loss": 0.5634,
      "num_input_tokens_seen": 18360904,
      "step": 28055
    },
    {
      "epoch": 14.70649895178197,
      "grad_norm": 0.4373205304145813,
      "learning_rate": 9.93381204225717e-06,
      "loss": 0.4947,
      "num_input_tokens_seen": 18363528,
      "step": 28060
    },
    {
      "epoch": 14.709119496855346,
      "grad_norm": 0.3000672459602356,
      "learning_rate": 9.924688973289782e-06,
      "loss": 0.4142,
      "num_input_tokens_seen": 18367688,
      "step": 28065
    },
    {
      "epoch": 14.71174004192872,
      "grad_norm": 0.34014925360679626,
      "learning_rate": 9.915569057905463e-06,
      "loss": 0.5212,
      "num_input_tokens_seen": 18370824,
      "step": 28070
    },
    {
      "epoch": 14.714360587002096,
      "grad_norm": 0.407189279794693,
      "learning_rate": 9.906452298012003e-06,
      "loss": 0.4275,
      "num_input_tokens_seen": 18373320,
      "step": 28075
    },
    {
      "epoch": 14.716981132075471,
      "grad_norm": 0.7871357202529907,
      "learning_rate": 9.897338695516529e-06,
      "loss": 0.5276,
      "num_input_tokens_seen": 18376360,
      "step": 28080
    },
    {
      "epoch": 14.719601677148846,
      "grad_norm": 0.32619282603263855,
      "learning_rate": 9.888228252325507e-06,
      "loss": 0.5013,
      "num_input_tokens_seen": 18382984,
      "step": 28085
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 0.5053386092185974,
      "learning_rate": 9.87912097034473e-06,
      "loss": 0.4065,
      "num_input_tokens_seen": 18387016,
      "step": 28090
    },
    {
      "epoch": 14.724842767295598,
      "grad_norm": 0.58530592918396,
      "learning_rate": 9.870016851479327e-06,
      "loss": 0.4194,
      "num_input_tokens_seen": 18390120,
      "step": 28095
    },
    {
      "epoch": 14.727463312368974,
      "grad_norm": 0.47792109847068787,
      "learning_rate": 9.860915897633791e-06,
      "loss": 0.4735,
      "num_input_tokens_seen": 18392936,
      "step": 28100
    },
    {
      "epoch": 14.730083857442349,
      "grad_norm": 0.41036102175712585,
      "learning_rate": 9.85181811071193e-06,
      "loss": 0.4699,
      "num_input_tokens_seen": 18395720,
      "step": 28105
    },
    {
      "epoch": 14.732704402515724,
      "grad_norm": 0.46315643191337585,
      "learning_rate": 9.842723492616913e-06,
      "loss": 0.3762,
      "num_input_tokens_seen": 18398312,
      "step": 28110
    },
    {
      "epoch": 14.735324947589099,
      "grad_norm": 0.43741312623023987,
      "learning_rate": 9.833632045251207e-06,
      "loss": 0.4225,
      "num_input_tokens_seen": 18401608,
      "step": 28115
    },
    {
      "epoch": 14.737945492662474,
      "grad_norm": 0.4869987666606903,
      "learning_rate": 9.824543770516656e-06,
      "loss": 0.506,
      "num_input_tokens_seen": 18404424,
      "step": 28120
    },
    {
      "epoch": 14.74056603773585,
      "grad_norm": 0.30837956070899963,
      "learning_rate": 9.815458670314412e-06,
      "loss": 0.4077,
      "num_input_tokens_seen": 18408040,
      "step": 28125
    },
    {
      "epoch": 14.743186582809225,
      "grad_norm": 0.49472135305404663,
      "learning_rate": 9.80637674654498e-06,
      "loss": 0.4561,
      "num_input_tokens_seen": 18410792,
      "step": 28130
    },
    {
      "epoch": 14.7458071278826,
      "grad_norm": 0.6297508478164673,
      "learning_rate": 9.797298001108193e-06,
      "loss": 0.4399,
      "num_input_tokens_seen": 18413320,
      "step": 28135
    },
    {
      "epoch": 14.748427672955975,
      "grad_norm": 0.7382582426071167,
      "learning_rate": 9.788222435903232e-06,
      "loss": 0.4155,
      "num_input_tokens_seen": 18416200,
      "step": 28140
    },
    {
      "epoch": 14.75104821802935,
      "grad_norm": 1.2294433116912842,
      "learning_rate": 9.779150052828596e-06,
      "loss": 0.5986,
      "num_input_tokens_seen": 18418760,
      "step": 28145
    },
    {
      "epoch": 14.753668763102725,
      "grad_norm": 0.7370238900184631,
      "learning_rate": 9.770080853782108e-06,
      "loss": 0.549,
      "num_input_tokens_seen": 18421832,
      "step": 28150
    },
    {
      "epoch": 14.7562893081761,
      "grad_norm": 0.5282382369041443,
      "learning_rate": 9.76101484066095e-06,
      "loss": 0.3529,
      "num_input_tokens_seen": 18425608,
      "step": 28155
    },
    {
      "epoch": 14.758909853249476,
      "grad_norm": 0.43193700909614563,
      "learning_rate": 9.751952015361637e-06,
      "loss": 0.4281,
      "num_input_tokens_seen": 18429256,
      "step": 28160
    },
    {
      "epoch": 14.76153039832285,
      "grad_norm": 0.4285932183265686,
      "learning_rate": 9.742892379780003e-06,
      "loss": 0.4063,
      "num_input_tokens_seen": 18434600,
      "step": 28165
    },
    {
      "epoch": 14.764150943396226,
      "grad_norm": 0.9052885174751282,
      "learning_rate": 9.733835935811237e-06,
      "loss": 0.4396,
      "num_input_tokens_seen": 18437416,
      "step": 28170
    },
    {
      "epoch": 14.766771488469601,
      "grad_norm": 0.541107714176178,
      "learning_rate": 9.724782685349809e-06,
      "loss": 0.4208,
      "num_input_tokens_seen": 18440776,
      "step": 28175
    },
    {
      "epoch": 14.769392033542976,
      "grad_norm": 0.45557188987731934,
      "learning_rate": 9.715732630289576e-06,
      "loss": 0.3255,
      "num_input_tokens_seen": 18443528,
      "step": 28180
    },
    {
      "epoch": 14.772012578616351,
      "grad_norm": 0.4915337562561035,
      "learning_rate": 9.706685772523701e-06,
      "loss": 0.4094,
      "num_input_tokens_seen": 18446504,
      "step": 28185
    },
    {
      "epoch": 14.774633123689728,
      "grad_norm": 0.3185359835624695,
      "learning_rate": 9.697642113944692e-06,
      "loss": 0.4193,
      "num_input_tokens_seen": 18449864,
      "step": 28190
    },
    {
      "epoch": 14.777253668763104,
      "grad_norm": 0.4887719452381134,
      "learning_rate": 9.688601656444381e-06,
      "loss": 0.4088,
      "num_input_tokens_seen": 18455816,
      "step": 28195
    },
    {
      "epoch": 14.779874213836479,
      "grad_norm": 0.6022867560386658,
      "learning_rate": 9.679564401913921e-06,
      "loss": 0.3771,
      "num_input_tokens_seen": 18459144,
      "step": 28200
    },
    {
      "epoch": 14.782494758909854,
      "grad_norm": 0.42504096031188965,
      "learning_rate": 9.670530352243798e-06,
      "loss": 0.5073,
      "num_input_tokens_seen": 18462056,
      "step": 28205
    },
    {
      "epoch": 14.785115303983229,
      "grad_norm": 0.49150383472442627,
      "learning_rate": 9.661499509323835e-06,
      "loss": 0.5644,
      "num_input_tokens_seen": 18465352,
      "step": 28210
    },
    {
      "epoch": 14.787735849056604,
      "grad_norm": 0.4360470175743103,
      "learning_rate": 9.652471875043182e-06,
      "loss": 0.5483,
      "num_input_tokens_seen": 18468776,
      "step": 28215
    },
    {
      "epoch": 14.79035639412998,
      "grad_norm": 0.2999735176563263,
      "learning_rate": 9.64344745129033e-06,
      "loss": 0.5024,
      "num_input_tokens_seen": 18472296,
      "step": 28220
    },
    {
      "epoch": 14.792976939203355,
      "grad_norm": 0.4227762818336487,
      "learning_rate": 9.634426239953073e-06,
      "loss": 0.5176,
      "num_input_tokens_seen": 18476424,
      "step": 28225
    },
    {
      "epoch": 14.79559748427673,
      "grad_norm": 0.42287471890449524,
      "learning_rate": 9.62540824291854e-06,
      "loss": 0.3979,
      "num_input_tokens_seen": 18479912,
      "step": 28230
    },
    {
      "epoch": 14.798218029350105,
      "grad_norm": 0.2988883852958679,
      "learning_rate": 9.616393462073197e-06,
      "loss": 0.3427,
      "num_input_tokens_seen": 18482632,
      "step": 28235
    },
    {
      "epoch": 14.80083857442348,
      "grad_norm": 0.35356420278549194,
      "learning_rate": 9.607381899302839e-06,
      "loss": 0.3981,
      "num_input_tokens_seen": 18485928,
      "step": 28240
    },
    {
      "epoch": 14.803459119496855,
      "grad_norm": 0.3610307574272156,
      "learning_rate": 9.59837355649258e-06,
      "loss": 0.5111,
      "num_input_tokens_seen": 18489768,
      "step": 28245
    },
    {
      "epoch": 14.80607966457023,
      "grad_norm": 0.6785309910774231,
      "learning_rate": 9.589368435526868e-06,
      "loss": 0.4733,
      "num_input_tokens_seen": 18492968,
      "step": 28250
    },
    {
      "epoch": 14.808700209643606,
      "grad_norm": 0.7447938919067383,
      "learning_rate": 9.580366538289468e-06,
      "loss": 0.5334,
      "num_input_tokens_seen": 18496136,
      "step": 28255
    },
    {
      "epoch": 14.81132075471698,
      "grad_norm": 0.35439932346343994,
      "learning_rate": 9.57136786666346e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 18499880,
      "step": 28260
    },
    {
      "epoch": 14.813941299790356,
      "grad_norm": 0.5309034585952759,
      "learning_rate": 9.562372422531276e-06,
      "loss": 0.544,
      "num_input_tokens_seen": 18504488,
      "step": 28265
    },
    {
      "epoch": 14.816561844863731,
      "grad_norm": 0.3335842788219452,
      "learning_rate": 9.553380207774662e-06,
      "loss": 0.3724,
      "num_input_tokens_seen": 18507304,
      "step": 28270
    },
    {
      "epoch": 14.819182389937106,
      "grad_norm": 0.6090202927589417,
      "learning_rate": 9.54439122427469e-06,
      "loss": 0.4362,
      "num_input_tokens_seen": 18510152,
      "step": 28275
    },
    {
      "epoch": 14.821802935010481,
      "grad_norm": 0.4937222898006439,
      "learning_rate": 9.535405473911748e-06,
      "loss": 0.4727,
      "num_input_tokens_seen": 18513960,
      "step": 28280
    },
    {
      "epoch": 14.824423480083858,
      "grad_norm": 0.4678272306919098,
      "learning_rate": 9.526422958565546e-06,
      "loss": 0.5704,
      "num_input_tokens_seen": 18516520,
      "step": 28285
    },
    {
      "epoch": 14.827044025157234,
      "grad_norm": 0.6068041920661926,
      "learning_rate": 9.517443680115124e-06,
      "loss": 0.3917,
      "num_input_tokens_seen": 18519336,
      "step": 28290
    },
    {
      "epoch": 14.829664570230609,
      "grad_norm": 0.4957093894481659,
      "learning_rate": 9.50846764043885e-06,
      "loss": 0.4606,
      "num_input_tokens_seen": 18522024,
      "step": 28295
    },
    {
      "epoch": 14.832285115303984,
      "grad_norm": 0.627527117729187,
      "learning_rate": 9.499494841414422e-06,
      "loss": 0.5733,
      "num_input_tokens_seen": 18526504,
      "step": 28300
    },
    {
      "epoch": 14.834905660377359,
      "grad_norm": 0.386987566947937,
      "learning_rate": 9.490525284918822e-06,
      "loss": 0.4164,
      "num_input_tokens_seen": 18529736,
      "step": 28305
    },
    {
      "epoch": 14.837526205450734,
      "grad_norm": 0.5169863104820251,
      "learning_rate": 9.4815589728284e-06,
      "loss": 0.4498,
      "num_input_tokens_seen": 18532648,
      "step": 28310
    },
    {
      "epoch": 14.84014675052411,
      "grad_norm": 0.576026201248169,
      "learning_rate": 9.472595907018788e-06,
      "loss": 0.4384,
      "num_input_tokens_seen": 18535112,
      "step": 28315
    },
    {
      "epoch": 14.842767295597485,
      "grad_norm": 0.3266207277774811,
      "learning_rate": 9.463636089364968e-06,
      "loss": 0.4371,
      "num_input_tokens_seen": 18538376,
      "step": 28320
    },
    {
      "epoch": 14.84538784067086,
      "grad_norm": 0.3981381356716156,
      "learning_rate": 9.454679521741239e-06,
      "loss": 0.4744,
      "num_input_tokens_seen": 18541864,
      "step": 28325
    },
    {
      "epoch": 14.848008385744235,
      "grad_norm": 0.46694856882095337,
      "learning_rate": 9.445726206021194e-06,
      "loss": 0.4928,
      "num_input_tokens_seen": 18545320,
      "step": 28330
    },
    {
      "epoch": 14.85062893081761,
      "grad_norm": 0.9379913806915283,
      "learning_rate": 9.436776144077772e-06,
      "loss": 0.4763,
      "num_input_tokens_seen": 18547720,
      "step": 28335
    },
    {
      "epoch": 14.853249475890985,
      "grad_norm": 0.7325523495674133,
      "learning_rate": 9.427829337783237e-06,
      "loss": 0.4765,
      "num_input_tokens_seen": 18551880,
      "step": 28340
    },
    {
      "epoch": 14.85587002096436,
      "grad_norm": 0.5420301556587219,
      "learning_rate": 9.418885789009138e-06,
      "loss": 0.2964,
      "num_input_tokens_seen": 18554568,
      "step": 28345
    },
    {
      "epoch": 14.858490566037736,
      "grad_norm": 0.36649179458618164,
      "learning_rate": 9.409945499626374e-06,
      "loss": 0.5223,
      "num_input_tokens_seen": 18557448,
      "step": 28350
    },
    {
      "epoch": 14.86111111111111,
      "grad_norm": 1.206765055656433,
      "learning_rate": 9.40100847150516e-06,
      "loss": 0.4812,
      "num_input_tokens_seen": 18559656,
      "step": 28355
    },
    {
      "epoch": 14.863731656184486,
      "grad_norm": 0.3473484516143799,
      "learning_rate": 9.392074706515003e-06,
      "loss": 0.4279,
      "num_input_tokens_seen": 18563752,
      "step": 28360
    },
    {
      "epoch": 14.866352201257861,
      "grad_norm": 0.38256198167800903,
      "learning_rate": 9.383144206524753e-06,
      "loss": 0.4777,
      "num_input_tokens_seen": 18566696,
      "step": 28365
    },
    {
      "epoch": 14.868972746331236,
      "grad_norm": 0.33144277334213257,
      "learning_rate": 9.374216973402575e-06,
      "loss": 0.4809,
      "num_input_tokens_seen": 18570280,
      "step": 28370
    },
    {
      "epoch": 14.871593291404611,
      "grad_norm": 0.5061672925949097,
      "learning_rate": 9.365293009015926e-06,
      "loss": 0.3685,
      "num_input_tokens_seen": 18573416,
      "step": 28375
    },
    {
      "epoch": 14.874213836477988,
      "grad_norm": 0.44210898876190186,
      "learning_rate": 9.356372315231623e-06,
      "loss": 0.3057,
      "num_input_tokens_seen": 18576712,
      "step": 28380
    },
    {
      "epoch": 14.876834381551364,
      "grad_norm": 0.6422157883644104,
      "learning_rate": 9.347454893915749e-06,
      "loss": 0.4653,
      "num_input_tokens_seen": 18579912,
      "step": 28385
    },
    {
      "epoch": 14.879454926624739,
      "grad_norm": 0.7161491513252258,
      "learning_rate": 9.338540746933741e-06,
      "loss": 0.4582,
      "num_input_tokens_seen": 18582664,
      "step": 28390
    },
    {
      "epoch": 14.882075471698114,
      "grad_norm": 0.6659968495368958,
      "learning_rate": 9.329629876150339e-06,
      "loss": 0.4402,
      "num_input_tokens_seen": 18585704,
      "step": 28395
    },
    {
      "epoch": 14.884696016771489,
      "grad_norm": 0.8158625960350037,
      "learning_rate": 9.320722283429581e-06,
      "loss": 0.3005,
      "num_input_tokens_seen": 18588456,
      "step": 28400
    },
    {
      "epoch": 14.887316561844864,
      "grad_norm": 0.6022749543190002,
      "learning_rate": 9.311817970634854e-06,
      "loss": 0.3808,
      "num_input_tokens_seen": 18591848,
      "step": 28405
    },
    {
      "epoch": 14.88993710691824,
      "grad_norm": 0.6775140762329102,
      "learning_rate": 9.302916939628814e-06,
      "loss": 0.5669,
      "num_input_tokens_seen": 18594440,
      "step": 28410
    },
    {
      "epoch": 14.892557651991615,
      "grad_norm": 0.6996208429336548,
      "learning_rate": 9.29401919227347e-06,
      "loss": 0.5114,
      "num_input_tokens_seen": 18597960,
      "step": 28415
    },
    {
      "epoch": 14.89517819706499,
      "grad_norm": 0.41863101720809937,
      "learning_rate": 9.28512473043013e-06,
      "loss": 0.5551,
      "num_input_tokens_seen": 18600712,
      "step": 28420
    },
    {
      "epoch": 14.897798742138365,
      "grad_norm": 0.49676963686943054,
      "learning_rate": 9.276233555959419e-06,
      "loss": 0.4181,
      "num_input_tokens_seen": 18603944,
      "step": 28425
    },
    {
      "epoch": 14.90041928721174,
      "grad_norm": 0.5359501242637634,
      "learning_rate": 9.267345670721254e-06,
      "loss": 0.4065,
      "num_input_tokens_seen": 18607016,
      "step": 28430
    },
    {
      "epoch": 14.903039832285115,
      "grad_norm": 0.9326598644256592,
      "learning_rate": 9.258461076574895e-06,
      "loss": 0.5437,
      "num_input_tokens_seen": 18609832,
      "step": 28435
    },
    {
      "epoch": 14.90566037735849,
      "grad_norm": 0.4287639260292053,
      "learning_rate": 9.24957977537888e-06,
      "loss": 0.31,
      "num_input_tokens_seen": 18613064,
      "step": 28440
    },
    {
      "epoch": 14.908280922431866,
      "grad_norm": 0.41578206419944763,
      "learning_rate": 9.240701768991086e-06,
      "loss": 0.457,
      "num_input_tokens_seen": 18616648,
      "step": 28445
    },
    {
      "epoch": 14.91090146750524,
      "grad_norm": 0.4077843427658081,
      "learning_rate": 9.231827059268692e-06,
      "loss": 0.4777,
      "num_input_tokens_seen": 18619688,
      "step": 28450
    },
    {
      "epoch": 14.913522012578616,
      "grad_norm": 0.5932729244232178,
      "learning_rate": 9.222955648068193e-06,
      "loss": 0.4796,
      "num_input_tokens_seen": 18622568,
      "step": 28455
    },
    {
      "epoch": 14.916142557651991,
      "grad_norm": 0.48897621035575867,
      "learning_rate": 9.214087537245378e-06,
      "loss": 0.4035,
      "num_input_tokens_seen": 18626216,
      "step": 28460
    },
    {
      "epoch": 14.918763102725366,
      "grad_norm": 0.3364638388156891,
      "learning_rate": 9.205222728655344e-06,
      "loss": 0.4917,
      "num_input_tokens_seen": 18629512,
      "step": 28465
    },
    {
      "epoch": 14.921383647798741,
      "grad_norm": 0.3357117474079132,
      "learning_rate": 9.196361224152522e-06,
      "loss": 0.3933,
      "num_input_tokens_seen": 18632744,
      "step": 28470
    },
    {
      "epoch": 14.924004192872118,
      "grad_norm": 0.6808487772941589,
      "learning_rate": 9.187503025590632e-06,
      "loss": 0.4085,
      "num_input_tokens_seen": 18635368,
      "step": 28475
    },
    {
      "epoch": 14.926624737945493,
      "grad_norm": 0.4611366093158722,
      "learning_rate": 9.17864813482272e-06,
      "loss": 0.4882,
      "num_input_tokens_seen": 18638376,
      "step": 28480
    },
    {
      "epoch": 14.929245283018869,
      "grad_norm": 0.7380167841911316,
      "learning_rate": 9.169796553701122e-06,
      "loss": 0.3913,
      "num_input_tokens_seen": 18641736,
      "step": 28485
    },
    {
      "epoch": 14.931865828092244,
      "grad_norm": 0.6981306076049805,
      "learning_rate": 9.160948284077473e-06,
      "loss": 0.5317,
      "num_input_tokens_seen": 18644136,
      "step": 28490
    },
    {
      "epoch": 14.934486373165619,
      "grad_norm": 0.36026710271835327,
      "learning_rate": 9.152103327802738e-06,
      "loss": 0.4134,
      "num_input_tokens_seen": 18646504,
      "step": 28495
    },
    {
      "epoch": 14.937106918238994,
      "grad_norm": 1.1826105117797852,
      "learning_rate": 9.143261686727189e-06,
      "loss": 0.3985,
      "num_input_tokens_seen": 18650536,
      "step": 28500
    },
    {
      "epoch": 14.93972746331237,
      "grad_norm": 0.2592841386795044,
      "learning_rate": 9.134423362700392e-06,
      "loss": 0.5498,
      "num_input_tokens_seen": 18655944,
      "step": 28505
    },
    {
      "epoch": 14.942348008385745,
      "grad_norm": 0.5449547171592712,
      "learning_rate": 9.12558835757123e-06,
      "loss": 0.3148,
      "num_input_tokens_seen": 18659496,
      "step": 28510
    },
    {
      "epoch": 14.94496855345912,
      "grad_norm": 0.5113848447799683,
      "learning_rate": 9.116756673187878e-06,
      "loss": 0.5764,
      "num_input_tokens_seen": 18662152,
      "step": 28515
    },
    {
      "epoch": 14.947589098532495,
      "grad_norm": 2.026481866836548,
      "learning_rate": 9.107928311397817e-06,
      "loss": 0.5053,
      "num_input_tokens_seen": 18664648,
      "step": 28520
    },
    {
      "epoch": 14.95020964360587,
      "grad_norm": 0.30344998836517334,
      "learning_rate": 9.09910327404785e-06,
      "loss": 0.505,
      "num_input_tokens_seen": 18668968,
      "step": 28525
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.4203422963619232,
      "learning_rate": 9.090281562984068e-06,
      "loss": 0.3344,
      "num_input_tokens_seen": 18672104,
      "step": 28530
    },
    {
      "epoch": 14.95545073375262,
      "grad_norm": 0.4368760585784912,
      "learning_rate": 9.081463180051883e-06,
      "loss": 0.3543,
      "num_input_tokens_seen": 18674472,
      "step": 28535
    },
    {
      "epoch": 14.958071278825996,
      "grad_norm": 0.46161267161369324,
      "learning_rate": 9.072648127095996e-06,
      "loss": 0.3552,
      "num_input_tokens_seen": 18677576,
      "step": 28540
    },
    {
      "epoch": 14.96069182389937,
      "grad_norm": 0.3804023563861847,
      "learning_rate": 9.0638364059604e-06,
      "loss": 0.4349,
      "num_input_tokens_seen": 18681832,
      "step": 28545
    },
    {
      "epoch": 14.963312368972746,
      "grad_norm": 0.7257262468338013,
      "learning_rate": 9.055028018488417e-06,
      "loss": 0.5692,
      "num_input_tokens_seen": 18684104,
      "step": 28550
    },
    {
      "epoch": 14.965932914046121,
      "grad_norm": 0.4291524291038513,
      "learning_rate": 9.046222966522665e-06,
      "loss": 0.4007,
      "num_input_tokens_seen": 18687208,
      "step": 28555
    },
    {
      "epoch": 14.968553459119496,
      "grad_norm": 0.5874577164649963,
      "learning_rate": 9.037421251905059e-06,
      "loss": 0.5009,
      "num_input_tokens_seen": 18689640,
      "step": 28560
    },
    {
      "epoch": 14.971174004192871,
      "grad_norm": 0.4301137924194336,
      "learning_rate": 9.028622876476833e-06,
      "loss": 0.37,
      "num_input_tokens_seen": 18692616,
      "step": 28565
    },
    {
      "epoch": 14.973794549266248,
      "grad_norm": 0.33304402232170105,
      "learning_rate": 9.019827842078474e-06,
      "loss": 0.2766,
      "num_input_tokens_seen": 18696328,
      "step": 28570
    },
    {
      "epoch": 14.976415094339622,
      "grad_norm": 0.5744001865386963,
      "learning_rate": 9.011036150549817e-06,
      "loss": 0.3361,
      "num_input_tokens_seen": 18699240,
      "step": 28575
    },
    {
      "epoch": 14.979035639412999,
      "grad_norm": 0.6728928685188293,
      "learning_rate": 9.002247803729987e-06,
      "loss": 0.5086,
      "num_input_tokens_seen": 18701704,
      "step": 28580
    },
    {
      "epoch": 14.981656184486374,
      "grad_norm": 0.4718582034111023,
      "learning_rate": 8.993462803457404e-06,
      "loss": 0.3735,
      "num_input_tokens_seen": 18704936,
      "step": 28585
    },
    {
      "epoch": 14.984276729559749,
      "grad_norm": 0.7013393640518188,
      "learning_rate": 8.984681151569799e-06,
      "loss": 0.5121,
      "num_input_tokens_seen": 18708808,
      "step": 28590
    },
    {
      "epoch": 14.986897274633124,
      "grad_norm": 0.628915548324585,
      "learning_rate": 8.975902849904183e-06,
      "loss": 0.4433,
      "num_input_tokens_seen": 18711880,
      "step": 28595
    },
    {
      "epoch": 14.9895178197065,
      "grad_norm": 0.4174082279205322,
      "learning_rate": 8.967127900296874e-06,
      "loss": 0.3654,
      "num_input_tokens_seen": 18715304,
      "step": 28600
    },
    {
      "epoch": 14.992138364779874,
      "grad_norm": 0.5322051048278809,
      "learning_rate": 8.95835630458349e-06,
      "loss": 0.3897,
      "num_input_tokens_seen": 18718600,
      "step": 28605
    },
    {
      "epoch": 14.99475890985325,
      "grad_norm": 0.5563614368438721,
      "learning_rate": 8.949588064598958e-06,
      "loss": 0.4626,
      "num_input_tokens_seen": 18721256,
      "step": 28610
    },
    {
      "epoch": 14.997379454926625,
      "grad_norm": 0.42654311656951904,
      "learning_rate": 8.940823182177499e-06,
      "loss": 0.4305,
      "num_input_tokens_seen": 18724008,
      "step": 28615
    },
    {
      "epoch": 15.0,
      "grad_norm": 1.047997236251831,
      "learning_rate": 8.932061659152607e-06,
      "loss": 0.4314,
      "num_input_tokens_seen": 18726216,
      "step": 28620
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.44717922806739807,
      "eval_runtime": 9.6424,
      "eval_samples_per_second": 87.945,
      "eval_steps_per_second": 21.986,
      "num_input_tokens_seen": 18726216,
      "step": 28620
    },
    {
      "epoch": 15.002620545073375,
      "grad_norm": 0.6935693025588989,
      "learning_rate": 8.92330349735711e-06,
      "loss": 0.3866,
      "num_input_tokens_seen": 18729384,
      "step": 28625
    },
    {
      "epoch": 15.00524109014675,
      "grad_norm": 0.3562239408493042,
      "learning_rate": 8.914548698623104e-06,
      "loss": 0.4479,
      "num_input_tokens_seen": 18731944,
      "step": 28630
    },
    {
      "epoch": 15.007861635220126,
      "grad_norm": 0.5604026913642883,
      "learning_rate": 8.905797264781995e-06,
      "loss": 0.4356,
      "num_input_tokens_seen": 18735432,
      "step": 28635
    },
    {
      "epoch": 15.0104821802935,
      "grad_norm": 0.2900761365890503,
      "learning_rate": 8.89704919766449e-06,
      "loss": 0.4173,
      "num_input_tokens_seen": 18739208,
      "step": 28640
    },
    {
      "epoch": 15.013102725366876,
      "grad_norm": 0.6240299344062805,
      "learning_rate": 8.888304499100586e-06,
      "loss": 0.5461,
      "num_input_tokens_seen": 18742472,
      "step": 28645
    },
    {
      "epoch": 15.015723270440251,
      "grad_norm": 0.23646171391010284,
      "learning_rate": 8.879563170919566e-06,
      "loss": 0.3214,
      "num_input_tokens_seen": 18746280,
      "step": 28650
    },
    {
      "epoch": 15.018343815513626,
      "grad_norm": 0.7230652570724487,
      "learning_rate": 8.870825214950016e-06,
      "loss": 0.3201,
      "num_input_tokens_seen": 18748968,
      "step": 28655
    },
    {
      "epoch": 15.020964360587001,
      "grad_norm": 0.6061871647834778,
      "learning_rate": 8.862090633019818e-06,
      "loss": 0.5883,
      "num_input_tokens_seen": 18752008,
      "step": 28660
    },
    {
      "epoch": 15.023584905660377,
      "grad_norm": 0.23998911678791046,
      "learning_rate": 8.853359426956146e-06,
      "loss": 0.319,
      "num_input_tokens_seen": 18756264,
      "step": 28665
    },
    {
      "epoch": 15.026205450733753,
      "grad_norm": 0.7035515308380127,
      "learning_rate": 8.844631598585482e-06,
      "loss": 0.5314,
      "num_input_tokens_seen": 18759560,
      "step": 28670
    },
    {
      "epoch": 15.028825995807129,
      "grad_norm": 0.7351382970809937,
      "learning_rate": 8.835907149733569e-06,
      "loss": 0.3303,
      "num_input_tokens_seen": 18762280,
      "step": 28675
    },
    {
      "epoch": 15.031446540880504,
      "grad_norm": 0.42680805921554565,
      "learning_rate": 8.827186082225477e-06,
      "loss": 0.3445,
      "num_input_tokens_seen": 18765288,
      "step": 28680
    },
    {
      "epoch": 15.034067085953879,
      "grad_norm": 0.42038747668266296,
      "learning_rate": 8.81846839788554e-06,
      "loss": 0.4212,
      "num_input_tokens_seen": 18769096,
      "step": 28685
    },
    {
      "epoch": 15.036687631027254,
      "grad_norm": 0.27803605794906616,
      "learning_rate": 8.809754098537407e-06,
      "loss": 0.4789,
      "num_input_tokens_seen": 18771496,
      "step": 28690
    },
    {
      "epoch": 15.03930817610063,
      "grad_norm": 0.8664239645004272,
      "learning_rate": 8.801043186004016e-06,
      "loss": 0.3877,
      "num_input_tokens_seen": 18774856,
      "step": 28695
    },
    {
      "epoch": 15.041928721174004,
      "grad_norm": 0.3413237929344177,
      "learning_rate": 8.792335662107575e-06,
      "loss": 0.598,
      "num_input_tokens_seen": 18777576,
      "step": 28700
    },
    {
      "epoch": 15.04454926624738,
      "grad_norm": 0.8957993388175964,
      "learning_rate": 8.783631528669605e-06,
      "loss": 0.455,
      "num_input_tokens_seen": 18780008,
      "step": 28705
    },
    {
      "epoch": 15.047169811320755,
      "grad_norm": 0.3540058434009552,
      "learning_rate": 8.774930787510924e-06,
      "loss": 0.5181,
      "num_input_tokens_seen": 18782984,
      "step": 28710
    },
    {
      "epoch": 15.04979035639413,
      "grad_norm": 0.42433691024780273,
      "learning_rate": 8.766233440451608e-06,
      "loss": 0.4281,
      "num_input_tokens_seen": 18785800,
      "step": 28715
    },
    {
      "epoch": 15.052410901467505,
      "grad_norm": 0.5948348045349121,
      "learning_rate": 8.75753948931105e-06,
      "loss": 0.3604,
      "num_input_tokens_seen": 18788040,
      "step": 28720
    },
    {
      "epoch": 15.05503144654088,
      "grad_norm": 0.48635175824165344,
      "learning_rate": 8.748848935907936e-06,
      "loss": 0.4649,
      "num_input_tokens_seen": 18791912,
      "step": 28725
    },
    {
      "epoch": 15.057651991614255,
      "grad_norm": 0.3303627073764801,
      "learning_rate": 8.740161782060214e-06,
      "loss": 0.3312,
      "num_input_tokens_seen": 18795208,
      "step": 28730
    },
    {
      "epoch": 15.06027253668763,
      "grad_norm": 0.3958640694618225,
      "learning_rate": 8.731478029585144e-06,
      "loss": 0.4026,
      "num_input_tokens_seen": 18797960,
      "step": 28735
    },
    {
      "epoch": 15.062893081761006,
      "grad_norm": 0.735343873500824,
      "learning_rate": 8.722797680299278e-06,
      "loss": 0.4667,
      "num_input_tokens_seen": 18801544,
      "step": 28740
    },
    {
      "epoch": 15.065513626834381,
      "grad_norm": 2.051295042037964,
      "learning_rate": 8.71412073601843e-06,
      "loss": 0.6869,
      "num_input_tokens_seen": 18804072,
      "step": 28745
    },
    {
      "epoch": 15.068134171907756,
      "grad_norm": 0.46018466353416443,
      "learning_rate": 8.705447198557731e-06,
      "loss": 0.4088,
      "num_input_tokens_seen": 18806696,
      "step": 28750
    },
    {
      "epoch": 15.070754716981131,
      "grad_norm": 0.3041419982910156,
      "learning_rate": 8.696777069731574e-06,
      "loss": 0.3491,
      "num_input_tokens_seen": 18809864,
      "step": 28755
    },
    {
      "epoch": 15.073375262054507,
      "grad_norm": 0.3872327506542206,
      "learning_rate": 8.688110351353654e-06,
      "loss": 0.52,
      "num_input_tokens_seen": 18814440,
      "step": 28760
    },
    {
      "epoch": 15.075995807127883,
      "grad_norm": 0.5539844036102295,
      "learning_rate": 8.679447045236962e-06,
      "loss": 0.5048,
      "num_input_tokens_seen": 18818024,
      "step": 28765
    },
    {
      "epoch": 15.078616352201259,
      "grad_norm": 0.4121147394180298,
      "learning_rate": 8.670787153193746e-06,
      "loss": 0.501,
      "num_input_tokens_seen": 18822216,
      "step": 28770
    },
    {
      "epoch": 15.081236897274634,
      "grad_norm": 0.6163171529769897,
      "learning_rate": 8.662130677035574e-06,
      "loss": 0.5132,
      "num_input_tokens_seen": 18824616,
      "step": 28775
    },
    {
      "epoch": 15.083857442348009,
      "grad_norm": 0.5258129239082336,
      "learning_rate": 8.65347761857326e-06,
      "loss": 0.3995,
      "num_input_tokens_seen": 18827624,
      "step": 28780
    },
    {
      "epoch": 15.086477987421384,
      "grad_norm": 0.6409345269203186,
      "learning_rate": 8.64482797961694e-06,
      "loss": 0.5033,
      "num_input_tokens_seen": 18830888,
      "step": 28785
    },
    {
      "epoch": 15.08909853249476,
      "grad_norm": 0.5175179839134216,
      "learning_rate": 8.636181761976016e-06,
      "loss": 0.4166,
      "num_input_tokens_seen": 18833640,
      "step": 28790
    },
    {
      "epoch": 15.091719077568134,
      "grad_norm": 0.9514361023902893,
      "learning_rate": 8.62753896745919e-06,
      "loss": 0.5131,
      "num_input_tokens_seen": 18836424,
      "step": 28795
    },
    {
      "epoch": 15.09433962264151,
      "grad_norm": 0.7255339622497559,
      "learning_rate": 8.618899597874413e-06,
      "loss": 0.4413,
      "num_input_tokens_seen": 18839496,
      "step": 28800
    },
    {
      "epoch": 15.096960167714885,
      "grad_norm": 0.5804357528686523,
      "learning_rate": 8.610263655028964e-06,
      "loss": 0.3783,
      "num_input_tokens_seen": 18841928,
      "step": 28805
    },
    {
      "epoch": 15.09958071278826,
      "grad_norm": 0.9741658568382263,
      "learning_rate": 8.601631140729366e-06,
      "loss": 0.3634,
      "num_input_tokens_seen": 18844648,
      "step": 28810
    },
    {
      "epoch": 15.102201257861635,
      "grad_norm": 0.6839268803596497,
      "learning_rate": 8.593002056781451e-06,
      "loss": 0.3993,
      "num_input_tokens_seen": 18847656,
      "step": 28815
    },
    {
      "epoch": 15.10482180293501,
      "grad_norm": 0.38076043128967285,
      "learning_rate": 8.584376404990326e-06,
      "loss": 0.4432,
      "num_input_tokens_seen": 18850760,
      "step": 28820
    },
    {
      "epoch": 15.107442348008385,
      "grad_norm": 0.4462469220161438,
      "learning_rate": 8.575754187160384e-06,
      "loss": 0.3643,
      "num_input_tokens_seen": 18853864,
      "step": 28825
    },
    {
      "epoch": 15.11006289308176,
      "grad_norm": 0.3809220790863037,
      "learning_rate": 8.567135405095291e-06,
      "loss": 0.4518,
      "num_input_tokens_seen": 18857992,
      "step": 28830
    },
    {
      "epoch": 15.112683438155136,
      "grad_norm": 0.4872756600379944,
      "learning_rate": 8.558520060597985e-06,
      "loss": 0.4238,
      "num_input_tokens_seen": 18861576,
      "step": 28835
    },
    {
      "epoch": 15.115303983228511,
      "grad_norm": 0.3783194124698639,
      "learning_rate": 8.54990815547071e-06,
      "loss": 0.4317,
      "num_input_tokens_seen": 18865736,
      "step": 28840
    },
    {
      "epoch": 15.117924528301886,
      "grad_norm": 0.3198864459991455,
      "learning_rate": 8.541299691514974e-06,
      "loss": 0.3695,
      "num_input_tokens_seen": 18869736,
      "step": 28845
    },
    {
      "epoch": 15.120545073375261,
      "grad_norm": 0.37309786677360535,
      "learning_rate": 8.53269467053158e-06,
      "loss": 0.5238,
      "num_input_tokens_seen": 18872520,
      "step": 28850
    },
    {
      "epoch": 15.123165618448636,
      "grad_norm": 0.36630722880363464,
      "learning_rate": 8.524093094320593e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 18876200,
      "step": 28855
    },
    {
      "epoch": 15.125786163522013,
      "grad_norm": 0.44152864813804626,
      "learning_rate": 8.515494964681357e-06,
      "loss": 0.4133,
      "num_input_tokens_seen": 18878856,
      "step": 28860
    },
    {
      "epoch": 15.128406708595389,
      "grad_norm": 0.42072153091430664,
      "learning_rate": 8.506900283412506e-06,
      "loss": 0.5259,
      "num_input_tokens_seen": 18883112,
      "step": 28865
    },
    {
      "epoch": 15.131027253668764,
      "grad_norm": 0.30310988426208496,
      "learning_rate": 8.498309052311953e-06,
      "loss": 0.4047,
      "num_input_tokens_seen": 18886344,
      "step": 28870
    },
    {
      "epoch": 15.133647798742139,
      "grad_norm": 0.3164421319961548,
      "learning_rate": 8.489721273176887e-06,
      "loss": 0.4801,
      "num_input_tokens_seen": 18890248,
      "step": 28875
    },
    {
      "epoch": 15.136268343815514,
      "grad_norm": 0.477510005235672,
      "learning_rate": 8.481136947803777e-06,
      "loss": 0.427,
      "num_input_tokens_seen": 18893512,
      "step": 28880
    },
    {
      "epoch": 15.13888888888889,
      "grad_norm": 0.7773237824440002,
      "learning_rate": 8.472556077988362e-06,
      "loss": 0.4496,
      "num_input_tokens_seen": 18896552,
      "step": 28885
    },
    {
      "epoch": 15.141509433962264,
      "grad_norm": 0.35735610127449036,
      "learning_rate": 8.46397866552565e-06,
      "loss": 0.4866,
      "num_input_tokens_seen": 18899912,
      "step": 28890
    },
    {
      "epoch": 15.14412997903564,
      "grad_norm": 0.9396243095397949,
      "learning_rate": 8.455404712209952e-06,
      "loss": 0.4762,
      "num_input_tokens_seen": 18903336,
      "step": 28895
    },
    {
      "epoch": 15.146750524109015,
      "grad_norm": 0.5326788425445557,
      "learning_rate": 8.446834219834836e-06,
      "loss": 0.4251,
      "num_input_tokens_seen": 18905736,
      "step": 28900
    },
    {
      "epoch": 15.14937106918239,
      "grad_norm": 0.40384581685066223,
      "learning_rate": 8.43826719019316e-06,
      "loss": 0.2861,
      "num_input_tokens_seen": 18909864,
      "step": 28905
    },
    {
      "epoch": 15.151991614255765,
      "grad_norm": 0.6025369167327881,
      "learning_rate": 8.429703625077043e-06,
      "loss": 0.4479,
      "num_input_tokens_seen": 18914280,
      "step": 28910
    },
    {
      "epoch": 15.15461215932914,
      "grad_norm": 0.5479709506034851,
      "learning_rate": 8.421143526277875e-06,
      "loss": 0.4797,
      "num_input_tokens_seen": 18917160,
      "step": 28915
    },
    {
      "epoch": 15.157232704402515,
      "grad_norm": 0.5299652218818665,
      "learning_rate": 8.41258689558634e-06,
      "loss": 0.4301,
      "num_input_tokens_seen": 18920616,
      "step": 28920
    },
    {
      "epoch": 15.15985324947589,
      "grad_norm": 0.3741520643234253,
      "learning_rate": 8.404033734792386e-06,
      "loss": 0.5139,
      "num_input_tokens_seen": 18924616,
      "step": 28925
    },
    {
      "epoch": 15.162473794549266,
      "grad_norm": 1.2986366748809814,
      "learning_rate": 8.395484045685248e-06,
      "loss": 0.4969,
      "num_input_tokens_seen": 18928488,
      "step": 28930
    },
    {
      "epoch": 15.165094339622641,
      "grad_norm": 0.6293283700942993,
      "learning_rate": 8.386937830053412e-06,
      "loss": 0.4412,
      "num_input_tokens_seen": 18931240,
      "step": 28935
    },
    {
      "epoch": 15.167714884696016,
      "grad_norm": 0.6143767833709717,
      "learning_rate": 8.37839508968464e-06,
      "loss": 0.712,
      "num_input_tokens_seen": 18933928,
      "step": 28940
    },
    {
      "epoch": 15.170335429769391,
      "grad_norm": 0.4632165729999542,
      "learning_rate": 8.369855826365988e-06,
      "loss": 0.3958,
      "num_input_tokens_seen": 18936776,
      "step": 28945
    },
    {
      "epoch": 15.172955974842766,
      "grad_norm": 0.308805376291275,
      "learning_rate": 8.361320041883772e-06,
      "loss": 0.4037,
      "num_input_tokens_seen": 18939912,
      "step": 28950
    },
    {
      "epoch": 15.175576519916143,
      "grad_norm": 0.7858838438987732,
      "learning_rate": 8.352787738023576e-06,
      "loss": 0.3669,
      "num_input_tokens_seen": 18942600,
      "step": 28955
    },
    {
      "epoch": 15.178197064989519,
      "grad_norm": 1.0016072988510132,
      "learning_rate": 8.344258916570274e-06,
      "loss": 0.4173,
      "num_input_tokens_seen": 18945640,
      "step": 28960
    },
    {
      "epoch": 15.180817610062894,
      "grad_norm": 0.35030341148376465,
      "learning_rate": 8.335733579307988e-06,
      "loss": 0.3562,
      "num_input_tokens_seen": 18950152,
      "step": 28965
    },
    {
      "epoch": 15.183438155136269,
      "grad_norm": 0.4349398910999298,
      "learning_rate": 8.327211728020113e-06,
      "loss": 0.4027,
      "num_input_tokens_seen": 18952520,
      "step": 28970
    },
    {
      "epoch": 15.186058700209644,
      "grad_norm": 0.38092464208602905,
      "learning_rate": 8.31869336448933e-06,
      "loss": 0.5408,
      "num_input_tokens_seen": 18955336,
      "step": 28975
    },
    {
      "epoch": 15.18867924528302,
      "grad_norm": 0.4483329653739929,
      "learning_rate": 8.310178490497586e-06,
      "loss": 0.3938,
      "num_input_tokens_seen": 18958216,
      "step": 28980
    },
    {
      "epoch": 15.191299790356394,
      "grad_norm": 0.4317742586135864,
      "learning_rate": 8.301667107826103e-06,
      "loss": 0.5657,
      "num_input_tokens_seen": 18961128,
      "step": 28985
    },
    {
      "epoch": 15.19392033542977,
      "grad_norm": 0.5763766765594482,
      "learning_rate": 8.293159218255345e-06,
      "loss": 0.4724,
      "num_input_tokens_seen": 18964200,
      "step": 28990
    },
    {
      "epoch": 15.196540880503145,
      "grad_norm": 0.6291733980178833,
      "learning_rate": 8.284654823565088e-06,
      "loss": 0.5936,
      "num_input_tokens_seen": 18967368,
      "step": 28995
    },
    {
      "epoch": 15.19916142557652,
      "grad_norm": 0.5423244833946228,
      "learning_rate": 8.276153925534333e-06,
      "loss": 0.3914,
      "num_input_tokens_seen": 18970696,
      "step": 29000
    },
    {
      "epoch": 15.201781970649895,
      "grad_norm": 0.5064682364463806,
      "learning_rate": 8.267656525941383e-06,
      "loss": 0.4849,
      "num_input_tokens_seen": 18974280,
      "step": 29005
    },
    {
      "epoch": 15.20440251572327,
      "grad_norm": 0.5785381197929382,
      "learning_rate": 8.259162626563801e-06,
      "loss": 0.3526,
      "num_input_tokens_seen": 18977320,
      "step": 29010
    },
    {
      "epoch": 15.207023060796645,
      "grad_norm": 0.3286251127719879,
      "learning_rate": 8.250672229178402e-06,
      "loss": 0.4009,
      "num_input_tokens_seen": 18980712,
      "step": 29015
    },
    {
      "epoch": 15.20964360587002,
      "grad_norm": 0.39334535598754883,
      "learning_rate": 8.242185335561295e-06,
      "loss": 0.4846,
      "num_input_tokens_seen": 18984168,
      "step": 29020
    },
    {
      "epoch": 15.212264150943396,
      "grad_norm": 0.3642756938934326,
      "learning_rate": 8.23370194748782e-06,
      "loss": 0.3879,
      "num_input_tokens_seen": 18987528,
      "step": 29025
    },
    {
      "epoch": 15.214884696016771,
      "grad_norm": 0.4908362329006195,
      "learning_rate": 8.22522206673262e-06,
      "loss": 0.4365,
      "num_input_tokens_seen": 18991560,
      "step": 29030
    },
    {
      "epoch": 15.217505241090146,
      "grad_norm": 0.8708427548408508,
      "learning_rate": 8.216745695069589e-06,
      "loss": 0.4662,
      "num_input_tokens_seen": 18996360,
      "step": 29035
    },
    {
      "epoch": 15.220125786163521,
      "grad_norm": 0.5840235352516174,
      "learning_rate": 8.208272834271894e-06,
      "loss": 0.4412,
      "num_input_tokens_seen": 18999304,
      "step": 29040
    },
    {
      "epoch": 15.222746331236896,
      "grad_norm": 0.7317090630531311,
      "learning_rate": 8.19980348611194e-06,
      "loss": 0.4256,
      "num_input_tokens_seen": 19002888,
      "step": 29045
    },
    {
      "epoch": 15.225366876310272,
      "grad_norm": 0.46226316690444946,
      "learning_rate": 8.191337652361439e-06,
      "loss": 0.5053,
      "num_input_tokens_seen": 19006760,
      "step": 29050
    },
    {
      "epoch": 15.227987421383649,
      "grad_norm": 0.6501890420913696,
      "learning_rate": 8.182875334791332e-06,
      "loss": 0.4684,
      "num_input_tokens_seen": 19009832,
      "step": 29055
    },
    {
      "epoch": 15.230607966457024,
      "grad_norm": 0.37612348794937134,
      "learning_rate": 8.174416535171841e-06,
      "loss": 0.3448,
      "num_input_tokens_seen": 19013000,
      "step": 29060
    },
    {
      "epoch": 15.233228511530399,
      "grad_norm": 0.7874695062637329,
      "learning_rate": 8.165961255272467e-06,
      "loss": 0.491,
      "num_input_tokens_seen": 19017288,
      "step": 29065
    },
    {
      "epoch": 15.235849056603774,
      "grad_norm": 0.6903502941131592,
      "learning_rate": 8.157509496861935e-06,
      "loss": 0.3586,
      "num_input_tokens_seen": 19020936,
      "step": 29070
    },
    {
      "epoch": 15.23846960167715,
      "grad_norm": 0.6156759262084961,
      "learning_rate": 8.149061261708266e-06,
      "loss": 0.4346,
      "num_input_tokens_seen": 19024488,
      "step": 29075
    },
    {
      "epoch": 15.241090146750524,
      "grad_norm": 0.7398204803466797,
      "learning_rate": 8.140616551578745e-06,
      "loss": 0.4899,
      "num_input_tokens_seen": 19027752,
      "step": 29080
    },
    {
      "epoch": 15.2437106918239,
      "grad_norm": 0.6017060279846191,
      "learning_rate": 8.13217536823989e-06,
      "loss": 0.4038,
      "num_input_tokens_seen": 19030152,
      "step": 29085
    },
    {
      "epoch": 15.246331236897275,
      "grad_norm": 0.714084804058075,
      "learning_rate": 8.12373771345752e-06,
      "loss": 0.4261,
      "num_input_tokens_seen": 19033032,
      "step": 29090
    },
    {
      "epoch": 15.24895178197065,
      "grad_norm": 0.4256148636341095,
      "learning_rate": 8.115303588996676e-06,
      "loss": 0.4736,
      "num_input_tokens_seen": 19036424,
      "step": 29095
    },
    {
      "epoch": 15.251572327044025,
      "grad_norm": 0.46767041087150574,
      "learning_rate": 8.10687299662169e-06,
      "loss": 0.3306,
      "num_input_tokens_seen": 19039592,
      "step": 29100
    },
    {
      "epoch": 15.2541928721174,
      "grad_norm": 0.46631142497062683,
      "learning_rate": 8.098445938096147e-06,
      "loss": 0.3317,
      "num_input_tokens_seen": 19041992,
      "step": 29105
    },
    {
      "epoch": 15.256813417190775,
      "grad_norm": 0.6015570759773254,
      "learning_rate": 8.090022415182898e-06,
      "loss": 0.4157,
      "num_input_tokens_seen": 19044968,
      "step": 29110
    },
    {
      "epoch": 15.25943396226415,
      "grad_norm": 0.5748265385627747,
      "learning_rate": 8.081602429644033e-06,
      "loss": 0.4206,
      "num_input_tokens_seen": 19047656,
      "step": 29115
    },
    {
      "epoch": 15.262054507337526,
      "grad_norm": 0.8975422382354736,
      "learning_rate": 8.073185983240932e-06,
      "loss": 0.5481,
      "num_input_tokens_seen": 19051272,
      "step": 29120
    },
    {
      "epoch": 15.264675052410901,
      "grad_norm": 0.5653550028800964,
      "learning_rate": 8.064773077734206e-06,
      "loss": 0.4426,
      "num_input_tokens_seen": 19054088,
      "step": 29125
    },
    {
      "epoch": 15.267295597484276,
      "grad_norm": 0.6262518167495728,
      "learning_rate": 8.05636371488374e-06,
      "loss": 0.4305,
      "num_input_tokens_seen": 19058248,
      "step": 29130
    },
    {
      "epoch": 15.269916142557651,
      "grad_norm": 0.4656822681427002,
      "learning_rate": 8.047957896448696e-06,
      "loss": 0.4567,
      "num_input_tokens_seen": 19063240,
      "step": 29135
    },
    {
      "epoch": 15.272536687631026,
      "grad_norm": 0.4423675239086151,
      "learning_rate": 8.039555624187451e-06,
      "loss": 0.3896,
      "num_input_tokens_seen": 19065768,
      "step": 29140
    },
    {
      "epoch": 15.275157232704402,
      "grad_norm": 0.6660606265068054,
      "learning_rate": 8.031156899857681e-06,
      "loss": 0.5393,
      "num_input_tokens_seen": 19069256,
      "step": 29145
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 0.5930132865905762,
      "learning_rate": 8.022761725216288e-06,
      "loss": 0.4904,
      "num_input_tokens_seen": 19072072,
      "step": 29150
    },
    {
      "epoch": 15.280398322851154,
      "grad_norm": 0.5881589651107788,
      "learning_rate": 8.014370102019456e-06,
      "loss": 0.3811,
      "num_input_tokens_seen": 19075272,
      "step": 29155
    },
    {
      "epoch": 15.283018867924529,
      "grad_norm": 0.4265514016151428,
      "learning_rate": 8.005982032022616e-06,
      "loss": 0.5349,
      "num_input_tokens_seen": 19078120,
      "step": 29160
    },
    {
      "epoch": 15.285639412997904,
      "grad_norm": 0.537724494934082,
      "learning_rate": 7.997597516980467e-06,
      "loss": 0.4174,
      "num_input_tokens_seen": 19080744,
      "step": 29165
    },
    {
      "epoch": 15.28825995807128,
      "grad_norm": 0.29970377683639526,
      "learning_rate": 7.989216558646942e-06,
      "loss": 0.3682,
      "num_input_tokens_seen": 19084200,
      "step": 29170
    },
    {
      "epoch": 15.290880503144654,
      "grad_norm": 0.3578774333000183,
      "learning_rate": 7.980839158775235e-06,
      "loss": 0.4213,
      "num_input_tokens_seen": 19087400,
      "step": 29175
    },
    {
      "epoch": 15.29350104821803,
      "grad_norm": 0.7310795187950134,
      "learning_rate": 7.972465319117814e-06,
      "loss": 0.3447,
      "num_input_tokens_seen": 19090472,
      "step": 29180
    },
    {
      "epoch": 15.296121593291405,
      "grad_norm": 0.34810659289360046,
      "learning_rate": 7.964095041426386e-06,
      "loss": 0.4983,
      "num_input_tokens_seen": 19093896,
      "step": 29185
    },
    {
      "epoch": 15.29874213836478,
      "grad_norm": 0.42759135365486145,
      "learning_rate": 7.95572832745192e-06,
      "loss": 0.5548,
      "num_input_tokens_seen": 19097384,
      "step": 29190
    },
    {
      "epoch": 15.301362683438155,
      "grad_norm": 0.4360002875328064,
      "learning_rate": 7.947365178944643e-06,
      "loss": 0.4202,
      "num_input_tokens_seen": 19101160,
      "step": 29195
    },
    {
      "epoch": 15.30398322851153,
      "grad_norm": 0.37640902400016785,
      "learning_rate": 7.939005597654025e-06,
      "loss": 0.3828,
      "num_input_tokens_seen": 19105064,
      "step": 29200
    },
    {
      "epoch": 15.306603773584905,
      "grad_norm": 0.6827409863471985,
      "learning_rate": 7.930649585328787e-06,
      "loss": 0.5211,
      "num_input_tokens_seen": 19107272,
      "step": 29205
    },
    {
      "epoch": 15.30922431865828,
      "grad_norm": 0.34948381781578064,
      "learning_rate": 7.922297143716918e-06,
      "loss": 0.4897,
      "num_input_tokens_seen": 19111144,
      "step": 29210
    },
    {
      "epoch": 15.311844863731656,
      "grad_norm": 0.4623952805995941,
      "learning_rate": 7.913948274565652e-06,
      "loss": 0.3573,
      "num_input_tokens_seen": 19114056,
      "step": 29215
    },
    {
      "epoch": 15.314465408805031,
      "grad_norm": 0.44595563411712646,
      "learning_rate": 7.905602979621491e-06,
      "loss": 0.3314,
      "num_input_tokens_seen": 19117416,
      "step": 29220
    },
    {
      "epoch": 15.317085953878406,
      "grad_norm": 0.46466901898384094,
      "learning_rate": 7.897261260630159e-06,
      "loss": 0.3724,
      "num_input_tokens_seen": 19120680,
      "step": 29225
    },
    {
      "epoch": 15.319706498951781,
      "grad_norm": 0.28708669543266296,
      "learning_rate": 7.888923119336647e-06,
      "loss": 0.3495,
      "num_input_tokens_seen": 19124136,
      "step": 29230
    },
    {
      "epoch": 15.322327044025156,
      "grad_norm": 0.5180231332778931,
      "learning_rate": 7.880588557485203e-06,
      "loss": 0.6405,
      "num_input_tokens_seen": 19126984,
      "step": 29235
    },
    {
      "epoch": 15.324947589098532,
      "grad_norm": 0.5706523656845093,
      "learning_rate": 7.872257576819325e-06,
      "loss": 0.4104,
      "num_input_tokens_seen": 19129704,
      "step": 29240
    },
    {
      "epoch": 15.327568134171909,
      "grad_norm": 0.6005495190620422,
      "learning_rate": 7.863930179081769e-06,
      "loss": 0.5576,
      "num_input_tokens_seen": 19132232,
      "step": 29245
    },
    {
      "epoch": 15.330188679245284,
      "grad_norm": 0.43642765283584595,
      "learning_rate": 7.855606366014517e-06,
      "loss": 0.4213,
      "num_input_tokens_seen": 19135016,
      "step": 29250
    },
    {
      "epoch": 15.332809224318659,
      "grad_norm": 0.3275730013847351,
      "learning_rate": 7.847286139358814e-06,
      "loss": 0.3737,
      "num_input_tokens_seen": 19137576,
      "step": 29255
    },
    {
      "epoch": 15.335429769392034,
      "grad_norm": 0.5355511903762817,
      "learning_rate": 7.838969500855162e-06,
      "loss": 0.4571,
      "num_input_tokens_seen": 19140488,
      "step": 29260
    },
    {
      "epoch": 15.33805031446541,
      "grad_norm": 0.4079970121383667,
      "learning_rate": 7.830656452243307e-06,
      "loss": 0.4555,
      "num_input_tokens_seen": 19144456,
      "step": 29265
    },
    {
      "epoch": 15.340670859538784,
      "grad_norm": 0.4211590588092804,
      "learning_rate": 7.822346995262241e-06,
      "loss": 0.3518,
      "num_input_tokens_seen": 19148200,
      "step": 29270
    },
    {
      "epoch": 15.34329140461216,
      "grad_norm": 0.3557562530040741,
      "learning_rate": 7.814041131650223e-06,
      "loss": 0.4849,
      "num_input_tokens_seen": 19152040,
      "step": 29275
    },
    {
      "epoch": 15.345911949685535,
      "grad_norm": 0.5378115177154541,
      "learning_rate": 7.805738863144731e-06,
      "loss": 0.4889,
      "num_input_tokens_seen": 19155272,
      "step": 29280
    },
    {
      "epoch": 15.34853249475891,
      "grad_norm": 0.5786949396133423,
      "learning_rate": 7.797440191482496e-06,
      "loss": 0.6009,
      "num_input_tokens_seen": 19158216,
      "step": 29285
    },
    {
      "epoch": 15.351153039832285,
      "grad_norm": 0.4714966118335724,
      "learning_rate": 7.789145118399518e-06,
      "loss": 0.3816,
      "num_input_tokens_seen": 19161800,
      "step": 29290
    },
    {
      "epoch": 15.35377358490566,
      "grad_norm": 0.9295991063117981,
      "learning_rate": 7.780853645631032e-06,
      "loss": 0.3053,
      "num_input_tokens_seen": 19165160,
      "step": 29295
    },
    {
      "epoch": 15.356394129979035,
      "grad_norm": 0.39507079124450684,
      "learning_rate": 7.772565774911522e-06,
      "loss": 0.4681,
      "num_input_tokens_seen": 19167752,
      "step": 29300
    },
    {
      "epoch": 15.35901467505241,
      "grad_norm": 0.21616216003894806,
      "learning_rate": 7.76428150797471e-06,
      "loss": 0.3837,
      "num_input_tokens_seen": 19171432,
      "step": 29305
    },
    {
      "epoch": 15.361635220125786,
      "grad_norm": 0.6450009346008301,
      "learning_rate": 7.756000846553562e-06,
      "loss": 0.3389,
      "num_input_tokens_seen": 19175720,
      "step": 29310
    },
    {
      "epoch": 15.364255765199161,
      "grad_norm": 0.4073147773742676,
      "learning_rate": 7.74772379238031e-06,
      "loss": 0.4392,
      "num_input_tokens_seen": 19178696,
      "step": 29315
    },
    {
      "epoch": 15.366876310272536,
      "grad_norm": 0.45130857825279236,
      "learning_rate": 7.739450347186417e-06,
      "loss": 0.4456,
      "num_input_tokens_seen": 19181480,
      "step": 29320
    },
    {
      "epoch": 15.369496855345911,
      "grad_norm": 0.5890026688575745,
      "learning_rate": 7.731180512702599e-06,
      "loss": 0.3868,
      "num_input_tokens_seen": 19184136,
      "step": 29325
    },
    {
      "epoch": 15.372117400419286,
      "grad_norm": 0.4224233627319336,
      "learning_rate": 7.722914290658795e-06,
      "loss": 0.6009,
      "num_input_tokens_seen": 19187144,
      "step": 29330
    },
    {
      "epoch": 15.374737945492662,
      "grad_norm": 0.3455183804035187,
      "learning_rate": 7.714651682784224e-06,
      "loss": 0.5019,
      "num_input_tokens_seen": 19190440,
      "step": 29335
    },
    {
      "epoch": 15.377358490566039,
      "grad_norm": 0.38831281661987305,
      "learning_rate": 7.706392690807313e-06,
      "loss": 0.4241,
      "num_input_tokens_seen": 19193736,
      "step": 29340
    },
    {
      "epoch": 15.379979035639414,
      "grad_norm": 0.37749406695365906,
      "learning_rate": 7.69813731645575e-06,
      "loss": 0.348,
      "num_input_tokens_seen": 19197416,
      "step": 29345
    },
    {
      "epoch": 15.382599580712789,
      "grad_norm": 0.4152107238769531,
      "learning_rate": 7.689885561456475e-06,
      "loss": 0.463,
      "num_input_tokens_seen": 19200232,
      "step": 29350
    },
    {
      "epoch": 15.385220125786164,
      "grad_norm": 1.50778067111969,
      "learning_rate": 7.681637427535663e-06,
      "loss": 0.3441,
      "num_input_tokens_seen": 19202824,
      "step": 29355
    },
    {
      "epoch": 15.38784067085954,
      "grad_norm": 0.37317854166030884,
      "learning_rate": 7.673392916418715e-06,
      "loss": 0.328,
      "num_input_tokens_seen": 19205832,
      "step": 29360
    },
    {
      "epoch": 15.390461215932914,
      "grad_norm": 0.5649587512016296,
      "learning_rate": 7.665152029830303e-06,
      "loss": 0.4848,
      "num_input_tokens_seen": 19209128,
      "step": 29365
    },
    {
      "epoch": 15.39308176100629,
      "grad_norm": 0.35799431800842285,
      "learning_rate": 7.656914769494314e-06,
      "loss": 0.4444,
      "num_input_tokens_seen": 19211912,
      "step": 29370
    },
    {
      "epoch": 15.395702306079665,
      "grad_norm": 0.725031316280365,
      "learning_rate": 7.648681137133892e-06,
      "loss": 0.5017,
      "num_input_tokens_seen": 19215080,
      "step": 29375
    },
    {
      "epoch": 15.39832285115304,
      "grad_norm": 0.7580263018608093,
      "learning_rate": 7.640451134471432e-06,
      "loss": 0.4985,
      "num_input_tokens_seen": 19218056,
      "step": 29380
    },
    {
      "epoch": 15.400943396226415,
      "grad_norm": 0.5991779565811157,
      "learning_rate": 7.632224763228538e-06,
      "loss": 0.381,
      "num_input_tokens_seen": 19221096,
      "step": 29385
    },
    {
      "epoch": 15.40356394129979,
      "grad_norm": 0.4394937753677368,
      "learning_rate": 7.6240020251260896e-06,
      "loss": 0.4616,
      "num_input_tokens_seen": 19223816,
      "step": 29390
    },
    {
      "epoch": 15.406184486373165,
      "grad_norm": 0.3211691081523895,
      "learning_rate": 7.615782921884174e-06,
      "loss": 0.3964,
      "num_input_tokens_seen": 19226184,
      "step": 29395
    },
    {
      "epoch": 15.40880503144654,
      "grad_norm": 0.48351550102233887,
      "learning_rate": 7.607567455222142e-06,
      "loss": 0.4372,
      "num_input_tokens_seen": 19229032,
      "step": 29400
    },
    {
      "epoch": 15.411425576519916,
      "grad_norm": 0.4611111879348755,
      "learning_rate": 7.599355626858582e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 19232584,
      "step": 29405
    },
    {
      "epoch": 15.414046121593291,
      "grad_norm": 0.6086786389350891,
      "learning_rate": 7.591147438511298e-06,
      "loss": 0.4291,
      "num_input_tokens_seen": 19235592,
      "step": 29410
    },
    {
      "epoch": 15.416666666666666,
      "grad_norm": 0.6023597121238708,
      "learning_rate": 7.5829428918973625e-06,
      "loss": 0.523,
      "num_input_tokens_seen": 19239304,
      "step": 29415
    },
    {
      "epoch": 15.419287211740041,
      "grad_norm": 0.3710349202156067,
      "learning_rate": 7.574741988733075e-06,
      "loss": 0.3792,
      "num_input_tokens_seen": 19241800,
      "step": 29420
    },
    {
      "epoch": 15.421907756813416,
      "grad_norm": 0.4597957134246826,
      "learning_rate": 7.56654473073396e-06,
      "loss": 0.3812,
      "num_input_tokens_seen": 19245256,
      "step": 29425
    },
    {
      "epoch": 15.424528301886792,
      "grad_norm": 1.3680155277252197,
      "learning_rate": 7.558351119614796e-06,
      "loss": 0.6007,
      "num_input_tokens_seen": 19247880,
      "step": 29430
    },
    {
      "epoch": 15.427148846960169,
      "grad_norm": 0.4053923785686493,
      "learning_rate": 7.5501611570896e-06,
      "loss": 0.4788,
      "num_input_tokens_seen": 19251240,
      "step": 29435
    },
    {
      "epoch": 15.429769392033544,
      "grad_norm": 0.39542996883392334,
      "learning_rate": 7.541974844871602e-06,
      "loss": 0.5018,
      "num_input_tokens_seen": 19254856,
      "step": 29440
    },
    {
      "epoch": 15.432389937106919,
      "grad_norm": 0.42864328622817993,
      "learning_rate": 7.5337921846732965e-06,
      "loss": 0.4487,
      "num_input_tokens_seen": 19258376,
      "step": 29445
    },
    {
      "epoch": 15.435010482180294,
      "grad_norm": 0.5078399181365967,
      "learning_rate": 7.525613178206409e-06,
      "loss": 0.5175,
      "num_input_tokens_seen": 19262312,
      "step": 29450
    },
    {
      "epoch": 15.43763102725367,
      "grad_norm": 0.22923767566680908,
      "learning_rate": 7.517437827181878e-06,
      "loss": 0.5872,
      "num_input_tokens_seen": 19266216,
      "step": 29455
    },
    {
      "epoch": 15.440251572327044,
      "grad_norm": 0.4829672873020172,
      "learning_rate": 7.509266133309908e-06,
      "loss": 0.6594,
      "num_input_tokens_seen": 19269256,
      "step": 29460
    },
    {
      "epoch": 15.44287211740042,
      "grad_norm": 0.43986740708351135,
      "learning_rate": 7.501098098299914e-06,
      "loss": 0.424,
      "num_input_tokens_seen": 19272008,
      "step": 29465
    },
    {
      "epoch": 15.445492662473795,
      "grad_norm": 0.41105106472969055,
      "learning_rate": 7.492933723860557e-06,
      "loss": 0.4891,
      "num_input_tokens_seen": 19275304,
      "step": 29470
    },
    {
      "epoch": 15.44811320754717,
      "grad_norm": 0.9629775881767273,
      "learning_rate": 7.4847730116997334e-06,
      "loss": 0.3696,
      "num_input_tokens_seen": 19277320,
      "step": 29475
    },
    {
      "epoch": 15.450733752620545,
      "grad_norm": 0.4723563492298126,
      "learning_rate": 7.476615963524583e-06,
      "loss": 0.5856,
      "num_input_tokens_seen": 19280680,
      "step": 29480
    },
    {
      "epoch": 15.45335429769392,
      "grad_norm": 0.3246888518333435,
      "learning_rate": 7.468462581041452e-06,
      "loss": 0.4114,
      "num_input_tokens_seen": 19283560,
      "step": 29485
    },
    {
      "epoch": 15.455974842767295,
      "grad_norm": 0.5209487676620483,
      "learning_rate": 7.460312865955932e-06,
      "loss": 0.4388,
      "num_input_tokens_seen": 19287048,
      "step": 29490
    },
    {
      "epoch": 15.45859538784067,
      "grad_norm": 0.40895289182662964,
      "learning_rate": 7.4521668199728584e-06,
      "loss": 0.4716,
      "num_input_tokens_seen": 19289672,
      "step": 29495
    },
    {
      "epoch": 15.461215932914046,
      "grad_norm": 0.5345129370689392,
      "learning_rate": 7.4440244447962884e-06,
      "loss": 0.4654,
      "num_input_tokens_seen": 19292712,
      "step": 29500
    },
    {
      "epoch": 15.463836477987421,
      "grad_norm": 0.3265443444252014,
      "learning_rate": 7.435885742129523e-06,
      "loss": 0.5004,
      "num_input_tokens_seen": 19296520,
      "step": 29505
    },
    {
      "epoch": 15.466457023060796,
      "grad_norm": 0.2664758265018463,
      "learning_rate": 7.427750713675071e-06,
      "loss": 0.4542,
      "num_input_tokens_seen": 19299816,
      "step": 29510
    },
    {
      "epoch": 15.469077568134171,
      "grad_norm": 0.5001009702682495,
      "learning_rate": 7.419619361134702e-06,
      "loss": 0.3227,
      "num_input_tokens_seen": 19302472,
      "step": 29515
    },
    {
      "epoch": 15.471698113207546,
      "grad_norm": 0.5982657074928284,
      "learning_rate": 7.411491686209387e-06,
      "loss": 0.4413,
      "num_input_tokens_seen": 19305640,
      "step": 29520
    },
    {
      "epoch": 15.474318658280922,
      "grad_norm": 0.49250468611717224,
      "learning_rate": 7.4033676905993525e-06,
      "loss": 0.4162,
      "num_input_tokens_seen": 19308680,
      "step": 29525
    },
    {
      "epoch": 15.476939203354299,
      "grad_norm": 0.3998391032218933,
      "learning_rate": 7.3952473760040444e-06,
      "loss": 0.356,
      "num_input_tokens_seen": 19311240,
      "step": 29530
    },
    {
      "epoch": 15.479559748427674,
      "grad_norm": 0.4228455126285553,
      "learning_rate": 7.387130744122148e-06,
      "loss": 0.4045,
      "num_input_tokens_seen": 19314632,
      "step": 29535
    },
    {
      "epoch": 15.482180293501049,
      "grad_norm": 0.5434148907661438,
      "learning_rate": 7.379017796651558e-06,
      "loss": 0.478,
      "num_input_tokens_seen": 19317736,
      "step": 29540
    },
    {
      "epoch": 15.484800838574424,
      "grad_norm": 1.192858099937439,
      "learning_rate": 7.3709085352894085e-06,
      "loss": 0.4286,
      "num_input_tokens_seen": 19320936,
      "step": 29545
    },
    {
      "epoch": 15.4874213836478,
      "grad_norm": 0.5987410545349121,
      "learning_rate": 7.362802961732071e-06,
      "loss": 0.3586,
      "num_input_tokens_seen": 19324360,
      "step": 29550
    },
    {
      "epoch": 15.490041928721174,
      "grad_norm": 0.4377870261669159,
      "learning_rate": 7.354701077675136e-06,
      "loss": 0.4266,
      "num_input_tokens_seen": 19328296,
      "step": 29555
    },
    {
      "epoch": 15.49266247379455,
      "grad_norm": 0.6499455571174622,
      "learning_rate": 7.346602884813439e-06,
      "loss": 0.4316,
      "num_input_tokens_seen": 19331336,
      "step": 29560
    },
    {
      "epoch": 15.495283018867925,
      "grad_norm": 0.33702242374420166,
      "learning_rate": 7.338508384841014e-06,
      "loss": 0.5155,
      "num_input_tokens_seen": 19334760,
      "step": 29565
    },
    {
      "epoch": 15.4979035639413,
      "grad_norm": 0.6114332675933838,
      "learning_rate": 7.3304175794511345e-06,
      "loss": 0.5365,
      "num_input_tokens_seen": 19337320,
      "step": 29570
    },
    {
      "epoch": 15.500524109014675,
      "grad_norm": 0.8291729688644409,
      "learning_rate": 7.3223304703363135e-06,
      "loss": 0.5252,
      "num_input_tokens_seen": 19340584,
      "step": 29575
    },
    {
      "epoch": 15.50314465408805,
      "grad_norm": 0.6195691823959351,
      "learning_rate": 7.314247059188281e-06,
      "loss": 0.5383,
      "num_input_tokens_seen": 19344904,
      "step": 29580
    },
    {
      "epoch": 15.505765199161425,
      "grad_norm": 0.5528520941734314,
      "learning_rate": 7.306167347697992e-06,
      "loss": 0.4211,
      "num_input_tokens_seen": 19348648,
      "step": 29585
    },
    {
      "epoch": 15.5083857442348,
      "grad_norm": 0.40831616520881653,
      "learning_rate": 7.29809133755564e-06,
      "loss": 0.4468,
      "num_input_tokens_seen": 19351816,
      "step": 29590
    },
    {
      "epoch": 15.511006289308176,
      "grad_norm": 0.5165854692459106,
      "learning_rate": 7.290019030450629e-06,
      "loss": 0.4604,
      "num_input_tokens_seen": 19355144,
      "step": 29595
    },
    {
      "epoch": 15.51362683438155,
      "grad_norm": 0.4501517713069916,
      "learning_rate": 7.2819504280715794e-06,
      "loss": 0.4125,
      "num_input_tokens_seen": 19357608,
      "step": 29600
    },
    {
      "epoch": 15.516247379454926,
      "grad_norm": 0.7078660130500793,
      "learning_rate": 7.273885532106364e-06,
      "loss": 0.5113,
      "num_input_tokens_seen": 19360936,
      "step": 29605
    },
    {
      "epoch": 15.518867924528301,
      "grad_norm": 0.27808713912963867,
      "learning_rate": 7.265824344242064e-06,
      "loss": 0.5077,
      "num_input_tokens_seen": 19364232,
      "step": 29610
    },
    {
      "epoch": 15.521488469601676,
      "grad_norm": 0.30340859293937683,
      "learning_rate": 7.257766866165e-06,
      "loss": 0.513,
      "num_input_tokens_seen": 19367720,
      "step": 29615
    },
    {
      "epoch": 15.524109014675052,
      "grad_norm": 0.8491610288619995,
      "learning_rate": 7.24971309956069e-06,
      "loss": 0.441,
      "num_input_tokens_seen": 19370408,
      "step": 29620
    },
    {
      "epoch": 15.526729559748428,
      "grad_norm": 0.43395113945007324,
      "learning_rate": 7.241663046113887e-06,
      "loss": 0.4553,
      "num_input_tokens_seen": 19373576,
      "step": 29625
    },
    {
      "epoch": 15.529350104821804,
      "grad_norm": 0.3952174782752991,
      "learning_rate": 7.233616707508576e-06,
      "loss": 0.4054,
      "num_input_tokens_seen": 19376232,
      "step": 29630
    },
    {
      "epoch": 15.531970649895179,
      "grad_norm": 0.5290385484695435,
      "learning_rate": 7.225574085427961e-06,
      "loss": 0.5952,
      "num_input_tokens_seen": 19379144,
      "step": 29635
    },
    {
      "epoch": 15.534591194968554,
      "grad_norm": 0.670487642288208,
      "learning_rate": 7.217535181554474e-06,
      "loss": 0.4082,
      "num_input_tokens_seen": 19381992,
      "step": 29640
    },
    {
      "epoch": 15.53721174004193,
      "grad_norm": 0.6097529530525208,
      "learning_rate": 7.209499997569747e-06,
      "loss": 0.5664,
      "num_input_tokens_seen": 19385416,
      "step": 29645
    },
    {
      "epoch": 15.539832285115304,
      "grad_norm": 0.5828827619552612,
      "learning_rate": 7.201468535154663e-06,
      "loss": 0.4184,
      "num_input_tokens_seen": 19388328,
      "step": 29650
    },
    {
      "epoch": 15.54245283018868,
      "grad_norm": 0.39000204205513,
      "learning_rate": 7.193440795989295e-06,
      "loss": 0.3655,
      "num_input_tokens_seen": 19391624,
      "step": 29655
    },
    {
      "epoch": 15.545073375262055,
      "grad_norm": 0.6924719214439392,
      "learning_rate": 7.185416781752965e-06,
      "loss": 0.4854,
      "num_input_tokens_seen": 19394376,
      "step": 29660
    },
    {
      "epoch": 15.54769392033543,
      "grad_norm": 0.4408862292766571,
      "learning_rate": 7.177396494124206e-06,
      "loss": 0.4789,
      "num_input_tokens_seen": 19397928,
      "step": 29665
    },
    {
      "epoch": 15.550314465408805,
      "grad_norm": 0.5833641290664673,
      "learning_rate": 7.169379934780779e-06,
      "loss": 0.4623,
      "num_input_tokens_seen": 19400936,
      "step": 29670
    },
    {
      "epoch": 15.55293501048218,
      "grad_norm": 0.42716866731643677,
      "learning_rate": 7.161367105399644e-06,
      "loss": 0.5302,
      "num_input_tokens_seen": 19404360,
      "step": 29675
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 0.7726572155952454,
      "learning_rate": 7.153358007656991e-06,
      "loss": 0.5146,
      "num_input_tokens_seen": 19406568,
      "step": 29680
    },
    {
      "epoch": 15.55817610062893,
      "grad_norm": 0.3333294689655304,
      "learning_rate": 7.145352643228237e-06,
      "loss": 0.4777,
      "num_input_tokens_seen": 19410184,
      "step": 29685
    },
    {
      "epoch": 15.560796645702306,
      "grad_norm": 0.3951112926006317,
      "learning_rate": 7.137351013788013e-06,
      "loss": 0.412,
      "num_input_tokens_seen": 19413480,
      "step": 29690
    },
    {
      "epoch": 15.56341719077568,
      "grad_norm": 0.6853533983230591,
      "learning_rate": 7.129353121010177e-06,
      "loss": 0.3738,
      "num_input_tokens_seen": 19416328,
      "step": 29695
    },
    {
      "epoch": 15.566037735849056,
      "grad_norm": 0.6543521285057068,
      "learning_rate": 7.121358966567779e-06,
      "loss": 0.5273,
      "num_input_tokens_seen": 19419656,
      "step": 29700
    },
    {
      "epoch": 15.568658280922431,
      "grad_norm": 0.4532952606678009,
      "learning_rate": 7.113368552133126e-06,
      "loss": 0.3794,
      "num_input_tokens_seen": 19424456,
      "step": 29705
    },
    {
      "epoch": 15.571278825995806,
      "grad_norm": 0.4741433262825012,
      "learning_rate": 7.105381879377701e-06,
      "loss": 0.4963,
      "num_input_tokens_seen": 19427304,
      "step": 29710
    },
    {
      "epoch": 15.573899371069182,
      "grad_norm": 0.5837791562080383,
      "learning_rate": 7.097398949972234e-06,
      "loss": 0.3471,
      "num_input_tokens_seen": 19429704,
      "step": 29715
    },
    {
      "epoch": 15.576519916142558,
      "grad_norm": 0.3491816520690918,
      "learning_rate": 7.089419765586672e-06,
      "loss": 0.3746,
      "num_input_tokens_seen": 19433000,
      "step": 29720
    },
    {
      "epoch": 15.579140461215934,
      "grad_norm": 0.3139180839061737,
      "learning_rate": 7.081444327890152e-06,
      "loss": 0.3331,
      "num_input_tokens_seen": 19437096,
      "step": 29725
    },
    {
      "epoch": 15.581761006289309,
      "grad_norm": 0.5541657209396362,
      "learning_rate": 7.073472638551054e-06,
      "loss": 0.5454,
      "num_input_tokens_seen": 19440232,
      "step": 29730
    },
    {
      "epoch": 15.584381551362684,
      "grad_norm": 0.6991321444511414,
      "learning_rate": 7.065504699236969e-06,
      "loss": 0.4054,
      "num_input_tokens_seen": 19443528,
      "step": 29735
    },
    {
      "epoch": 15.58700209643606,
      "grad_norm": 0.5139695405960083,
      "learning_rate": 7.057540511614688e-06,
      "loss": 0.3847,
      "num_input_tokens_seen": 19448200,
      "step": 29740
    },
    {
      "epoch": 15.589622641509434,
      "grad_norm": 0.6130852103233337,
      "learning_rate": 7.049580077350232e-06,
      "loss": 0.3305,
      "num_input_tokens_seen": 19450792,
      "step": 29745
    },
    {
      "epoch": 15.59224318658281,
      "grad_norm": 0.5608651041984558,
      "learning_rate": 7.041623398108843e-06,
      "loss": 0.346,
      "num_input_tokens_seen": 19456744,
      "step": 29750
    },
    {
      "epoch": 15.594863731656185,
      "grad_norm": 0.36625272035598755,
      "learning_rate": 7.033670475554949e-06,
      "loss": 0.4808,
      "num_input_tokens_seen": 19459368,
      "step": 29755
    },
    {
      "epoch": 15.59748427672956,
      "grad_norm": 0.5470812916755676,
      "learning_rate": 7.025721311352224e-06,
      "loss": 0.4387,
      "num_input_tokens_seen": 19462952,
      "step": 29760
    },
    {
      "epoch": 15.600104821802935,
      "grad_norm": 0.33719027042388916,
      "learning_rate": 7.017775907163546e-06,
      "loss": 0.5012,
      "num_input_tokens_seen": 19467720,
      "step": 29765
    },
    {
      "epoch": 15.60272536687631,
      "grad_norm": 0.380952924489975,
      "learning_rate": 7.009834264650989e-06,
      "loss": 0.3354,
      "num_input_tokens_seen": 19470952,
      "step": 29770
    },
    {
      "epoch": 15.605345911949685,
      "grad_norm": 0.38904014229774475,
      "learning_rate": 7.001896385475867e-06,
      "loss": 0.4548,
      "num_input_tokens_seen": 19474280,
      "step": 29775
    },
    {
      "epoch": 15.60796645702306,
      "grad_norm": 0.7819775342941284,
      "learning_rate": 6.993962271298682e-06,
      "loss": 0.4252,
      "num_input_tokens_seen": 19478312,
      "step": 29780
    },
    {
      "epoch": 15.610587002096436,
      "grad_norm": 1.1707286834716797,
      "learning_rate": 6.986031923779166e-06,
      "loss": 0.4439,
      "num_input_tokens_seen": 19482536,
      "step": 29785
    },
    {
      "epoch": 15.61320754716981,
      "grad_norm": 0.910822868347168,
      "learning_rate": 6.978105344576264e-06,
      "loss": 0.528,
      "num_input_tokens_seen": 19486696,
      "step": 29790
    },
    {
      "epoch": 15.615828092243186,
      "grad_norm": 0.3205655515193939,
      "learning_rate": 6.970182535348111e-06,
      "loss": 0.311,
      "num_input_tokens_seen": 19490792,
      "step": 29795
    },
    {
      "epoch": 15.618448637316561,
      "grad_norm": 0.49092626571655273,
      "learning_rate": 6.962263497752086e-06,
      "loss": 0.3919,
      "num_input_tokens_seen": 19493768,
      "step": 29800
    },
    {
      "epoch": 15.621069182389936,
      "grad_norm": 0.5175452828407288,
      "learning_rate": 6.9543482334447436e-06,
      "loss": 0.4936,
      "num_input_tokens_seen": 19496968,
      "step": 29805
    },
    {
      "epoch": 15.623689727463312,
      "grad_norm": 0.3538512885570526,
      "learning_rate": 6.946436744081875e-06,
      "loss": 0.394,
      "num_input_tokens_seen": 19500136,
      "step": 29810
    },
    {
      "epoch": 15.626310272536688,
      "grad_norm": 0.5025473237037659,
      "learning_rate": 6.938529031318472e-06,
      "loss": 0.3948,
      "num_input_tokens_seen": 19503432,
      "step": 29815
    },
    {
      "epoch": 15.628930817610064,
      "grad_norm": 0.5352917313575745,
      "learning_rate": 6.9306250968087485e-06,
      "loss": 0.4367,
      "num_input_tokens_seen": 19506856,
      "step": 29820
    },
    {
      "epoch": 15.631551362683439,
      "grad_norm": 0.5885750651359558,
      "learning_rate": 6.922724942206102e-06,
      "loss": 0.4933,
      "num_input_tokens_seen": 19513512,
      "step": 29825
    },
    {
      "epoch": 15.634171907756814,
      "grad_norm": 0.6809061765670776,
      "learning_rate": 6.914828569163167e-06,
      "loss": 0.3212,
      "num_input_tokens_seen": 19515624,
      "step": 29830
    },
    {
      "epoch": 15.63679245283019,
      "grad_norm": 0.45931923389434814,
      "learning_rate": 6.906935979331763e-06,
      "loss": 0.5074,
      "num_input_tokens_seen": 19518696,
      "step": 29835
    },
    {
      "epoch": 15.639412997903564,
      "grad_norm": 0.831092357635498,
      "learning_rate": 6.8990471743629356e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 19521608,
      "step": 29840
    },
    {
      "epoch": 15.64203354297694,
      "grad_norm": 0.6622728109359741,
      "learning_rate": 6.89116215590693e-06,
      "loss": 0.4673,
      "num_input_tokens_seen": 19525416,
      "step": 29845
    },
    {
      "epoch": 15.644654088050315,
      "grad_norm": 0.4402795135974884,
      "learning_rate": 6.8832809256132146e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 19528584,
      "step": 29850
    },
    {
      "epoch": 15.64727463312369,
      "grad_norm": 0.2279849350452423,
      "learning_rate": 6.875403485130444e-06,
      "loss": 0.3758,
      "num_input_tokens_seen": 19531784,
      "step": 29855
    },
    {
      "epoch": 15.649895178197065,
      "grad_norm": 1.0612256526947021,
      "learning_rate": 6.867529836106479e-06,
      "loss": 0.387,
      "num_input_tokens_seen": 19534728,
      "step": 29860
    },
    {
      "epoch": 15.65251572327044,
      "grad_norm": 0.7917433977127075,
      "learning_rate": 6.8596599801884045e-06,
      "loss": 0.4959,
      "num_input_tokens_seen": 19537288,
      "step": 29865
    },
    {
      "epoch": 15.655136268343815,
      "grad_norm": 0.4501040577888489,
      "learning_rate": 6.851793919022509e-06,
      "loss": 0.3487,
      "num_input_tokens_seen": 19540296,
      "step": 29870
    },
    {
      "epoch": 15.65775681341719,
      "grad_norm": 0.7410935759544373,
      "learning_rate": 6.843931654254285e-06,
      "loss": 0.4164,
      "num_input_tokens_seen": 19542504,
      "step": 29875
    },
    {
      "epoch": 15.660377358490566,
      "grad_norm": 0.41440045833587646,
      "learning_rate": 6.836073187528425e-06,
      "loss": 0.3759,
      "num_input_tokens_seen": 19546184,
      "step": 29880
    },
    {
      "epoch": 15.66299790356394,
      "grad_norm": 0.4788266122341156,
      "learning_rate": 6.828218520488821e-06,
      "loss": 0.3895,
      "num_input_tokens_seen": 19549096,
      "step": 29885
    },
    {
      "epoch": 15.665618448637316,
      "grad_norm": 0.34093570709228516,
      "learning_rate": 6.820367654778589e-06,
      "loss": 0.4275,
      "num_input_tokens_seen": 19552296,
      "step": 29890
    },
    {
      "epoch": 15.668238993710691,
      "grad_norm": 0.5311493873596191,
      "learning_rate": 6.812520592040039e-06,
      "loss": 0.5573,
      "num_input_tokens_seen": 19556520,
      "step": 29895
    },
    {
      "epoch": 15.670859538784066,
      "grad_norm": 0.5316265225410461,
      "learning_rate": 6.804677333914689e-06,
      "loss": 0.4429,
      "num_input_tokens_seen": 19560712,
      "step": 29900
    },
    {
      "epoch": 15.673480083857442,
      "grad_norm": 0.45667678117752075,
      "learning_rate": 6.796837882043261e-06,
      "loss": 0.5001,
      "num_input_tokens_seen": 19564072,
      "step": 29905
    },
    {
      "epoch": 15.676100628930818,
      "grad_norm": 0.3129226267337799,
      "learning_rate": 6.7890022380656785e-06,
      "loss": 0.5241,
      "num_input_tokens_seen": 19567176,
      "step": 29910
    },
    {
      "epoch": 15.678721174004194,
      "grad_norm": 0.4515344202518463,
      "learning_rate": 6.781170403621056e-06,
      "loss": 0.4317,
      "num_input_tokens_seen": 19570248,
      "step": 29915
    },
    {
      "epoch": 15.681341719077569,
      "grad_norm": 0.35269367694854736,
      "learning_rate": 6.773342380347736e-06,
      "loss": 0.4694,
      "num_input_tokens_seen": 19573000,
      "step": 29920
    },
    {
      "epoch": 15.683962264150944,
      "grad_norm": 0.8319022059440613,
      "learning_rate": 6.765518169883248e-06,
      "loss": 0.4878,
      "num_input_tokens_seen": 19575272,
      "step": 29925
    },
    {
      "epoch": 15.68658280922432,
      "grad_norm": 0.9374271631240845,
      "learning_rate": 6.757697773864338e-06,
      "loss": 0.4473,
      "num_input_tokens_seen": 19577544,
      "step": 29930
    },
    {
      "epoch": 15.689203354297694,
      "grad_norm": 0.3699159622192383,
      "learning_rate": 6.749881193926932e-06,
      "loss": 0.3285,
      "num_input_tokens_seen": 19581192,
      "step": 29935
    },
    {
      "epoch": 15.69182389937107,
      "grad_norm": 0.5981929302215576,
      "learning_rate": 6.742068431706167e-06,
      "loss": 0.3933,
      "num_input_tokens_seen": 19584456,
      "step": 29940
    },
    {
      "epoch": 15.694444444444445,
      "grad_norm": 0.712679922580719,
      "learning_rate": 6.734259488836386e-06,
      "loss": 0.4955,
      "num_input_tokens_seen": 19587176,
      "step": 29945
    },
    {
      "epoch": 15.69706498951782,
      "grad_norm": 0.4318007230758667,
      "learning_rate": 6.726454366951135e-06,
      "loss": 0.4157,
      "num_input_tokens_seen": 19590664,
      "step": 29950
    },
    {
      "epoch": 15.699685534591195,
      "grad_norm": 0.6434649229049683,
      "learning_rate": 6.7186530676831616e-06,
      "loss": 0.4471,
      "num_input_tokens_seen": 19594792,
      "step": 29955
    },
    {
      "epoch": 15.70230607966457,
      "grad_norm": 0.44071370363235474,
      "learning_rate": 6.710855592664403e-06,
      "loss": 0.4352,
      "num_input_tokens_seen": 19597352,
      "step": 29960
    },
    {
      "epoch": 15.704926624737945,
      "grad_norm": 0.5636324882507324,
      "learning_rate": 6.703061943525993e-06,
      "loss": 0.3793,
      "num_input_tokens_seen": 19600168,
      "step": 29965
    },
    {
      "epoch": 15.70754716981132,
      "grad_norm": 0.383491575717926,
      "learning_rate": 6.69527212189828e-06,
      "loss": 0.5007,
      "num_input_tokens_seen": 19603560,
      "step": 29970
    },
    {
      "epoch": 15.710167714884696,
      "grad_norm": 0.3362613320350647,
      "learning_rate": 6.687486129410811e-06,
      "loss": 0.3393,
      "num_input_tokens_seen": 19607240,
      "step": 29975
    },
    {
      "epoch": 15.71278825995807,
      "grad_norm": 0.5290386080741882,
      "learning_rate": 6.679703967692322e-06,
      "loss": 0.35,
      "num_input_tokens_seen": 19610088,
      "step": 29980
    },
    {
      "epoch": 15.715408805031446,
      "grad_norm": 0.6649067997932434,
      "learning_rate": 6.67192563837076e-06,
      "loss": 0.5091,
      "num_input_tokens_seen": 19613064,
      "step": 29985
    },
    {
      "epoch": 15.718029350104821,
      "grad_norm": 0.5016653537750244,
      "learning_rate": 6.664151143073258e-06,
      "loss": 0.3764,
      "num_input_tokens_seen": 19616264,
      "step": 29990
    },
    {
      "epoch": 15.720649895178196,
      "grad_norm": 0.33840981125831604,
      "learning_rate": 6.656380483426141e-06,
      "loss": 0.4375,
      "num_input_tokens_seen": 19619560,
      "step": 29995
    },
    {
      "epoch": 15.723270440251572,
      "grad_norm": 0.6333931684494019,
      "learning_rate": 6.648613661054956e-06,
      "loss": 0.4475,
      "num_input_tokens_seen": 19622088,
      "step": 30000
    },
    {
      "epoch": 15.725890985324948,
      "grad_norm": 0.41164398193359375,
      "learning_rate": 6.6408506775844256e-06,
      "loss": 0.3972,
      "num_input_tokens_seen": 19626312,
      "step": 30005
    },
    {
      "epoch": 15.728511530398324,
      "grad_norm": 0.7543476223945618,
      "learning_rate": 6.633091534638492e-06,
      "loss": 0.378,
      "num_input_tokens_seen": 19628840,
      "step": 30010
    },
    {
      "epoch": 15.731132075471699,
      "grad_norm": 0.540401816368103,
      "learning_rate": 6.625336233840257e-06,
      "loss": 0.4104,
      "num_input_tokens_seen": 19631720,
      "step": 30015
    },
    {
      "epoch": 15.733752620545074,
      "grad_norm": 0.5547587871551514,
      "learning_rate": 6.617584776812064e-06,
      "loss": 0.5306,
      "num_input_tokens_seen": 19634312,
      "step": 30020
    },
    {
      "epoch": 15.73637316561845,
      "grad_norm": 0.7153112292289734,
      "learning_rate": 6.6098371651754085e-06,
      "loss": 0.4562,
      "num_input_tokens_seen": 19637064,
      "step": 30025
    },
    {
      "epoch": 15.738993710691824,
      "grad_norm": 0.2740249037742615,
      "learning_rate": 6.602093400551012e-06,
      "loss": 0.4575,
      "num_input_tokens_seen": 19640552,
      "step": 30030
    },
    {
      "epoch": 15.7416142557652,
      "grad_norm": 0.8640535473823547,
      "learning_rate": 6.59435348455879e-06,
      "loss": 0.4638,
      "num_input_tokens_seen": 19643624,
      "step": 30035
    },
    {
      "epoch": 15.744234800838575,
      "grad_norm": 0.35346752405166626,
      "learning_rate": 6.586617418817828e-06,
      "loss": 0.3518,
      "num_input_tokens_seen": 19646280,
      "step": 30040
    },
    {
      "epoch": 15.74685534591195,
      "grad_norm": 0.5711010098457336,
      "learning_rate": 6.578885204946439e-06,
      "loss": 0.4189,
      "num_input_tokens_seen": 19649160,
      "step": 30045
    },
    {
      "epoch": 15.749475890985325,
      "grad_norm": 0.6904095411300659,
      "learning_rate": 6.571156844562098e-06,
      "loss": 0.504,
      "num_input_tokens_seen": 19652616,
      "step": 30050
    },
    {
      "epoch": 15.7520964360587,
      "grad_norm": 1.0941593647003174,
      "learning_rate": 6.563432339281497e-06,
      "loss": 0.3636,
      "num_input_tokens_seen": 19655208,
      "step": 30055
    },
    {
      "epoch": 15.754716981132075,
      "grad_norm": 0.4153323173522949,
      "learning_rate": 6.555711690720517e-06,
      "loss": 0.4655,
      "num_input_tokens_seen": 19657928,
      "step": 30060
    },
    {
      "epoch": 15.75733752620545,
      "grad_norm": 0.5034831166267395,
      "learning_rate": 6.547994900494234e-06,
      "loss": 0.4202,
      "num_input_tokens_seen": 19661544,
      "step": 30065
    },
    {
      "epoch": 15.759958071278826,
      "grad_norm": 0.3876982033252716,
      "learning_rate": 6.540281970216899e-06,
      "loss": 0.4632,
      "num_input_tokens_seen": 19664776,
      "step": 30070
    },
    {
      "epoch": 15.7625786163522,
      "grad_norm": 0.44748029112815857,
      "learning_rate": 6.5325729015019845e-06,
      "loss": 0.4341,
      "num_input_tokens_seen": 19668136,
      "step": 30075
    },
    {
      "epoch": 15.765199161425576,
      "grad_norm": 0.45212846994400024,
      "learning_rate": 6.524867695962122e-06,
      "loss": 0.6628,
      "num_input_tokens_seen": 19671592,
      "step": 30080
    },
    {
      "epoch": 15.767819706498951,
      "grad_norm": 0.6153603792190552,
      "learning_rate": 6.517166355209165e-06,
      "loss": 0.4766,
      "num_input_tokens_seen": 19676136,
      "step": 30085
    },
    {
      "epoch": 15.770440251572326,
      "grad_norm": 0.3343924582004547,
      "learning_rate": 6.50946888085415e-06,
      "loss": 0.4138,
      "num_input_tokens_seen": 19679368,
      "step": 30090
    },
    {
      "epoch": 15.773060796645701,
      "grad_norm": 0.26791879534721375,
      "learning_rate": 6.5017752745072855e-06,
      "loss": 0.4009,
      "num_input_tokens_seen": 19682856,
      "step": 30095
    },
    {
      "epoch": 15.775681341719078,
      "grad_norm": 2.378281593322754,
      "learning_rate": 6.494085537777994e-06,
      "loss": 0.6078,
      "num_input_tokens_seen": 19685480,
      "step": 30100
    },
    {
      "epoch": 15.778301886792454,
      "grad_norm": 0.3880050480365753,
      "learning_rate": 6.48639967227489e-06,
      "loss": 0.334,
      "num_input_tokens_seen": 19688104,
      "step": 30105
    },
    {
      "epoch": 15.780922431865829,
      "grad_norm": 0.842960774898529,
      "learning_rate": 6.47871767960575e-06,
      "loss": 0.4327,
      "num_input_tokens_seen": 19691656,
      "step": 30110
    },
    {
      "epoch": 15.783542976939204,
      "grad_norm": 0.5218769311904907,
      "learning_rate": 6.471039561377581e-06,
      "loss": 0.4417,
      "num_input_tokens_seen": 19695336,
      "step": 30115
    },
    {
      "epoch": 15.786163522012579,
      "grad_norm": 0.3633198142051697,
      "learning_rate": 6.463365319196538e-06,
      "loss": 0.4548,
      "num_input_tokens_seen": 19698088,
      "step": 30120
    },
    {
      "epoch": 15.788784067085954,
      "grad_norm": 0.3879700303077698,
      "learning_rate": 6.4556949546679905e-06,
      "loss": 0.4392,
      "num_input_tokens_seen": 19701384,
      "step": 30125
    },
    {
      "epoch": 15.79140461215933,
      "grad_norm": 0.7001583576202393,
      "learning_rate": 6.448028469396497e-06,
      "loss": 0.3857,
      "num_input_tokens_seen": 19705800,
      "step": 30130
    },
    {
      "epoch": 15.794025157232705,
      "grad_norm": 0.8766027688980103,
      "learning_rate": 6.440365864985801e-06,
      "loss": 0.3497,
      "num_input_tokens_seen": 19709384,
      "step": 30135
    },
    {
      "epoch": 15.79664570230608,
      "grad_norm": 0.5054000616073608,
      "learning_rate": 6.432707143038818e-06,
      "loss": 0.3934,
      "num_input_tokens_seen": 19712072,
      "step": 30140
    },
    {
      "epoch": 15.799266247379455,
      "grad_norm": 0.35706350207328796,
      "learning_rate": 6.4250523051576834e-06,
      "loss": 0.4369,
      "num_input_tokens_seen": 19715528,
      "step": 30145
    },
    {
      "epoch": 15.80188679245283,
      "grad_norm": 0.301704466342926,
      "learning_rate": 6.417401352943686e-06,
      "loss": 0.486,
      "num_input_tokens_seen": 19719016,
      "step": 30150
    },
    {
      "epoch": 15.804507337526205,
      "grad_norm": 0.4448402523994446,
      "learning_rate": 6.409754287997322e-06,
      "loss": 0.3484,
      "num_input_tokens_seen": 19722344,
      "step": 30155
    },
    {
      "epoch": 15.80712788259958,
      "grad_norm": 0.3388840854167938,
      "learning_rate": 6.40211111191828e-06,
      "loss": 0.517,
      "num_input_tokens_seen": 19725768,
      "step": 30160
    },
    {
      "epoch": 15.809748427672956,
      "grad_norm": 0.5711359977722168,
      "learning_rate": 6.394471826305409e-06,
      "loss": 0.5942,
      "num_input_tokens_seen": 19728200,
      "step": 30165
    },
    {
      "epoch": 15.81236897274633,
      "grad_norm": 0.5289974212646484,
      "learning_rate": 6.3868364327567795e-06,
      "loss": 0.4561,
      "num_input_tokens_seen": 19730920,
      "step": 30170
    },
    {
      "epoch": 15.814989517819706,
      "grad_norm": 0.5042347311973572,
      "learning_rate": 6.379204932869606e-06,
      "loss": 0.4857,
      "num_input_tokens_seen": 19733800,
      "step": 30175
    },
    {
      "epoch": 15.817610062893081,
      "grad_norm": 0.8598082065582275,
      "learning_rate": 6.371577328240327e-06,
      "loss": 0.4109,
      "num_input_tokens_seen": 19736040,
      "step": 30180
    },
    {
      "epoch": 15.820230607966456,
      "grad_norm": 0.8315336108207703,
      "learning_rate": 6.363953620464547e-06,
      "loss": 0.4533,
      "num_input_tokens_seen": 19739112,
      "step": 30185
    },
    {
      "epoch": 15.822851153039831,
      "grad_norm": 0.651733934879303,
      "learning_rate": 6.356333811137064e-06,
      "loss": 0.4921,
      "num_input_tokens_seen": 19742984,
      "step": 30190
    },
    {
      "epoch": 15.825471698113208,
      "grad_norm": 0.5113019943237305,
      "learning_rate": 6.34871790185185e-06,
      "loss": 0.6426,
      "num_input_tokens_seen": 19746120,
      "step": 30195
    },
    {
      "epoch": 15.828092243186584,
      "grad_norm": 0.3381820321083069,
      "learning_rate": 6.341105894202057e-06,
      "loss": 0.4385,
      "num_input_tokens_seen": 19749992,
      "step": 30200
    },
    {
      "epoch": 15.830712788259959,
      "grad_norm": 0.4887646436691284,
      "learning_rate": 6.333497789780041e-06,
      "loss": 0.5544,
      "num_input_tokens_seen": 19753128,
      "step": 30205
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.25376155972480774,
      "learning_rate": 6.325893590177329e-06,
      "loss": 0.5544,
      "num_input_tokens_seen": 19756328,
      "step": 30210
    },
    {
      "epoch": 15.835953878406709,
      "grad_norm": 0.6592461466789246,
      "learning_rate": 6.318293296984631e-06,
      "loss": 0.4074,
      "num_input_tokens_seen": 19760456,
      "step": 30215
    },
    {
      "epoch": 15.838574423480084,
      "grad_norm": 0.450019896030426,
      "learning_rate": 6.3106969117918495e-06,
      "loss": 0.3579,
      "num_input_tokens_seen": 19764168,
      "step": 30220
    },
    {
      "epoch": 15.84119496855346,
      "grad_norm": 0.6128926277160645,
      "learning_rate": 6.303104436188057e-06,
      "loss": 0.574,
      "num_input_tokens_seen": 19768040,
      "step": 30225
    },
    {
      "epoch": 15.843815513626835,
      "grad_norm": 0.5061340928077698,
      "learning_rate": 6.2955158717615036e-06,
      "loss": 0.4241,
      "num_input_tokens_seen": 19771144,
      "step": 30230
    },
    {
      "epoch": 15.84643605870021,
      "grad_norm": 0.5670109391212463,
      "learning_rate": 6.287931220099638e-06,
      "loss": 0.4453,
      "num_input_tokens_seen": 19774536,
      "step": 30235
    },
    {
      "epoch": 15.849056603773585,
      "grad_norm": 0.3573871850967407,
      "learning_rate": 6.280350482789082e-06,
      "loss": 0.3378,
      "num_input_tokens_seen": 19778024,
      "step": 30240
    },
    {
      "epoch": 15.85167714884696,
      "grad_norm": 0.4599284529685974,
      "learning_rate": 6.272773661415645e-06,
      "loss": 0.4932,
      "num_input_tokens_seen": 19780936,
      "step": 30245
    },
    {
      "epoch": 15.854297693920335,
      "grad_norm": 0.45962008833885193,
      "learning_rate": 6.26520075756431e-06,
      "loss": 0.3549,
      "num_input_tokens_seen": 19784392,
      "step": 30250
    },
    {
      "epoch": 15.85691823899371,
      "grad_norm": 0.5651799440383911,
      "learning_rate": 6.2576317728192304e-06,
      "loss": 0.3961,
      "num_input_tokens_seen": 19787336,
      "step": 30255
    },
    {
      "epoch": 15.859538784067086,
      "grad_norm": 0.39126917719841003,
      "learning_rate": 6.250066708763761e-06,
      "loss": 0.4876,
      "num_input_tokens_seen": 19790664,
      "step": 30260
    },
    {
      "epoch": 15.86215932914046,
      "grad_norm": 0.5173476934432983,
      "learning_rate": 6.242505566980422e-06,
      "loss": 0.699,
      "num_input_tokens_seen": 19793320,
      "step": 30265
    },
    {
      "epoch": 15.864779874213836,
      "grad_norm": 0.44715937972068787,
      "learning_rate": 6.234948349050931e-06,
      "loss": 0.4179,
      "num_input_tokens_seen": 19796360,
      "step": 30270
    },
    {
      "epoch": 15.867400419287211,
      "grad_norm": 2.2479889392852783,
      "learning_rate": 6.227395056556162e-06,
      "loss": 0.4574,
      "num_input_tokens_seen": 19803272,
      "step": 30275
    },
    {
      "epoch": 15.870020964360586,
      "grad_norm": 0.5009881854057312,
      "learning_rate": 6.219845691076173e-06,
      "loss": 0.4929,
      "num_input_tokens_seen": 19812264,
      "step": 30280
    },
    {
      "epoch": 15.872641509433961,
      "grad_norm": 0.39252954721450806,
      "learning_rate": 6.212300254190206e-06,
      "loss": 0.4384,
      "num_input_tokens_seen": 19815144,
      "step": 30285
    },
    {
      "epoch": 15.875262054507338,
      "grad_norm": 0.310672789812088,
      "learning_rate": 6.204758747476688e-06,
      "loss": 0.5838,
      "num_input_tokens_seen": 19819368,
      "step": 30290
    },
    {
      "epoch": 15.877882599580714,
      "grad_norm": 0.6895864009857178,
      "learning_rate": 6.1972211725132095e-06,
      "loss": 0.377,
      "num_input_tokens_seen": 19821736,
      "step": 30295
    },
    {
      "epoch": 15.880503144654089,
      "grad_norm": 0.4434283375740051,
      "learning_rate": 6.189687530876559e-06,
      "loss": 0.6449,
      "num_input_tokens_seen": 19825320,
      "step": 30300
    },
    {
      "epoch": 15.883123689727464,
      "grad_norm": 1.0485032796859741,
      "learning_rate": 6.182157824142676e-06,
      "loss": 0.4844,
      "num_input_tokens_seen": 19829000,
      "step": 30305
    },
    {
      "epoch": 15.885744234800839,
      "grad_norm": 0.49312272667884827,
      "learning_rate": 6.174632053886681e-06,
      "loss": 0.4199,
      "num_input_tokens_seen": 19831592,
      "step": 30310
    },
    {
      "epoch": 15.888364779874214,
      "grad_norm": 0.4098449647426605,
      "learning_rate": 6.167110221682893e-06,
      "loss": 0.3554,
      "num_input_tokens_seen": 19834664,
      "step": 30315
    },
    {
      "epoch": 15.89098532494759,
      "grad_norm": 0.9201260209083557,
      "learning_rate": 6.159592329104788e-06,
      "loss": 0.4401,
      "num_input_tokens_seen": 19837640,
      "step": 30320
    },
    {
      "epoch": 15.893605870020965,
      "grad_norm": 0.36318761110305786,
      "learning_rate": 6.152078377725032e-06,
      "loss": 0.3966,
      "num_input_tokens_seen": 19841192,
      "step": 30325
    },
    {
      "epoch": 15.89622641509434,
      "grad_norm": 0.4006713628768921,
      "learning_rate": 6.144568369115453e-06,
      "loss": 0.469,
      "num_input_tokens_seen": 19845032,
      "step": 30330
    },
    {
      "epoch": 15.898846960167715,
      "grad_norm": 0.39231741428375244,
      "learning_rate": 6.137062304847046e-06,
      "loss": 0.3534,
      "num_input_tokens_seen": 19848232,
      "step": 30335
    },
    {
      "epoch": 15.90146750524109,
      "grad_norm": 0.41856229305267334,
      "learning_rate": 6.129560186490008e-06,
      "loss": 0.5033,
      "num_input_tokens_seen": 19851336,
      "step": 30340
    },
    {
      "epoch": 15.904088050314465,
      "grad_norm": 0.5221222639083862,
      "learning_rate": 6.122062015613694e-06,
      "loss": 0.4644,
      "num_input_tokens_seen": 19856040,
      "step": 30345
    },
    {
      "epoch": 15.90670859538784,
      "grad_norm": 0.3838379383087158,
      "learning_rate": 6.114567793786641e-06,
      "loss": 0.476,
      "num_input_tokens_seen": 19859176,
      "step": 30350
    },
    {
      "epoch": 15.909329140461216,
      "grad_norm": 0.6037972569465637,
      "learning_rate": 6.107077522576543e-06,
      "loss": 0.3472,
      "num_input_tokens_seen": 19862056,
      "step": 30355
    },
    {
      "epoch": 15.91194968553459,
      "grad_norm": 0.4990224838256836,
      "learning_rate": 6.099591203550292e-06,
      "loss": 0.3981,
      "num_input_tokens_seen": 19864968,
      "step": 30360
    },
    {
      "epoch": 15.914570230607966,
      "grad_norm": 0.3223462402820587,
      "learning_rate": 6.092108838273927e-06,
      "loss": 0.3709,
      "num_input_tokens_seen": 19868456,
      "step": 30365
    },
    {
      "epoch": 15.917190775681341,
      "grad_norm": 0.5404132604598999,
      "learning_rate": 6.08463042831268e-06,
      "loss": 0.4477,
      "num_input_tokens_seen": 19872424,
      "step": 30370
    },
    {
      "epoch": 15.919811320754716,
      "grad_norm": 0.7865062952041626,
      "learning_rate": 6.0771559752309496e-06,
      "loss": 0.4801,
      "num_input_tokens_seen": 19875400,
      "step": 30375
    },
    {
      "epoch": 15.922431865828091,
      "grad_norm": 0.5267395973205566,
      "learning_rate": 6.069685480592313e-06,
      "loss": 0.3627,
      "num_input_tokens_seen": 19878184,
      "step": 30380
    },
    {
      "epoch": 15.925052410901468,
      "grad_norm": 0.6081675291061401,
      "learning_rate": 6.062218945959497e-06,
      "loss": 0.5118,
      "num_input_tokens_seen": 19880744,
      "step": 30385
    },
    {
      "epoch": 15.927672955974844,
      "grad_norm": 0.3800152540206909,
      "learning_rate": 6.054756372894435e-06,
      "loss": 0.4643,
      "num_input_tokens_seen": 19883784,
      "step": 30390
    },
    {
      "epoch": 15.930293501048219,
      "grad_norm": 0.7630696892738342,
      "learning_rate": 6.04729776295819e-06,
      "loss": 0.4972,
      "num_input_tokens_seen": 19887304,
      "step": 30395
    },
    {
      "epoch": 15.932914046121594,
      "grad_norm": 0.5378018617630005,
      "learning_rate": 6.0398431177110306e-06,
      "loss": 0.514,
      "num_input_tokens_seen": 19890312,
      "step": 30400
    },
    {
      "epoch": 15.935534591194969,
      "grad_norm": 0.41445374488830566,
      "learning_rate": 6.032392438712389e-06,
      "loss": 0.3909,
      "num_input_tokens_seen": 19893704,
      "step": 30405
    },
    {
      "epoch": 15.938155136268344,
      "grad_norm": 0.8305946588516235,
      "learning_rate": 6.024945727520847e-06,
      "loss": 0.3583,
      "num_input_tokens_seen": 19896552,
      "step": 30410
    },
    {
      "epoch": 15.94077568134172,
      "grad_norm": 0.677929699420929,
      "learning_rate": 6.01750298569419e-06,
      "loss": 0.5136,
      "num_input_tokens_seen": 19899592,
      "step": 30415
    },
    {
      "epoch": 15.943396226415095,
      "grad_norm": 0.607421338558197,
      "learning_rate": 6.010064214789335e-06,
      "loss": 0.5216,
      "num_input_tokens_seen": 19902952,
      "step": 30420
    },
    {
      "epoch": 15.94601677148847,
      "grad_norm": 0.5002890229225159,
      "learning_rate": 6.002629416362399e-06,
      "loss": 0.4191,
      "num_input_tokens_seen": 19906344,
      "step": 30425
    },
    {
      "epoch": 15.948637316561845,
      "grad_norm": 0.5640715956687927,
      "learning_rate": 5.995198591968662e-06,
      "loss": 0.4568,
      "num_input_tokens_seen": 19909864,
      "step": 30430
    },
    {
      "epoch": 15.95125786163522,
      "grad_norm": 0.4006171226501465,
      "learning_rate": 5.987771743162554e-06,
      "loss": 0.42,
      "num_input_tokens_seen": 19913160,
      "step": 30435
    },
    {
      "epoch": 15.953878406708595,
      "grad_norm": 0.7395240664482117,
      "learning_rate": 5.9803488714976955e-06,
      "loss": 0.4281,
      "num_input_tokens_seen": 19917160,
      "step": 30440
    },
    {
      "epoch": 15.95649895178197,
      "grad_norm": 0.39666444063186646,
      "learning_rate": 5.972929978526872e-06,
      "loss": 0.4113,
      "num_input_tokens_seen": 19921672,
      "step": 30445
    },
    {
      "epoch": 15.959119496855346,
      "grad_norm": 0.4449063539505005,
      "learning_rate": 5.965515065802019e-06,
      "loss": 0.4122,
      "num_input_tokens_seen": 19924392,
      "step": 30450
    },
    {
      "epoch": 15.96174004192872,
      "grad_norm": 0.44615113735198975,
      "learning_rate": 5.958104134874254e-06,
      "loss": 0.411,
      "num_input_tokens_seen": 19927720,
      "step": 30455
    },
    {
      "epoch": 15.964360587002096,
      "grad_norm": 0.6934599876403809,
      "learning_rate": 5.950697187293872e-06,
      "loss": 0.4734,
      "num_input_tokens_seen": 19931112,
      "step": 30460
    },
    {
      "epoch": 15.966981132075471,
      "grad_norm": 0.3689444363117218,
      "learning_rate": 5.943294224610305e-06,
      "loss": 0.389,
      "num_input_tokens_seen": 19934152,
      "step": 30465
    },
    {
      "epoch": 15.969601677148846,
      "grad_norm": 0.6529605984687805,
      "learning_rate": 5.935895248372175e-06,
      "loss": 0.5037,
      "num_input_tokens_seen": 19937160,
      "step": 30470
    },
    {
      "epoch": 15.972222222222221,
      "grad_norm": 0.33909234404563904,
      "learning_rate": 5.928500260127273e-06,
      "loss": 0.5093,
      "num_input_tokens_seen": 19940424,
      "step": 30475
    },
    {
      "epoch": 15.974842767295598,
      "grad_norm": 1.3320471048355103,
      "learning_rate": 5.921109261422531e-06,
      "loss": 0.344,
      "num_input_tokens_seen": 19942728,
      "step": 30480
    },
    {
      "epoch": 15.977463312368974,
      "grad_norm": 0.6093885898590088,
      "learning_rate": 5.913722253804071e-06,
      "loss": 0.4961,
      "num_input_tokens_seen": 19946056,
      "step": 30485
    },
    {
      "epoch": 15.980083857442349,
      "grad_norm": 0.35948607325553894,
      "learning_rate": 5.906339238817163e-06,
      "loss": 0.5308,
      "num_input_tokens_seen": 19950632,
      "step": 30490
    },
    {
      "epoch": 15.982704402515724,
      "grad_norm": 0.45498329401016235,
      "learning_rate": 5.8989602180062554e-06,
      "loss": 0.3901,
      "num_input_tokens_seen": 19954280,
      "step": 30495
    },
    {
      "epoch": 15.985324947589099,
      "grad_norm": 0.7784779667854309,
      "learning_rate": 5.891585192914953e-06,
      "loss": 0.3311,
      "num_input_tokens_seen": 19957768,
      "step": 30500
    },
    {
      "epoch": 15.987945492662474,
      "grad_norm": 0.6094395518302917,
      "learning_rate": 5.884214165086036e-06,
      "loss": 0.4781,
      "num_input_tokens_seen": 19960328,
      "step": 30505
    },
    {
      "epoch": 15.99056603773585,
      "grad_norm": 0.4061218202114105,
      "learning_rate": 5.876847136061428e-06,
      "loss": 0.4185,
      "num_input_tokens_seen": 19963272,
      "step": 30510
    },
    {
      "epoch": 15.993186582809225,
      "grad_norm": 0.34578055143356323,
      "learning_rate": 5.869484107382228e-06,
      "loss": 0.337,
      "num_input_tokens_seen": 19966376,
      "step": 30515
    },
    {
      "epoch": 15.9958071278826,
      "grad_norm": 0.4812397062778473,
      "learning_rate": 5.862125080588696e-06,
      "loss": 0.4757,
      "num_input_tokens_seen": 19968968,
      "step": 30520
    },
    {
      "epoch": 15.998427672955975,
      "grad_norm": 0.5647005438804626,
      "learning_rate": 5.854770057220263e-06,
      "loss": 0.316,
      "num_input_tokens_seen": 19971624,
      "step": 30525
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.44694584608078003,
      "eval_runtime": 9.682,
      "eval_samples_per_second": 87.585,
      "eval_steps_per_second": 21.896,
      "num_input_tokens_seen": 19973408,
      "step": 30528
    },
    {
      "epoch": 16.00104821802935,
      "grad_norm": 0.6635236740112305,
      "learning_rate": 5.8474190388155216e-06,
      "loss": 0.4454,
      "num_input_tokens_seen": 19974496,
      "step": 30530
    },
    {
      "epoch": 16.003668763102727,
      "grad_norm": 0.4371119439601898,
      "learning_rate": 5.840072026912205e-06,
      "loss": 0.4176,
      "num_input_tokens_seen": 19977888,
      "step": 30535
    },
    {
      "epoch": 16.0062893081761,
      "grad_norm": 1.1244789361953735,
      "learning_rate": 5.83272902304724e-06,
      "loss": 0.4067,
      "num_input_tokens_seen": 19980480,
      "step": 30540
    },
    {
      "epoch": 16.008909853249477,
      "grad_norm": 0.47525733709335327,
      "learning_rate": 5.825390028756683e-06,
      "loss": 0.4998,
      "num_input_tokens_seen": 19983200,
      "step": 30545
    },
    {
      "epoch": 16.01153039832285,
      "grad_norm": 0.4703885018825531,
      "learning_rate": 5.818055045575776e-06,
      "loss": 0.393,
      "num_input_tokens_seen": 19987200,
      "step": 30550
    },
    {
      "epoch": 16.014150943396228,
      "grad_norm": 0.7593360543251038,
      "learning_rate": 5.810724075038917e-06,
      "loss": 0.4474,
      "num_input_tokens_seen": 19989920,
      "step": 30555
    },
    {
      "epoch": 16.0167714884696,
      "grad_norm": 0.3959534168243408,
      "learning_rate": 5.803397118679665e-06,
      "loss": 0.3964,
      "num_input_tokens_seen": 19993376,
      "step": 30560
    },
    {
      "epoch": 16.019392033542978,
      "grad_norm": 0.6977052688598633,
      "learning_rate": 5.796074178030727e-06,
      "loss": 0.4051,
      "num_input_tokens_seen": 19995264,
      "step": 30565
    },
    {
      "epoch": 16.02201257861635,
      "grad_norm": 0.475078284740448,
      "learning_rate": 5.788755254623976e-06,
      "loss": 0.5816,
      "num_input_tokens_seen": 19997824,
      "step": 30570
    },
    {
      "epoch": 16.02463312368973,
      "grad_norm": 0.5770791172981262,
      "learning_rate": 5.7814403499904525e-06,
      "loss": 0.4793,
      "num_input_tokens_seen": 20000640,
      "step": 30575
    },
    {
      "epoch": 16.0272536687631,
      "grad_norm": 0.7088234424591064,
      "learning_rate": 5.774129465660352e-06,
      "loss": 0.3781,
      "num_input_tokens_seen": 20003264,
      "step": 30580
    },
    {
      "epoch": 16.02987421383648,
      "grad_norm": 0.44263115525245667,
      "learning_rate": 5.7668226031630246e-06,
      "loss": 0.4449,
      "num_input_tokens_seen": 20006688,
      "step": 30585
    },
    {
      "epoch": 16.032494758909852,
      "grad_norm": 0.8370557427406311,
      "learning_rate": 5.759519764026999e-06,
      "loss": 0.4707,
      "num_input_tokens_seen": 20009280,
      "step": 30590
    },
    {
      "epoch": 16.03511530398323,
      "grad_norm": 0.7110258936882019,
      "learning_rate": 5.7522209497799185e-06,
      "loss": 0.449,
      "num_input_tokens_seen": 20012960,
      "step": 30595
    },
    {
      "epoch": 16.037735849056602,
      "grad_norm": 0.7666118144989014,
      "learning_rate": 5.7449261619486246e-06,
      "loss": 0.4967,
      "num_input_tokens_seen": 20015616,
      "step": 30600
    },
    {
      "epoch": 16.04035639412998,
      "grad_norm": 0.36716428399086,
      "learning_rate": 5.737635402059105e-06,
      "loss": 0.5161,
      "num_input_tokens_seen": 20018048,
      "step": 30605
    },
    {
      "epoch": 16.042976939203353,
      "grad_norm": 0.6439269781112671,
      "learning_rate": 5.730348671636501e-06,
      "loss": 0.4291,
      "num_input_tokens_seen": 20021152,
      "step": 30610
    },
    {
      "epoch": 16.04559748427673,
      "grad_norm": 0.5868062973022461,
      "learning_rate": 5.723065972205124e-06,
      "loss": 0.3899,
      "num_input_tokens_seen": 20024032,
      "step": 30615
    },
    {
      "epoch": 16.048218029350103,
      "grad_norm": 0.4951799511909485,
      "learning_rate": 5.715787305288423e-06,
      "loss": 0.4223,
      "num_input_tokens_seen": 20026944,
      "step": 30620
    },
    {
      "epoch": 16.05083857442348,
      "grad_norm": 0.7191483378410339,
      "learning_rate": 5.7085126724090024e-06,
      "loss": 0.4634,
      "num_input_tokens_seen": 20029536,
      "step": 30625
    },
    {
      "epoch": 16.053459119496857,
      "grad_norm": 0.6860195398330688,
      "learning_rate": 5.7012420750886445e-06,
      "loss": 0.421,
      "num_input_tokens_seen": 20032768,
      "step": 30630
    },
    {
      "epoch": 16.05607966457023,
      "grad_norm": 0.43562600016593933,
      "learning_rate": 5.693975514848271e-06,
      "loss": 0.4616,
      "num_input_tokens_seen": 20035712,
      "step": 30635
    },
    {
      "epoch": 16.058700209643607,
      "grad_norm": 0.6525528430938721,
      "learning_rate": 5.686712993207971e-06,
      "loss": 0.4819,
      "num_input_tokens_seen": 20038336,
      "step": 30640
    },
    {
      "epoch": 16.06132075471698,
      "grad_norm": 0.38456234335899353,
      "learning_rate": 5.679454511686977e-06,
      "loss": 0.4368,
      "num_input_tokens_seen": 20042688,
      "step": 30645
    },
    {
      "epoch": 16.063941299790358,
      "grad_norm": 0.4005119800567627,
      "learning_rate": 5.672200071803671e-06,
      "loss": 0.4667,
      "num_input_tokens_seen": 20045568,
      "step": 30650
    },
    {
      "epoch": 16.06656184486373,
      "grad_norm": 0.4916720986366272,
      "learning_rate": 5.664949675075607e-06,
      "loss": 0.4648,
      "num_input_tokens_seen": 20048928,
      "step": 30655
    },
    {
      "epoch": 16.069182389937108,
      "grad_norm": 0.36524245142936707,
      "learning_rate": 5.657703323019486e-06,
      "loss": 0.4531,
      "num_input_tokens_seen": 20051840,
      "step": 30660
    },
    {
      "epoch": 16.07180293501048,
      "grad_norm": 0.28373852372169495,
      "learning_rate": 5.65046101715116e-06,
      "loss": 0.4548,
      "num_input_tokens_seen": 20054592,
      "step": 30665
    },
    {
      "epoch": 16.07442348008386,
      "grad_norm": 0.5670201778411865,
      "learning_rate": 5.643222758985644e-06,
      "loss": 0.6,
      "num_input_tokens_seen": 20058624,
      "step": 30670
    },
    {
      "epoch": 16.07704402515723,
      "grad_norm": 0.5225298404693604,
      "learning_rate": 5.635988550037094e-06,
      "loss": 0.3656,
      "num_input_tokens_seen": 20061632,
      "step": 30675
    },
    {
      "epoch": 16.07966457023061,
      "grad_norm": 0.2951929569244385,
      "learning_rate": 5.628758391818817e-06,
      "loss": 0.3484,
      "num_input_tokens_seen": 20065472,
      "step": 30680
    },
    {
      "epoch": 16.082285115303982,
      "grad_norm": 0.4153038561344147,
      "learning_rate": 5.621532285843287e-06,
      "loss": 0.4407,
      "num_input_tokens_seen": 20068128,
      "step": 30685
    },
    {
      "epoch": 16.08490566037736,
      "grad_norm": 0.5011256337165833,
      "learning_rate": 5.614310233622119e-06,
      "loss": 0.4599,
      "num_input_tokens_seen": 20071872,
      "step": 30690
    },
    {
      "epoch": 16.087526205450732,
      "grad_norm": 0.5142343640327454,
      "learning_rate": 5.607092236666093e-06,
      "loss": 0.5245,
      "num_input_tokens_seen": 20075232,
      "step": 30695
    },
    {
      "epoch": 16.09014675052411,
      "grad_norm": 0.35393500328063965,
      "learning_rate": 5.599878296485125e-06,
      "loss": 0.3894,
      "num_input_tokens_seen": 20078272,
      "step": 30700
    },
    {
      "epoch": 16.092767295597483,
      "grad_norm": 0.6927416920661926,
      "learning_rate": 5.592668414588284e-06,
      "loss": 0.5352,
      "num_input_tokens_seen": 20082304,
      "step": 30705
    },
    {
      "epoch": 16.09538784067086,
      "grad_norm": 0.46241670846939087,
      "learning_rate": 5.585462592483801e-06,
      "loss": 0.4165,
      "num_input_tokens_seen": 20086112,
      "step": 30710
    },
    {
      "epoch": 16.098008385744233,
      "grad_norm": 0.5069248080253601,
      "learning_rate": 5.578260831679047e-06,
      "loss": 0.3407,
      "num_input_tokens_seen": 20090272,
      "step": 30715
    },
    {
      "epoch": 16.10062893081761,
      "grad_norm": 0.5978959798812866,
      "learning_rate": 5.571063133680562e-06,
      "loss": 0.5154,
      "num_input_tokens_seen": 20093280,
      "step": 30720
    },
    {
      "epoch": 16.103249475890987,
      "grad_norm": 0.4265366494655609,
      "learning_rate": 5.563869499994004e-06,
      "loss": 0.3466,
      "num_input_tokens_seen": 20096128,
      "step": 30725
    },
    {
      "epoch": 16.10587002096436,
      "grad_norm": 0.8714920282363892,
      "learning_rate": 5.556679932124215e-06,
      "loss": 0.4314,
      "num_input_tokens_seen": 20099136,
      "step": 30730
    },
    {
      "epoch": 16.108490566037737,
      "grad_norm": 0.27980294823646545,
      "learning_rate": 5.549494431575159e-06,
      "loss": 0.4681,
      "num_input_tokens_seen": 20102240,
      "step": 30735
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 1.0857689380645752,
      "learning_rate": 5.542312999849961e-06,
      "loss": 0.5928,
      "num_input_tokens_seen": 20104960,
      "step": 30740
    },
    {
      "epoch": 16.113731656184488,
      "grad_norm": 0.4908713698387146,
      "learning_rate": 5.5351356384509e-06,
      "loss": 0.4272,
      "num_input_tokens_seen": 20108288,
      "step": 30745
    },
    {
      "epoch": 16.11635220125786,
      "grad_norm": 0.6895716190338135,
      "learning_rate": 5.527962348879401e-06,
      "loss": 0.4517,
      "num_input_tokens_seen": 20111072,
      "step": 30750
    },
    {
      "epoch": 16.118972746331238,
      "grad_norm": 0.42419761419296265,
      "learning_rate": 5.520793132636027e-06,
      "loss": 0.5426,
      "num_input_tokens_seen": 20114240,
      "step": 30755
    },
    {
      "epoch": 16.12159329140461,
      "grad_norm": 0.2692822813987732,
      "learning_rate": 5.513627991220502e-06,
      "loss": 0.3839,
      "num_input_tokens_seen": 20117920,
      "step": 30760
    },
    {
      "epoch": 16.12421383647799,
      "grad_norm": 0.44370266795158386,
      "learning_rate": 5.506466926131684e-06,
      "loss": 0.5093,
      "num_input_tokens_seen": 20121056,
      "step": 30765
    },
    {
      "epoch": 16.12683438155136,
      "grad_norm": 0.556686282157898,
      "learning_rate": 5.4993099388675916e-06,
      "loss": 0.4988,
      "num_input_tokens_seen": 20124160,
      "step": 30770
    },
    {
      "epoch": 16.12945492662474,
      "grad_norm": 0.37643128633499146,
      "learning_rate": 5.49215703092539e-06,
      "loss": 0.431,
      "num_input_tokens_seen": 20127264,
      "step": 30775
    },
    {
      "epoch": 16.132075471698112,
      "grad_norm": 0.47630301117897034,
      "learning_rate": 5.4850082038013755e-06,
      "loss": 0.4731,
      "num_input_tokens_seen": 20129760,
      "step": 30780
    },
    {
      "epoch": 16.13469601677149,
      "grad_norm": 0.3785385489463806,
      "learning_rate": 5.4778634589910115e-06,
      "loss": 0.4294,
      "num_input_tokens_seen": 20133056,
      "step": 30785
    },
    {
      "epoch": 16.137316561844862,
      "grad_norm": 0.47743022441864014,
      "learning_rate": 5.4707227979888834e-06,
      "loss": 0.6548,
      "num_input_tokens_seen": 20139776,
      "step": 30790
    },
    {
      "epoch": 16.13993710691824,
      "grad_norm": 0.682373046875,
      "learning_rate": 5.463586222288747e-06,
      "loss": 0.4166,
      "num_input_tokens_seen": 20143680,
      "step": 30795
    },
    {
      "epoch": 16.142557651991613,
      "grad_norm": 0.6517452001571655,
      "learning_rate": 5.456453733383493e-06,
      "loss": 0.4352,
      "num_input_tokens_seen": 20147776,
      "step": 30800
    },
    {
      "epoch": 16.14517819706499,
      "grad_norm": 0.3296968638896942,
      "learning_rate": 5.449325332765151e-06,
      "loss": 0.4269,
      "num_input_tokens_seen": 20151680,
      "step": 30805
    },
    {
      "epoch": 16.147798742138363,
      "grad_norm": 0.38578304648399353,
      "learning_rate": 5.4422010219249e-06,
      "loss": 0.3532,
      "num_input_tokens_seen": 20155712,
      "step": 30810
    },
    {
      "epoch": 16.15041928721174,
      "grad_norm": 0.7671103477478027,
      "learning_rate": 5.435080802353077e-06,
      "loss": 0.3527,
      "num_input_tokens_seen": 20158592,
      "step": 30815
    },
    {
      "epoch": 16.153039832285117,
      "grad_norm": 0.31446531414985657,
      "learning_rate": 5.4279646755391325e-06,
      "loss": 0.3899,
      "num_input_tokens_seen": 20162752,
      "step": 30820
    },
    {
      "epoch": 16.15566037735849,
      "grad_norm": 0.445536732673645,
      "learning_rate": 5.420852642971691e-06,
      "loss": 0.554,
      "num_input_tokens_seen": 20165440,
      "step": 30825
    },
    {
      "epoch": 16.158280922431867,
      "grad_norm": 0.3845553994178772,
      "learning_rate": 5.41374470613851e-06,
      "loss": 0.4988,
      "num_input_tokens_seen": 20167936,
      "step": 30830
    },
    {
      "epoch": 16.16090146750524,
      "grad_norm": 0.6471143960952759,
      "learning_rate": 5.40664086652648e-06,
      "loss": 0.2623,
      "num_input_tokens_seen": 20170400,
      "step": 30835
    },
    {
      "epoch": 16.163522012578618,
      "grad_norm": 0.42293116450309753,
      "learning_rate": 5.399541125621646e-06,
      "loss": 0.4199,
      "num_input_tokens_seen": 20173792,
      "step": 30840
    },
    {
      "epoch": 16.16614255765199,
      "grad_norm": 0.4741238057613373,
      "learning_rate": 5.392445484909203e-06,
      "loss": 0.4129,
      "num_input_tokens_seen": 20177088,
      "step": 30845
    },
    {
      "epoch": 16.168763102725368,
      "grad_norm": 0.5891104936599731,
      "learning_rate": 5.385353945873464e-06,
      "loss": 0.3954,
      "num_input_tokens_seen": 20181088,
      "step": 30850
    },
    {
      "epoch": 16.17138364779874,
      "grad_norm": 0.6415672898292542,
      "learning_rate": 5.378266509997909e-06,
      "loss": 0.3543,
      "num_input_tokens_seen": 20183808,
      "step": 30855
    },
    {
      "epoch": 16.17400419287212,
      "grad_norm": 0.44753095507621765,
      "learning_rate": 5.371183178765135e-06,
      "loss": 0.4133,
      "num_input_tokens_seen": 20186656,
      "step": 30860
    },
    {
      "epoch": 16.17662473794549,
      "grad_norm": 0.42330652475357056,
      "learning_rate": 5.364103953656904e-06,
      "loss": 0.3558,
      "num_input_tokens_seen": 20190688,
      "step": 30865
    },
    {
      "epoch": 16.17924528301887,
      "grad_norm": 0.31687071919441223,
      "learning_rate": 5.35702883615411e-06,
      "loss": 0.4637,
      "num_input_tokens_seen": 20194560,
      "step": 30870
    },
    {
      "epoch": 16.181865828092242,
      "grad_norm": 0.9177676439285278,
      "learning_rate": 5.34995782773679e-06,
      "loss": 0.5143,
      "num_input_tokens_seen": 20197184,
      "step": 30875
    },
    {
      "epoch": 16.18448637316562,
      "grad_norm": 0.42368873953819275,
      "learning_rate": 5.3428909298841115e-06,
      "loss": 0.4179,
      "num_input_tokens_seen": 20200640,
      "step": 30880
    },
    {
      "epoch": 16.187106918238992,
      "grad_norm": 0.4791402220726013,
      "learning_rate": 5.335828144074384e-06,
      "loss": 0.6155,
      "num_input_tokens_seen": 20203488,
      "step": 30885
    },
    {
      "epoch": 16.18972746331237,
      "grad_norm": 0.4486720561981201,
      "learning_rate": 5.3287694717850695e-06,
      "loss": 0.3541,
      "num_input_tokens_seen": 20208800,
      "step": 30890
    },
    {
      "epoch": 16.192348008385743,
      "grad_norm": 0.453930139541626,
      "learning_rate": 5.321714914492759e-06,
      "loss": 0.326,
      "num_input_tokens_seen": 20213024,
      "step": 30895
    },
    {
      "epoch": 16.19496855345912,
      "grad_norm": 0.8910936117172241,
      "learning_rate": 5.314664473673195e-06,
      "loss": 0.5383,
      "num_input_tokens_seen": 20215968,
      "step": 30900
    },
    {
      "epoch": 16.197589098532493,
      "grad_norm": 0.27274155616760254,
      "learning_rate": 5.307618150801239e-06,
      "loss": 0.4153,
      "num_input_tokens_seen": 20219264,
      "step": 30905
    },
    {
      "epoch": 16.20020964360587,
      "grad_norm": 0.4316534101963043,
      "learning_rate": 5.300575947350897e-06,
      "loss": 0.4537,
      "num_input_tokens_seen": 20222208,
      "step": 30910
    },
    {
      "epoch": 16.202830188679247,
      "grad_norm": 0.4897924065589905,
      "learning_rate": 5.293537864795325e-06,
      "loss": 0.4033,
      "num_input_tokens_seen": 20225760,
      "step": 30915
    },
    {
      "epoch": 16.20545073375262,
      "grad_norm": 0.4967251121997833,
      "learning_rate": 5.2865039046068075e-06,
      "loss": 0.4875,
      "num_input_tokens_seen": 20229120,
      "step": 30920
    },
    {
      "epoch": 16.208071278825997,
      "grad_norm": 0.5469145178794861,
      "learning_rate": 5.279474068256767e-06,
      "loss": 0.5544,
      "num_input_tokens_seen": 20231744,
      "step": 30925
    },
    {
      "epoch": 16.21069182389937,
      "grad_norm": 1.045669674873352,
      "learning_rate": 5.272448357215779e-06,
      "loss": 0.5626,
      "num_input_tokens_seen": 20235808,
      "step": 30930
    },
    {
      "epoch": 16.213312368972748,
      "grad_norm": 0.7720370888710022,
      "learning_rate": 5.26542677295353e-06,
      "loss": 0.5378,
      "num_input_tokens_seen": 20238784,
      "step": 30935
    },
    {
      "epoch": 16.21593291404612,
      "grad_norm": 0.43741971254348755,
      "learning_rate": 5.258409316938848e-06,
      "loss": 0.4585,
      "num_input_tokens_seen": 20241952,
      "step": 30940
    },
    {
      "epoch": 16.218553459119498,
      "grad_norm": 0.9225888252258301,
      "learning_rate": 5.251395990639713e-06,
      "loss": 0.579,
      "num_input_tokens_seen": 20244608,
      "step": 30945
    },
    {
      "epoch": 16.22117400419287,
      "grad_norm": 0.4067034125328064,
      "learning_rate": 5.244386795523235e-06,
      "loss": 0.3685,
      "num_input_tokens_seen": 20247424,
      "step": 30950
    },
    {
      "epoch": 16.22379454926625,
      "grad_norm": 0.4262027144432068,
      "learning_rate": 5.23738173305566e-06,
      "loss": 0.3769,
      "num_input_tokens_seen": 20250240,
      "step": 30955
    },
    {
      "epoch": 16.22641509433962,
      "grad_norm": 0.5716176629066467,
      "learning_rate": 5.230380804702362e-06,
      "loss": 0.3204,
      "num_input_tokens_seen": 20253440,
      "step": 30960
    },
    {
      "epoch": 16.229035639413,
      "grad_norm": 0.5732386708259583,
      "learning_rate": 5.223384011927851e-06,
      "loss": 0.4573,
      "num_input_tokens_seen": 20257440,
      "step": 30965
    },
    {
      "epoch": 16.231656184486372,
      "grad_norm": 1.4010967016220093,
      "learning_rate": 5.216391356195779e-06,
      "loss": 0.6121,
      "num_input_tokens_seen": 20260320,
      "step": 30970
    },
    {
      "epoch": 16.23427672955975,
      "grad_norm": 0.6524578332901001,
      "learning_rate": 5.209402838968935e-06,
      "loss": 0.6973,
      "num_input_tokens_seen": 20263424,
      "step": 30975
    },
    {
      "epoch": 16.236897274633122,
      "grad_norm": 0.3057844042778015,
      "learning_rate": 5.202418461709233e-06,
      "loss": 0.3389,
      "num_input_tokens_seen": 20266752,
      "step": 30980
    },
    {
      "epoch": 16.2395178197065,
      "grad_norm": 0.32131996750831604,
      "learning_rate": 5.195438225877738e-06,
      "loss": 0.4232,
      "num_input_tokens_seen": 20269792,
      "step": 30985
    },
    {
      "epoch": 16.242138364779873,
      "grad_norm": 0.9725826382637024,
      "learning_rate": 5.18846213293461e-06,
      "loss": 0.4611,
      "num_input_tokens_seen": 20272608,
      "step": 30990
    },
    {
      "epoch": 16.24475890985325,
      "grad_norm": 0.6965963840484619,
      "learning_rate": 5.181490184339185e-06,
      "loss": 0.3211,
      "num_input_tokens_seen": 20278688,
      "step": 30995
    },
    {
      "epoch": 16.247379454926623,
      "grad_norm": 0.7823581695556641,
      "learning_rate": 5.17452238154991e-06,
      "loss": 0.3183,
      "num_input_tokens_seen": 20280736,
      "step": 31000
    },
    {
      "epoch": 16.25,
      "grad_norm": 0.4464915692806244,
      "learning_rate": 5.1675587260243715e-06,
      "loss": 0.5,
      "num_input_tokens_seen": 20284800,
      "step": 31005
    },
    {
      "epoch": 16.252620545073377,
      "grad_norm": 0.5802856087684631,
      "learning_rate": 5.160599219219295e-06,
      "loss": 0.59,
      "num_input_tokens_seen": 20288608,
      "step": 31010
    },
    {
      "epoch": 16.25524109014675,
      "grad_norm": 0.7874293923377991,
      "learning_rate": 5.1536438625905185e-06,
      "loss": 0.4894,
      "num_input_tokens_seen": 20291712,
      "step": 31015
    },
    {
      "epoch": 16.257861635220127,
      "grad_norm": 0.9156151413917542,
      "learning_rate": 5.146692657593025e-06,
      "loss": 0.4364,
      "num_input_tokens_seen": 20294880,
      "step": 31020
    },
    {
      "epoch": 16.2604821802935,
      "grad_norm": 0.5871492624282837,
      "learning_rate": 5.139745605680926e-06,
      "loss": 0.3663,
      "num_input_tokens_seen": 20297824,
      "step": 31025
    },
    {
      "epoch": 16.263102725366878,
      "grad_norm": 0.6502978205680847,
      "learning_rate": 5.132802708307469e-06,
      "loss": 0.4486,
      "num_input_tokens_seen": 20300672,
      "step": 31030
    },
    {
      "epoch": 16.26572327044025,
      "grad_norm": 0.3345726728439331,
      "learning_rate": 5.125863966925035e-06,
      "loss": 0.4017,
      "num_input_tokens_seen": 20303232,
      "step": 31035
    },
    {
      "epoch": 16.268343815513628,
      "grad_norm": 0.4479732811450958,
      "learning_rate": 5.118929382985116e-06,
      "loss": 0.4646,
      "num_input_tokens_seen": 20307232,
      "step": 31040
    },
    {
      "epoch": 16.270964360587,
      "grad_norm": 0.4313116669654846,
      "learning_rate": 5.111998957938363e-06,
      "loss": 0.447,
      "num_input_tokens_seen": 20310912,
      "step": 31045
    },
    {
      "epoch": 16.27358490566038,
      "grad_norm": 0.41991129517555237,
      "learning_rate": 5.105072693234525e-06,
      "loss": 0.5023,
      "num_input_tokens_seen": 20313760,
      "step": 31050
    },
    {
      "epoch": 16.27620545073375,
      "grad_norm": 1.281815528869629,
      "learning_rate": 5.098150590322506e-06,
      "loss": 0.4112,
      "num_input_tokens_seen": 20316960,
      "step": 31055
    },
    {
      "epoch": 16.27882599580713,
      "grad_norm": 0.8025782704353333,
      "learning_rate": 5.091232650650335e-06,
      "loss": 0.4128,
      "num_input_tokens_seen": 20319552,
      "step": 31060
    },
    {
      "epoch": 16.281446540880502,
      "grad_norm": 0.5631641149520874,
      "learning_rate": 5.0843188756651645e-06,
      "loss": 0.4849,
      "num_input_tokens_seen": 20322656,
      "step": 31065
    },
    {
      "epoch": 16.28406708595388,
      "grad_norm": 0.5282813310623169,
      "learning_rate": 5.077409266813276e-06,
      "loss": 0.5139,
      "num_input_tokens_seen": 20325376,
      "step": 31070
    },
    {
      "epoch": 16.286687631027252,
      "grad_norm": 0.4027068614959717,
      "learning_rate": 5.0705038255400765e-06,
      "loss": 0.4141,
      "num_input_tokens_seen": 20328960,
      "step": 31075
    },
    {
      "epoch": 16.28930817610063,
      "grad_norm": 0.3621020019054413,
      "learning_rate": 5.063602553290106e-06,
      "loss": 0.3497,
      "num_input_tokens_seen": 20332608,
      "step": 31080
    },
    {
      "epoch": 16.291928721174003,
      "grad_norm": 0.49422019720077515,
      "learning_rate": 5.056705451507035e-06,
      "loss": 0.392,
      "num_input_tokens_seen": 20335744,
      "step": 31085
    },
    {
      "epoch": 16.29454926624738,
      "grad_norm": 0.48607972264289856,
      "learning_rate": 5.049812521633668e-06,
      "loss": 0.4364,
      "num_input_tokens_seen": 20339520,
      "step": 31090
    },
    {
      "epoch": 16.297169811320753,
      "grad_norm": 0.7394883036613464,
      "learning_rate": 5.042923765111907e-06,
      "loss": 0.5812,
      "num_input_tokens_seen": 20341952,
      "step": 31095
    },
    {
      "epoch": 16.29979035639413,
      "grad_norm": 0.2833108901977539,
      "learning_rate": 5.036039183382821e-06,
      "loss": 0.3524,
      "num_input_tokens_seen": 20346112,
      "step": 31100
    },
    {
      "epoch": 16.302410901467507,
      "grad_norm": 0.39090240001678467,
      "learning_rate": 5.02915877788657e-06,
      "loss": 0.4704,
      "num_input_tokens_seen": 20348224,
      "step": 31105
    },
    {
      "epoch": 16.30503144654088,
      "grad_norm": 0.564131498336792,
      "learning_rate": 5.0222825500624625e-06,
      "loss": 0.4682,
      "num_input_tokens_seen": 20351040,
      "step": 31110
    },
    {
      "epoch": 16.307651991614257,
      "grad_norm": 0.2284570336341858,
      "learning_rate": 5.015410501348935e-06,
      "loss": 0.2982,
      "num_input_tokens_seen": 20356032,
      "step": 31115
    },
    {
      "epoch": 16.31027253668763,
      "grad_norm": 0.5560289621353149,
      "learning_rate": 5.008542633183527e-06,
      "loss": 0.3733,
      "num_input_tokens_seen": 20359936,
      "step": 31120
    },
    {
      "epoch": 16.312893081761008,
      "grad_norm": 0.5905187726020813,
      "learning_rate": 5.001678947002927e-06,
      "loss": 0.3342,
      "num_input_tokens_seen": 20362880,
      "step": 31125
    },
    {
      "epoch": 16.31551362683438,
      "grad_norm": 0.8485694527626038,
      "learning_rate": 4.994819444242943e-06,
      "loss": 0.4093,
      "num_input_tokens_seen": 20366784,
      "step": 31130
    },
    {
      "epoch": 16.318134171907758,
      "grad_norm": 0.3291422724723816,
      "learning_rate": 4.9879641263384956e-06,
      "loss": 0.5218,
      "num_input_tokens_seen": 20370496,
      "step": 31135
    },
    {
      "epoch": 16.32075471698113,
      "grad_norm": 0.36855417490005493,
      "learning_rate": 4.981112994723644e-06,
      "loss": 0.3996,
      "num_input_tokens_seen": 20374016,
      "step": 31140
    },
    {
      "epoch": 16.32337526205451,
      "grad_norm": 0.3001963198184967,
      "learning_rate": 4.974266050831572e-06,
      "loss": 0.6646,
      "num_input_tokens_seen": 20377280,
      "step": 31145
    },
    {
      "epoch": 16.32599580712788,
      "grad_norm": 0.7942404747009277,
      "learning_rate": 4.967423296094567e-06,
      "loss": 0.4357,
      "num_input_tokens_seen": 20379872,
      "step": 31150
    },
    {
      "epoch": 16.32861635220126,
      "grad_norm": 0.685187578201294,
      "learning_rate": 4.960584731944076e-06,
      "loss": 0.4971,
      "num_input_tokens_seen": 20382080,
      "step": 31155
    },
    {
      "epoch": 16.331236897274632,
      "grad_norm": 0.34232449531555176,
      "learning_rate": 4.953750359810627e-06,
      "loss": 0.4447,
      "num_input_tokens_seen": 20384992,
      "step": 31160
    },
    {
      "epoch": 16.33385744234801,
      "grad_norm": 0.3623727560043335,
      "learning_rate": 4.946920181123904e-06,
      "loss": 0.3706,
      "num_input_tokens_seen": 20389280,
      "step": 31165
    },
    {
      "epoch": 16.336477987421382,
      "grad_norm": 0.5233638882637024,
      "learning_rate": 4.940094197312709e-06,
      "loss": 0.475,
      "num_input_tokens_seen": 20391808,
      "step": 31170
    },
    {
      "epoch": 16.33909853249476,
      "grad_norm": 0.41210392117500305,
      "learning_rate": 4.933272409804946e-06,
      "loss": 0.3842,
      "num_input_tokens_seen": 20394304,
      "step": 31175
    },
    {
      "epoch": 16.341719077568133,
      "grad_norm": 0.36599090695381165,
      "learning_rate": 4.926454820027663e-06,
      "loss": 0.3572,
      "num_input_tokens_seen": 20398528,
      "step": 31180
    },
    {
      "epoch": 16.34433962264151,
      "grad_norm": 0.639545738697052,
      "learning_rate": 4.919641429407024e-06,
      "loss": 0.5484,
      "num_input_tokens_seen": 20400864,
      "step": 31185
    },
    {
      "epoch": 16.346960167714883,
      "grad_norm": 0.3660264313220978,
      "learning_rate": 4.912832239368304e-06,
      "loss": 0.3495,
      "num_input_tokens_seen": 20403904,
      "step": 31190
    },
    {
      "epoch": 16.34958071278826,
      "grad_norm": 0.48486021161079407,
      "learning_rate": 4.906027251335918e-06,
      "loss": 0.3999,
      "num_input_tokens_seen": 20407360,
      "step": 31195
    },
    {
      "epoch": 16.352201257861637,
      "grad_norm": 0.4845607876777649,
      "learning_rate": 4.899226466733381e-06,
      "loss": 0.3721,
      "num_input_tokens_seen": 20409728,
      "step": 31200
    },
    {
      "epoch": 16.35482180293501,
      "grad_norm": 0.5049707889556885,
      "learning_rate": 4.892429886983346e-06,
      "loss": 0.3755,
      "num_input_tokens_seen": 20412800,
      "step": 31205
    },
    {
      "epoch": 16.357442348008387,
      "grad_norm": 0.516979992389679,
      "learning_rate": 4.88563751350758e-06,
      "loss": 0.3602,
      "num_input_tokens_seen": 20415680,
      "step": 31210
    },
    {
      "epoch": 16.36006289308176,
      "grad_norm": 0.43255218863487244,
      "learning_rate": 4.878849347726977e-06,
      "loss": 0.422,
      "num_input_tokens_seen": 20419744,
      "step": 31215
    },
    {
      "epoch": 16.362683438155138,
      "grad_norm": 0.38967734575271606,
      "learning_rate": 4.872065391061526e-06,
      "loss": 0.6897,
      "num_input_tokens_seen": 20422304,
      "step": 31220
    },
    {
      "epoch": 16.36530398322851,
      "grad_norm": 0.9351868033409119,
      "learning_rate": 4.865285644930373e-06,
      "loss": 0.3652,
      "num_input_tokens_seen": 20424832,
      "step": 31225
    },
    {
      "epoch": 16.367924528301888,
      "grad_norm": 0.4550139605998993,
      "learning_rate": 4.858510110751746e-06,
      "loss": 0.3947,
      "num_input_tokens_seen": 20428256,
      "step": 31230
    },
    {
      "epoch": 16.37054507337526,
      "grad_norm": 0.7738524079322815,
      "learning_rate": 4.851738789943019e-06,
      "loss": 0.3859,
      "num_input_tokens_seen": 20430784,
      "step": 31235
    },
    {
      "epoch": 16.37316561844864,
      "grad_norm": 0.4279380440711975,
      "learning_rate": 4.844971683920674e-06,
      "loss": 0.4772,
      "num_input_tokens_seen": 20433760,
      "step": 31240
    },
    {
      "epoch": 16.37578616352201,
      "grad_norm": 0.5267482995986938,
      "learning_rate": 4.838208794100316e-06,
      "loss": 0.36,
      "num_input_tokens_seen": 20436416,
      "step": 31245
    },
    {
      "epoch": 16.37840670859539,
      "grad_norm": 0.8221210241317749,
      "learning_rate": 4.831450121896661e-06,
      "loss": 0.3568,
      "num_input_tokens_seen": 20439328,
      "step": 31250
    },
    {
      "epoch": 16.381027253668762,
      "grad_norm": 0.3509054183959961,
      "learning_rate": 4.824695668723542e-06,
      "loss": 0.5558,
      "num_input_tokens_seen": 20442016,
      "step": 31255
    },
    {
      "epoch": 16.38364779874214,
      "grad_norm": 0.33723729848861694,
      "learning_rate": 4.817945435993915e-06,
      "loss": 0.4353,
      "num_input_tokens_seen": 20444800,
      "step": 31260
    },
    {
      "epoch": 16.386268343815512,
      "grad_norm": 0.8600237965583801,
      "learning_rate": 4.811199425119853e-06,
      "loss": 0.3148,
      "num_input_tokens_seen": 20447456,
      "step": 31265
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.30663493275642395,
      "learning_rate": 4.804457637512552e-06,
      "loss": 0.3973,
      "num_input_tokens_seen": 20451904,
      "step": 31270
    },
    {
      "epoch": 16.391509433962263,
      "grad_norm": 0.5147437453269958,
      "learning_rate": 4.79772007458231e-06,
      "loss": 0.5097,
      "num_input_tokens_seen": 20455520,
      "step": 31275
    },
    {
      "epoch": 16.39412997903564,
      "grad_norm": 0.3874985873699188,
      "learning_rate": 4.7909867377385424e-06,
      "loss": 0.3439,
      "num_input_tokens_seen": 20459232,
      "step": 31280
    },
    {
      "epoch": 16.396750524109013,
      "grad_norm": 0.5185250639915466,
      "learning_rate": 4.784257628389794e-06,
      "loss": 0.5792,
      "num_input_tokens_seen": 20462400,
      "step": 31285
    },
    {
      "epoch": 16.39937106918239,
      "grad_norm": 0.3147609531879425,
      "learning_rate": 4.7775327479437166e-06,
      "loss": 0.5994,
      "num_input_tokens_seen": 20465120,
      "step": 31290
    },
    {
      "epoch": 16.401991614255767,
      "grad_norm": 0.836302638053894,
      "learning_rate": 4.770812097807076e-06,
      "loss": 0.4902,
      "num_input_tokens_seen": 20467616,
      "step": 31295
    },
    {
      "epoch": 16.40461215932914,
      "grad_norm": 0.463893324136734,
      "learning_rate": 4.764095679385766e-06,
      "loss": 0.4768,
      "num_input_tokens_seen": 20472192,
      "step": 31300
    },
    {
      "epoch": 16.407232704402517,
      "grad_norm": 0.3628605008125305,
      "learning_rate": 4.7573834940847745e-06,
      "loss": 0.3761,
      "num_input_tokens_seen": 20475776,
      "step": 31305
    },
    {
      "epoch": 16.40985324947589,
      "grad_norm": 0.31006723642349243,
      "learning_rate": 4.750675543308212e-06,
      "loss": 0.4674,
      "num_input_tokens_seen": 20479456,
      "step": 31310
    },
    {
      "epoch": 16.412473794549268,
      "grad_norm": 0.4709627628326416,
      "learning_rate": 4.743971828459307e-06,
      "loss": 0.4748,
      "num_input_tokens_seen": 20481952,
      "step": 31315
    },
    {
      "epoch": 16.41509433962264,
      "grad_norm": 0.38641828298568726,
      "learning_rate": 4.7372723509404025e-06,
      "loss": 0.4495,
      "num_input_tokens_seen": 20486016,
      "step": 31320
    },
    {
      "epoch": 16.417714884696018,
      "grad_norm": 0.4316524863243103,
      "learning_rate": 4.7305771121529615e-06,
      "loss": 0.3531,
      "num_input_tokens_seen": 20493312,
      "step": 31325
    },
    {
      "epoch": 16.42033542976939,
      "grad_norm": 0.4634840488433838,
      "learning_rate": 4.723886113497542e-06,
      "loss": 0.4381,
      "num_input_tokens_seen": 20496224,
      "step": 31330
    },
    {
      "epoch": 16.42295597484277,
      "grad_norm": 0.4940658509731293,
      "learning_rate": 4.7171993563738195e-06,
      "loss": 0.5136,
      "num_input_tokens_seen": 20499776,
      "step": 31335
    },
    {
      "epoch": 16.42557651991614,
      "grad_norm": 0.5746532082557678,
      "learning_rate": 4.7105168421805924e-06,
      "loss": 0.3821,
      "num_input_tokens_seen": 20502816,
      "step": 31340
    },
    {
      "epoch": 16.42819706498952,
      "grad_norm": 0.3963151276111603,
      "learning_rate": 4.703838572315767e-06,
      "loss": 0.4362,
      "num_input_tokens_seen": 20506240,
      "step": 31345
    },
    {
      "epoch": 16.430817610062892,
      "grad_norm": 0.44291484355926514,
      "learning_rate": 4.697164548176367e-06,
      "loss": 0.5163,
      "num_input_tokens_seen": 20509408,
      "step": 31350
    },
    {
      "epoch": 16.43343815513627,
      "grad_norm": 0.6862174272537231,
      "learning_rate": 4.690494771158519e-06,
      "loss": 0.3449,
      "num_input_tokens_seen": 20511680,
      "step": 31355
    },
    {
      "epoch": 16.436058700209642,
      "grad_norm": 0.4238582253456116,
      "learning_rate": 4.683829242657453e-06,
      "loss": 0.3841,
      "num_input_tokens_seen": 20515072,
      "step": 31360
    },
    {
      "epoch": 16.43867924528302,
      "grad_norm": 0.5750885009765625,
      "learning_rate": 4.677167964067533e-06,
      "loss": 0.4519,
      "num_input_tokens_seen": 20517856,
      "step": 31365
    },
    {
      "epoch": 16.441299790356393,
      "grad_norm": 0.4468299448490143,
      "learning_rate": 4.670510936782216e-06,
      "loss": 0.3859,
      "num_input_tokens_seen": 20521344,
      "step": 31370
    },
    {
      "epoch": 16.44392033542977,
      "grad_norm": 0.2960250675678253,
      "learning_rate": 4.6638581621940815e-06,
      "loss": 0.8184,
      "num_input_tokens_seen": 20524896,
      "step": 31375
    },
    {
      "epoch": 16.446540880503143,
      "grad_norm": 0.39667245745658875,
      "learning_rate": 4.657209641694821e-06,
      "loss": 0.4175,
      "num_input_tokens_seen": 20528576,
      "step": 31380
    },
    {
      "epoch": 16.44916142557652,
      "grad_norm": 0.6574380397796631,
      "learning_rate": 4.65056537667522e-06,
      "loss": 0.3823,
      "num_input_tokens_seen": 20531712,
      "step": 31385
    },
    {
      "epoch": 16.451781970649897,
      "grad_norm": 0.43767333030700684,
      "learning_rate": 4.643925368525176e-06,
      "loss": 0.5254,
      "num_input_tokens_seen": 20534432,
      "step": 31390
    },
    {
      "epoch": 16.45440251572327,
      "grad_norm": 0.43803131580352783,
      "learning_rate": 4.637289618633711e-06,
      "loss": 0.5234,
      "num_input_tokens_seen": 20538752,
      "step": 31395
    },
    {
      "epoch": 16.457023060796647,
      "grad_norm": 0.4279201924800873,
      "learning_rate": 4.630658128388948e-06,
      "loss": 0.4056,
      "num_input_tokens_seen": 20541376,
      "step": 31400
    },
    {
      "epoch": 16.45964360587002,
      "grad_norm": 0.3113384246826172,
      "learning_rate": 4.624030899178125e-06,
      "loss": 0.5465,
      "num_input_tokens_seen": 20544256,
      "step": 31405
    },
    {
      "epoch": 16.462264150943398,
      "grad_norm": 0.5950334668159485,
      "learning_rate": 4.617407932387569e-06,
      "loss": 0.4147,
      "num_input_tokens_seen": 20548576,
      "step": 31410
    },
    {
      "epoch": 16.46488469601677,
      "grad_norm": 0.5864704251289368,
      "learning_rate": 4.610789229402743e-06,
      "loss": 0.4797,
      "num_input_tokens_seen": 20551968,
      "step": 31415
    },
    {
      "epoch": 16.467505241090148,
      "grad_norm": 0.467511922121048,
      "learning_rate": 4.604174791608193e-06,
      "loss": 0.4154,
      "num_input_tokens_seen": 20555008,
      "step": 31420
    },
    {
      "epoch": 16.47012578616352,
      "grad_norm": 0.31499072909355164,
      "learning_rate": 4.597564620387587e-06,
      "loss": 0.3504,
      "num_input_tokens_seen": 20558144,
      "step": 31425
    },
    {
      "epoch": 16.4727463312369,
      "grad_norm": 0.5201533436775208,
      "learning_rate": 4.590958717123708e-06,
      "loss": 0.8385,
      "num_input_tokens_seen": 20560704,
      "step": 31430
    },
    {
      "epoch": 16.47536687631027,
      "grad_norm": 0.45456254482269287,
      "learning_rate": 4.5843570831984174e-06,
      "loss": 0.532,
      "num_input_tokens_seen": 20564608,
      "step": 31435
    },
    {
      "epoch": 16.47798742138365,
      "grad_norm": 0.4778614342212677,
      "learning_rate": 4.577759719992719e-06,
      "loss": 0.4701,
      "num_input_tokens_seen": 20568096,
      "step": 31440
    },
    {
      "epoch": 16.480607966457022,
      "grad_norm": 0.3756287097930908,
      "learning_rate": 4.571166628886692e-06,
      "loss": 0.4135,
      "num_input_tokens_seen": 20571744,
      "step": 31445
    },
    {
      "epoch": 16.4832285115304,
      "grad_norm": 0.41664206981658936,
      "learning_rate": 4.564577811259541e-06,
      "loss": 0.4192,
      "num_input_tokens_seen": 20574368,
      "step": 31450
    },
    {
      "epoch": 16.485849056603772,
      "grad_norm": 0.5033436417579651,
      "learning_rate": 4.557993268489574e-06,
      "loss": 0.4053,
      "num_input_tokens_seen": 20578080,
      "step": 31455
    },
    {
      "epoch": 16.48846960167715,
      "grad_norm": 0.3974302411079407,
      "learning_rate": 4.551413001954205e-06,
      "loss": 0.4712,
      "num_input_tokens_seen": 20581248,
      "step": 31460
    },
    {
      "epoch": 16.491090146750523,
      "grad_norm": 0.46511662006378174,
      "learning_rate": 4.54483701302994e-06,
      "loss": 0.4063,
      "num_input_tokens_seen": 20585664,
      "step": 31465
    },
    {
      "epoch": 16.4937106918239,
      "grad_norm": 0.37413057684898376,
      "learning_rate": 4.538265303092415e-06,
      "loss": 0.4052,
      "num_input_tokens_seen": 20589024,
      "step": 31470
    },
    {
      "epoch": 16.496331236897273,
      "grad_norm": 0.31710830330848694,
      "learning_rate": 4.531697873516338e-06,
      "loss": 0.3428,
      "num_input_tokens_seen": 20592320,
      "step": 31475
    },
    {
      "epoch": 16.49895178197065,
      "grad_norm": 0.5497682094573975,
      "learning_rate": 4.525134725675556e-06,
      "loss": 0.3815,
      "num_input_tokens_seen": 20594848,
      "step": 31480
    },
    {
      "epoch": 16.501572327044027,
      "grad_norm": 0.3871712386608124,
      "learning_rate": 4.518575860943003e-06,
      "loss": 0.3907,
      "num_input_tokens_seen": 20597504,
      "step": 31485
    },
    {
      "epoch": 16.5041928721174,
      "grad_norm": 0.4350557029247284,
      "learning_rate": 4.512021280690712e-06,
      "loss": 0.5029,
      "num_input_tokens_seen": 20601152,
      "step": 31490
    },
    {
      "epoch": 16.506813417190777,
      "grad_norm": 0.6368108987808228,
      "learning_rate": 4.505470986289828e-06,
      "loss": 0.4178,
      "num_input_tokens_seen": 20603936,
      "step": 31495
    },
    {
      "epoch": 16.50943396226415,
      "grad_norm": 0.37852317094802856,
      "learning_rate": 4.498924979110605e-06,
      "loss": 0.5135,
      "num_input_tokens_seen": 20607872,
      "step": 31500
    },
    {
      "epoch": 16.512054507337528,
      "grad_norm": 0.3730285167694092,
      "learning_rate": 4.492383260522384e-06,
      "loss": 0.6242,
      "num_input_tokens_seen": 20612288,
      "step": 31505
    },
    {
      "epoch": 16.5146750524109,
      "grad_norm": 0.3737787902355194,
      "learning_rate": 4.485845831893626e-06,
      "loss": 0.3819,
      "num_input_tokens_seen": 20615072,
      "step": 31510
    },
    {
      "epoch": 16.517295597484278,
      "grad_norm": 0.503854513168335,
      "learning_rate": 4.479312694591878e-06,
      "loss": 0.4408,
      "num_input_tokens_seen": 20620736,
      "step": 31515
    },
    {
      "epoch": 16.51991614255765,
      "grad_norm": 0.4261833429336548,
      "learning_rate": 4.4727838499838e-06,
      "loss": 0.4204,
      "num_input_tokens_seen": 20624640,
      "step": 31520
    },
    {
      "epoch": 16.52253668763103,
      "grad_norm": 0.5115658640861511,
      "learning_rate": 4.4662592994351545e-06,
      "loss": 0.5354,
      "num_input_tokens_seen": 20627968,
      "step": 31525
    },
    {
      "epoch": 16.5251572327044,
      "grad_norm": 0.4937722682952881,
      "learning_rate": 4.4597390443108065e-06,
      "loss": 0.4683,
      "num_input_tokens_seen": 20631360,
      "step": 31530
    },
    {
      "epoch": 16.52777777777778,
      "grad_norm": 0.68111252784729,
      "learning_rate": 4.453223085974712e-06,
      "loss": 0.5382,
      "num_input_tokens_seen": 20633760,
      "step": 31535
    },
    {
      "epoch": 16.530398322851152,
      "grad_norm": 0.43278104066848755,
      "learning_rate": 4.4467114257899425e-06,
      "loss": 0.4528,
      "num_input_tokens_seen": 20636640,
      "step": 31540
    },
    {
      "epoch": 16.53301886792453,
      "grad_norm": 0.8349581360816956,
      "learning_rate": 4.44020406511865e-06,
      "loss": 0.397,
      "num_input_tokens_seen": 20639648,
      "step": 31545
    },
    {
      "epoch": 16.535639412997902,
      "grad_norm": 0.652252197265625,
      "learning_rate": 4.43370100532211e-06,
      "loss": 0.4601,
      "num_input_tokens_seen": 20642464,
      "step": 31550
    },
    {
      "epoch": 16.53825995807128,
      "grad_norm": 0.5330799221992493,
      "learning_rate": 4.427202247760695e-06,
      "loss": 0.4712,
      "num_input_tokens_seen": 20645760,
      "step": 31555
    },
    {
      "epoch": 16.540880503144653,
      "grad_norm": 0.5641380548477173,
      "learning_rate": 4.420707793793854e-06,
      "loss": 0.3865,
      "num_input_tokens_seen": 20648960,
      "step": 31560
    },
    {
      "epoch": 16.54350104821803,
      "grad_norm": 2.984631299972534,
      "learning_rate": 4.414217644780172e-06,
      "loss": 0.4913,
      "num_input_tokens_seen": 20652000,
      "step": 31565
    },
    {
      "epoch": 16.546121593291403,
      "grad_norm": 0.8132476210594177,
      "learning_rate": 4.407731802077294e-06,
      "loss": 0.4381,
      "num_input_tokens_seen": 20654592,
      "step": 31570
    },
    {
      "epoch": 16.54874213836478,
      "grad_norm": 0.5277867913246155,
      "learning_rate": 4.401250267041996e-06,
      "loss": 0.494,
      "num_input_tokens_seen": 20657472,
      "step": 31575
    },
    {
      "epoch": 16.551362683438157,
      "grad_norm": 0.31731507182121277,
      "learning_rate": 4.394773041030137e-06,
      "loss": 0.4607,
      "num_input_tokens_seen": 20660192,
      "step": 31580
    },
    {
      "epoch": 16.55398322851153,
      "grad_norm": 0.3674197196960449,
      "learning_rate": 4.388300125396685e-06,
      "loss": 0.3832,
      "num_input_tokens_seen": 20663488,
      "step": 31585
    },
    {
      "epoch": 16.556603773584907,
      "grad_norm": 0.44666174054145813,
      "learning_rate": 4.381831521495699e-06,
      "loss": 0.5169,
      "num_input_tokens_seen": 20666080,
      "step": 31590
    },
    {
      "epoch": 16.55922431865828,
      "grad_norm": 0.43076178431510925,
      "learning_rate": 4.37536723068033e-06,
      "loss": 0.3751,
      "num_input_tokens_seen": 20669056,
      "step": 31595
    },
    {
      "epoch": 16.561844863731658,
      "grad_norm": 0.708480179309845,
      "learning_rate": 4.368907254302837e-06,
      "loss": 0.3942,
      "num_input_tokens_seen": 20672320,
      "step": 31600
    },
    {
      "epoch": 16.56446540880503,
      "grad_norm": 0.7347479462623596,
      "learning_rate": 4.3624515937145725e-06,
      "loss": 0.5793,
      "num_input_tokens_seen": 20675424,
      "step": 31605
    },
    {
      "epoch": 16.567085953878408,
      "grad_norm": 0.5431678891181946,
      "learning_rate": 4.3560002502659905e-06,
      "loss": 0.3764,
      "num_input_tokens_seen": 20678624,
      "step": 31610
    },
    {
      "epoch": 16.56970649895178,
      "grad_norm": 0.5136486291885376,
      "learning_rate": 4.349553225306644e-06,
      "loss": 0.5164,
      "num_input_tokens_seen": 20682176,
      "step": 31615
    },
    {
      "epoch": 16.572327044025158,
      "grad_norm": 0.335332453250885,
      "learning_rate": 4.343110520185167e-06,
      "loss": 0.4483,
      "num_input_tokens_seen": 20686688,
      "step": 31620
    },
    {
      "epoch": 16.57494758909853,
      "grad_norm": 0.3576357066631317,
      "learning_rate": 4.336672136249298e-06,
      "loss": 0.386,
      "num_input_tokens_seen": 20689376,
      "step": 31625
    },
    {
      "epoch": 16.57756813417191,
      "grad_norm": 0.4243420660495758,
      "learning_rate": 4.33023807484588e-06,
      "loss": 0.4177,
      "num_input_tokens_seen": 20692960,
      "step": 31630
    },
    {
      "epoch": 16.580188679245282,
      "grad_norm": 0.8579074144363403,
      "learning_rate": 4.323808337320842e-06,
      "loss": 0.4569,
      "num_input_tokens_seen": 20695520,
      "step": 31635
    },
    {
      "epoch": 16.58280922431866,
      "grad_norm": 0.9770468473434448,
      "learning_rate": 4.31738292501922e-06,
      "loss": 0.3995,
      "num_input_tokens_seen": 20699136,
      "step": 31640
    },
    {
      "epoch": 16.585429769392032,
      "grad_norm": 0.3644993305206299,
      "learning_rate": 4.31096183928513e-06,
      "loss": 0.3379,
      "num_input_tokens_seen": 20703104,
      "step": 31645
    },
    {
      "epoch": 16.58805031446541,
      "grad_norm": 0.4357186257839203,
      "learning_rate": 4.304545081461783e-06,
      "loss": 0.4495,
      "num_input_tokens_seen": 20705984,
      "step": 31650
    },
    {
      "epoch": 16.590670859538783,
      "grad_norm": 0.2935159206390381,
      "learning_rate": 4.298132652891501e-06,
      "loss": 0.4175,
      "num_input_tokens_seen": 20712416,
      "step": 31655
    },
    {
      "epoch": 16.59329140461216,
      "grad_norm": 0.5790664553642273,
      "learning_rate": 4.291724554915688e-06,
      "loss": 0.4249,
      "num_input_tokens_seen": 20715456,
      "step": 31660
    },
    {
      "epoch": 16.595911949685533,
      "grad_norm": 0.613131046295166,
      "learning_rate": 4.285320788874853e-06,
      "loss": 0.3777,
      "num_input_tokens_seen": 20718144,
      "step": 31665
    },
    {
      "epoch": 16.59853249475891,
      "grad_norm": 0.4916026294231415,
      "learning_rate": 4.278921356108581e-06,
      "loss": 0.4167,
      "num_input_tokens_seen": 20720640,
      "step": 31670
    },
    {
      "epoch": 16.601153039832283,
      "grad_norm": 0.45291656255722046,
      "learning_rate": 4.272526257955556e-06,
      "loss": 0.4683,
      "num_input_tokens_seen": 20723424,
      "step": 31675
    },
    {
      "epoch": 16.60377358490566,
      "grad_norm": 0.5664371848106384,
      "learning_rate": 4.266135495753567e-06,
      "loss": 0.4598,
      "num_input_tokens_seen": 20726528,
      "step": 31680
    },
    {
      "epoch": 16.606394129979037,
      "grad_norm": 0.9460019469261169,
      "learning_rate": 4.259749070839491e-06,
      "loss": 0.4213,
      "num_input_tokens_seen": 20729216,
      "step": 31685
    },
    {
      "epoch": 16.60901467505241,
      "grad_norm": 0.3723750412464142,
      "learning_rate": 4.253366984549289e-06,
      "loss": 0.5267,
      "num_input_tokens_seen": 20733344,
      "step": 31690
    },
    {
      "epoch": 16.611635220125788,
      "grad_norm": 0.6443468928337097,
      "learning_rate": 4.24698923821803e-06,
      "loss": 0.5256,
      "num_input_tokens_seen": 20736224,
      "step": 31695
    },
    {
      "epoch": 16.61425576519916,
      "grad_norm": 0.5069889426231384,
      "learning_rate": 4.240615833179861e-06,
      "loss": 0.3795,
      "num_input_tokens_seen": 20739424,
      "step": 31700
    },
    {
      "epoch": 16.616876310272538,
      "grad_norm": 0.4824150800704956,
      "learning_rate": 4.234246770768016e-06,
      "loss": 0.3498,
      "num_input_tokens_seen": 20741824,
      "step": 31705
    },
    {
      "epoch": 16.61949685534591,
      "grad_norm": 0.42496511340141296,
      "learning_rate": 4.227882052314841e-06,
      "loss": 0.3789,
      "num_input_tokens_seen": 20744896,
      "step": 31710
    },
    {
      "epoch": 16.622117400419288,
      "grad_norm": 0.5414470434188843,
      "learning_rate": 4.2215216791517575e-06,
      "loss": 0.4037,
      "num_input_tokens_seen": 20747776,
      "step": 31715
    },
    {
      "epoch": 16.62473794549266,
      "grad_norm": 0.4410947561264038,
      "learning_rate": 4.215165652609293e-06,
      "loss": 0.4352,
      "num_input_tokens_seen": 20750496,
      "step": 31720
    },
    {
      "epoch": 16.62735849056604,
      "grad_norm": 0.6110008955001831,
      "learning_rate": 4.208813974017048e-06,
      "loss": 0.4667,
      "num_input_tokens_seen": 20753216,
      "step": 31725
    },
    {
      "epoch": 16.629979035639412,
      "grad_norm": 0.8420256972312927,
      "learning_rate": 4.202466644703712e-06,
      "loss": 0.5327,
      "num_input_tokens_seen": 20756608,
      "step": 31730
    },
    {
      "epoch": 16.63259958071279,
      "grad_norm": 0.6279153227806091,
      "learning_rate": 4.196123665997087e-06,
      "loss": 0.3736,
      "num_input_tokens_seen": 20759264,
      "step": 31735
    },
    {
      "epoch": 16.635220125786162,
      "grad_norm": 0.4356282353401184,
      "learning_rate": 4.189785039224045e-06,
      "loss": 0.335,
      "num_input_tokens_seen": 20762432,
      "step": 31740
    },
    {
      "epoch": 16.63784067085954,
      "grad_norm": 0.7832061052322388,
      "learning_rate": 4.183450765710567e-06,
      "loss": 0.4081,
      "num_input_tokens_seen": 20765760,
      "step": 31745
    },
    {
      "epoch": 16.640461215932913,
      "grad_norm": 0.2880796790122986,
      "learning_rate": 4.177120846781693e-06,
      "loss": 0.5146,
      "num_input_tokens_seen": 20769376,
      "step": 31750
    },
    {
      "epoch": 16.64308176100629,
      "grad_norm": 0.4278738498687744,
      "learning_rate": 4.170795283761583e-06,
      "loss": 0.4708,
      "num_input_tokens_seen": 20772608,
      "step": 31755
    },
    {
      "epoch": 16.645702306079663,
      "grad_norm": 0.5081526041030884,
      "learning_rate": 4.164474077973462e-06,
      "loss": 0.428,
      "num_input_tokens_seen": 20776288,
      "step": 31760
    },
    {
      "epoch": 16.64832285115304,
      "grad_norm": 0.7737395763397217,
      "learning_rate": 4.158157230739659e-06,
      "loss": 0.5026,
      "num_input_tokens_seen": 20780064,
      "step": 31765
    },
    {
      "epoch": 16.650943396226417,
      "grad_norm": 0.5538225769996643,
      "learning_rate": 4.151844743381589e-06,
      "loss": 0.4215,
      "num_input_tokens_seen": 20783040,
      "step": 31770
    },
    {
      "epoch": 16.65356394129979,
      "grad_norm": 0.48369380831718445,
      "learning_rate": 4.145536617219756e-06,
      "loss": 0.4469,
      "num_input_tokens_seen": 20785952,
      "step": 31775
    },
    {
      "epoch": 16.656184486373167,
      "grad_norm": 0.6791942119598389,
      "learning_rate": 4.139232853573735e-06,
      "loss": 0.4626,
      "num_input_tokens_seen": 20788640,
      "step": 31780
    },
    {
      "epoch": 16.65880503144654,
      "grad_norm": 0.4033765494823456,
      "learning_rate": 4.132933453762214e-06,
      "loss": 0.4271,
      "num_input_tokens_seen": 20792832,
      "step": 31785
    },
    {
      "epoch": 16.661425576519918,
      "grad_norm": 0.37199637293815613,
      "learning_rate": 4.126638419102946e-06,
      "loss": 0.3619,
      "num_input_tokens_seen": 20796512,
      "step": 31790
    },
    {
      "epoch": 16.66404612159329,
      "grad_norm": 0.3198448419570923,
      "learning_rate": 4.120347750912784e-06,
      "loss": 0.377,
      "num_input_tokens_seen": 20800160,
      "step": 31795
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.7781898975372314,
      "learning_rate": 4.1140614505076724e-06,
      "loss": 0.3773,
      "num_input_tokens_seen": 20802624,
      "step": 31800
    },
    {
      "epoch": 16.66928721174004,
      "grad_norm": 0.5184923410415649,
      "learning_rate": 4.10777951920262e-06,
      "loss": 0.2807,
      "num_input_tokens_seen": 20805984,
      "step": 31805
    },
    {
      "epoch": 16.671907756813418,
      "grad_norm": 0.3538147807121277,
      "learning_rate": 4.1015019583117455e-06,
      "loss": 0.42,
      "num_input_tokens_seen": 20809504,
      "step": 31810
    },
    {
      "epoch": 16.67452830188679,
      "grad_norm": 0.5316723585128784,
      "learning_rate": 4.095228769148232e-06,
      "loss": 0.4029,
      "num_input_tokens_seen": 20812736,
      "step": 31815
    },
    {
      "epoch": 16.67714884696017,
      "grad_norm": 0.757793664932251,
      "learning_rate": 4.088959953024368e-06,
      "loss": 0.4522,
      "num_input_tokens_seen": 20815680,
      "step": 31820
    },
    {
      "epoch": 16.679769392033542,
      "grad_norm": 0.6430827379226685,
      "learning_rate": 4.082695511251522e-06,
      "loss": 0.378,
      "num_input_tokens_seen": 20818880,
      "step": 31825
    },
    {
      "epoch": 16.68238993710692,
      "grad_norm": 0.5283839702606201,
      "learning_rate": 4.076435445140131e-06,
      "loss": 0.5244,
      "num_input_tokens_seen": 20821632,
      "step": 31830
    },
    {
      "epoch": 16.685010482180292,
      "grad_norm": 0.3327677547931671,
      "learning_rate": 4.070179755999737e-06,
      "loss": 0.4244,
      "num_input_tokens_seen": 20824512,
      "step": 31835
    },
    {
      "epoch": 16.68763102725367,
      "grad_norm": 0.2888546884059906,
      "learning_rate": 4.063928445138962e-06,
      "loss": 0.5843,
      "num_input_tokens_seen": 20828192,
      "step": 31840
    },
    {
      "epoch": 16.690251572327043,
      "grad_norm": 0.33836987614631653,
      "learning_rate": 4.057681513865499e-06,
      "loss": 0.4835,
      "num_input_tokens_seen": 20831072,
      "step": 31845
    },
    {
      "epoch": 16.69287211740042,
      "grad_norm": 0.37318429350852966,
      "learning_rate": 4.0514389634861426e-06,
      "loss": 0.4374,
      "num_input_tokens_seen": 20834880,
      "step": 31850
    },
    {
      "epoch": 16.695492662473793,
      "grad_norm": 0.3806522488594055,
      "learning_rate": 4.045200795306767e-06,
      "loss": 0.3217,
      "num_input_tokens_seen": 20840928,
      "step": 31855
    },
    {
      "epoch": 16.69811320754717,
      "grad_norm": 0.9251434206962585,
      "learning_rate": 4.038967010632314e-06,
      "loss": 0.4365,
      "num_input_tokens_seen": 20843808,
      "step": 31860
    },
    {
      "epoch": 16.700733752620543,
      "grad_norm": 0.6088883876800537,
      "learning_rate": 4.032737610766826e-06,
      "loss": 0.4028,
      "num_input_tokens_seen": 20847264,
      "step": 31865
    },
    {
      "epoch": 16.70335429769392,
      "grad_norm": 0.4941903352737427,
      "learning_rate": 4.026512597013429e-06,
      "loss": 0.445,
      "num_input_tokens_seen": 20851264,
      "step": 31870
    },
    {
      "epoch": 16.705974842767297,
      "grad_norm": 0.532319188117981,
      "learning_rate": 4.020291970674312e-06,
      "loss": 0.4221,
      "num_input_tokens_seen": 20854048,
      "step": 31875
    },
    {
      "epoch": 16.70859538784067,
      "grad_norm": 0.7948506474494934,
      "learning_rate": 4.0140757330507705e-06,
      "loss": 0.3733,
      "num_input_tokens_seen": 20857472,
      "step": 31880
    },
    {
      "epoch": 16.711215932914047,
      "grad_norm": 0.338950514793396,
      "learning_rate": 4.007863885443161e-06,
      "loss": 0.3877,
      "num_input_tokens_seen": 20860352,
      "step": 31885
    },
    {
      "epoch": 16.71383647798742,
      "grad_norm": 0.4580176770687103,
      "learning_rate": 4.001656429150935e-06,
      "loss": 0.3479,
      "num_input_tokens_seen": 20863808,
      "step": 31890
    },
    {
      "epoch": 16.716457023060798,
      "grad_norm": 0.7173757553100586,
      "learning_rate": 3.9954533654726235e-06,
      "loss": 0.5132,
      "num_input_tokens_seen": 20867488,
      "step": 31895
    },
    {
      "epoch": 16.71907756813417,
      "grad_norm": 0.40460142493247986,
      "learning_rate": 3.989254695705841e-06,
      "loss": 0.4624,
      "num_input_tokens_seen": 20870336,
      "step": 31900
    },
    {
      "epoch": 16.721698113207548,
      "grad_norm": 0.33937540650367737,
      "learning_rate": 3.983060421147272e-06,
      "loss": 0.4207,
      "num_input_tokens_seen": 20874624,
      "step": 31905
    },
    {
      "epoch": 16.72431865828092,
      "grad_norm": 0.5154498219490051,
      "learning_rate": 3.976870543092684e-06,
      "loss": 0.5985,
      "num_input_tokens_seen": 20879456,
      "step": 31910
    },
    {
      "epoch": 16.7269392033543,
      "grad_norm": 0.4207884967327118,
      "learning_rate": 3.970685062836932e-06,
      "loss": 0.5529,
      "num_input_tokens_seen": 20883040,
      "step": 31915
    },
    {
      "epoch": 16.729559748427672,
      "grad_norm": 0.36675339937210083,
      "learning_rate": 3.96450398167395e-06,
      "loss": 0.3833,
      "num_input_tokens_seen": 20886304,
      "step": 31920
    },
    {
      "epoch": 16.73218029350105,
      "grad_norm": 0.3940034508705139,
      "learning_rate": 3.958327300896758e-06,
      "loss": 0.3578,
      "num_input_tokens_seen": 20889888,
      "step": 31925
    },
    {
      "epoch": 16.734800838574422,
      "grad_norm": 0.7723459005355835,
      "learning_rate": 3.952155021797429e-06,
      "loss": 0.4322,
      "num_input_tokens_seen": 20892896,
      "step": 31930
    },
    {
      "epoch": 16.7374213836478,
      "grad_norm": 0.4561674892902374,
      "learning_rate": 3.945987145667151e-06,
      "loss": 0.439,
      "num_input_tokens_seen": 20896672,
      "step": 31935
    },
    {
      "epoch": 16.740041928721173,
      "grad_norm": 0.379696786403656,
      "learning_rate": 3.939823673796162e-06,
      "loss": 0.4283,
      "num_input_tokens_seen": 20900128,
      "step": 31940
    },
    {
      "epoch": 16.74266247379455,
      "grad_norm": 0.39330407977104187,
      "learning_rate": 3.933664607473789e-06,
      "loss": 0.4601,
      "num_input_tokens_seen": 20903872,
      "step": 31945
    },
    {
      "epoch": 16.745283018867923,
      "grad_norm": 0.4873635768890381,
      "learning_rate": 3.927509947988445e-06,
      "loss": 0.5374,
      "num_input_tokens_seen": 20907008,
      "step": 31950
    },
    {
      "epoch": 16.7479035639413,
      "grad_norm": 0.5343647599220276,
      "learning_rate": 3.9213596966276164e-06,
      "loss": 0.4898,
      "num_input_tokens_seen": 20910048,
      "step": 31955
    },
    {
      "epoch": 16.750524109014677,
      "grad_norm": 0.3996182978153229,
      "learning_rate": 3.9152138546778625e-06,
      "loss": 0.5083,
      "num_input_tokens_seen": 20914784,
      "step": 31960
    },
    {
      "epoch": 16.75314465408805,
      "grad_norm": 0.40025898814201355,
      "learning_rate": 3.9090724234248175e-06,
      "loss": 0.3846,
      "num_input_tokens_seen": 20918880,
      "step": 31965
    },
    {
      "epoch": 16.755765199161427,
      "grad_norm": 0.3831682801246643,
      "learning_rate": 3.902935404153202e-06,
      "loss": 0.3857,
      "num_input_tokens_seen": 20921696,
      "step": 31970
    },
    {
      "epoch": 16.7583857442348,
      "grad_norm": 0.5832064747810364,
      "learning_rate": 3.896802798146809e-06,
      "loss": 0.4409,
      "num_input_tokens_seen": 20925440,
      "step": 31975
    },
    {
      "epoch": 16.761006289308177,
      "grad_norm": 0.5079379677772522,
      "learning_rate": 3.890674606688524e-06,
      "loss": 0.3916,
      "num_input_tokens_seen": 20928032,
      "step": 31980
    },
    {
      "epoch": 16.76362683438155,
      "grad_norm": 0.628879189491272,
      "learning_rate": 3.88455083106028e-06,
      "loss": 0.4891,
      "num_input_tokens_seen": 20930688,
      "step": 31985
    },
    {
      "epoch": 16.766247379454928,
      "grad_norm": 0.47839048504829407,
      "learning_rate": 3.878431472543098e-06,
      "loss": 0.3641,
      "num_input_tokens_seen": 20933888,
      "step": 31990
    },
    {
      "epoch": 16.7688679245283,
      "grad_norm": 0.4726346433162689,
      "learning_rate": 3.872316532417086e-06,
      "loss": 0.3224,
      "num_input_tokens_seen": 20936896,
      "step": 31995
    },
    {
      "epoch": 16.771488469601678,
      "grad_norm": 0.4154132902622223,
      "learning_rate": 3.866206011961418e-06,
      "loss": 0.4402,
      "num_input_tokens_seen": 20940192,
      "step": 32000
    },
    {
      "epoch": 16.77410901467505,
      "grad_norm": 0.49058085680007935,
      "learning_rate": 3.860099912454346e-06,
      "loss": 0.4439,
      "num_input_tokens_seen": 20943168,
      "step": 32005
    },
    {
      "epoch": 16.77672955974843,
      "grad_norm": 0.43891793489456177,
      "learning_rate": 3.853998235173201e-06,
      "loss": 0.4746,
      "num_input_tokens_seen": 20946272,
      "step": 32010
    },
    {
      "epoch": 16.779350104821802,
      "grad_norm": 0.5684124231338501,
      "learning_rate": 3.847900981394379e-06,
      "loss": 0.4697,
      "num_input_tokens_seen": 20949024,
      "step": 32015
    },
    {
      "epoch": 16.78197064989518,
      "grad_norm": 0.49580979347229004,
      "learning_rate": 3.841808152393348e-06,
      "loss": 0.4498,
      "num_input_tokens_seen": 20952160,
      "step": 32020
    },
    {
      "epoch": 16.784591194968552,
      "grad_norm": 0.5688408017158508,
      "learning_rate": 3.835719749444666e-06,
      "loss": 0.4633,
      "num_input_tokens_seen": 20956192,
      "step": 32025
    },
    {
      "epoch": 16.78721174004193,
      "grad_norm": 0.6283749341964722,
      "learning_rate": 3.8296357738219606e-06,
      "loss": 0.3862,
      "num_input_tokens_seen": 20958816,
      "step": 32030
    },
    {
      "epoch": 16.789832285115303,
      "grad_norm": 0.48406973481178284,
      "learning_rate": 3.823556226797931e-06,
      "loss": 0.4093,
      "num_input_tokens_seen": 20961600,
      "step": 32035
    },
    {
      "epoch": 16.79245283018868,
      "grad_norm": 0.3691200315952301,
      "learning_rate": 3.817481109644344e-06,
      "loss": 0.547,
      "num_input_tokens_seen": 20964832,
      "step": 32040
    },
    {
      "epoch": 16.795073375262053,
      "grad_norm": 0.6538154482841492,
      "learning_rate": 3.811410423632042e-06,
      "loss": 0.4735,
      "num_input_tokens_seen": 20967232,
      "step": 32045
    },
    {
      "epoch": 16.79769392033543,
      "grad_norm": 0.5988561511039734,
      "learning_rate": 3.8053441700309473e-06,
      "loss": 0.4626,
      "num_input_tokens_seen": 20969952,
      "step": 32050
    },
    {
      "epoch": 16.800314465408803,
      "grad_norm": 0.5223396420478821,
      "learning_rate": 3.7992823501100507e-06,
      "loss": 0.4332,
      "num_input_tokens_seen": 20973856,
      "step": 32055
    },
    {
      "epoch": 16.80293501048218,
      "grad_norm": 0.29701298475265503,
      "learning_rate": 3.7932249651374207e-06,
      "loss": 0.5366,
      "num_input_tokens_seen": 20977472,
      "step": 32060
    },
    {
      "epoch": 16.805555555555557,
      "grad_norm": 0.3706912696361542,
      "learning_rate": 3.787172016380183e-06,
      "loss": 0.3975,
      "num_input_tokens_seen": 20981248,
      "step": 32065
    },
    {
      "epoch": 16.80817610062893,
      "grad_norm": 0.3999658524990082,
      "learning_rate": 3.7811235051045558e-06,
      "loss": 0.4354,
      "num_input_tokens_seen": 20985024,
      "step": 32070
    },
    {
      "epoch": 16.810796645702307,
      "grad_norm": 0.3868328034877777,
      "learning_rate": 3.775079432575809e-06,
      "loss": 0.3759,
      "num_input_tokens_seen": 20988128,
      "step": 32075
    },
    {
      "epoch": 16.81341719077568,
      "grad_norm": 0.6481267213821411,
      "learning_rate": 3.7690398000582964e-06,
      "loss": 0.4638,
      "num_input_tokens_seen": 20991552,
      "step": 32080
    },
    {
      "epoch": 16.816037735849058,
      "grad_norm": 0.9439833760261536,
      "learning_rate": 3.763004608815443e-06,
      "loss": 0.5439,
      "num_input_tokens_seen": 20995296,
      "step": 32085
    },
    {
      "epoch": 16.81865828092243,
      "grad_norm": 0.4769347608089447,
      "learning_rate": 3.756973860109747e-06,
      "loss": 0.3312,
      "num_input_tokens_seen": 20999136,
      "step": 32090
    },
    {
      "epoch": 16.821278825995808,
      "grad_norm": 0.5357992649078369,
      "learning_rate": 3.7509475552027663e-06,
      "loss": 0.4577,
      "num_input_tokens_seen": 21002592,
      "step": 32095
    },
    {
      "epoch": 16.82389937106918,
      "grad_norm": 0.3707086145877838,
      "learning_rate": 3.744925695355131e-06,
      "loss": 0.4278,
      "num_input_tokens_seen": 21005440,
      "step": 32100
    },
    {
      "epoch": 16.82651991614256,
      "grad_norm": 0.48106324672698975,
      "learning_rate": 3.7389082818265476e-06,
      "loss": 0.4715,
      "num_input_tokens_seen": 21008832,
      "step": 32105
    },
    {
      "epoch": 16.829140461215932,
      "grad_norm": 0.6020572185516357,
      "learning_rate": 3.732895315875792e-06,
      "loss": 0.3489,
      "num_input_tokens_seen": 21012128,
      "step": 32110
    },
    {
      "epoch": 16.83176100628931,
      "grad_norm": 0.7015138268470764,
      "learning_rate": 3.726886798760715e-06,
      "loss": 0.413,
      "num_input_tokens_seen": 21014944,
      "step": 32115
    },
    {
      "epoch": 16.834381551362682,
      "grad_norm": 0.2698211967945099,
      "learning_rate": 3.720882731738215e-06,
      "loss": 0.4011,
      "num_input_tokens_seen": 21018784,
      "step": 32120
    },
    {
      "epoch": 16.83700209643606,
      "grad_norm": 0.47351229190826416,
      "learning_rate": 3.714883116064291e-06,
      "loss": 0.3761,
      "num_input_tokens_seen": 21021728,
      "step": 32125
    },
    {
      "epoch": 16.839622641509433,
      "grad_norm": 0.48234108090400696,
      "learning_rate": 3.708887952993978e-06,
      "loss": 0.3607,
      "num_input_tokens_seen": 21024768,
      "step": 32130
    },
    {
      "epoch": 16.84224318658281,
      "grad_norm": 0.578090250492096,
      "learning_rate": 3.7028972437814057e-06,
      "loss": 0.4665,
      "num_input_tokens_seen": 21031456,
      "step": 32135
    },
    {
      "epoch": 16.844863731656183,
      "grad_norm": 0.3675305247306824,
      "learning_rate": 3.696910989679764e-06,
      "loss": 0.4133,
      "num_input_tokens_seen": 21034624,
      "step": 32140
    },
    {
      "epoch": 16.84748427672956,
      "grad_norm": 0.33652108907699585,
      "learning_rate": 3.6909291919413003e-06,
      "loss": 0.4273,
      "num_input_tokens_seen": 21038560,
      "step": 32145
    },
    {
      "epoch": 16.850104821802937,
      "grad_norm": 0.41234156489372253,
      "learning_rate": 3.684951851817342e-06,
      "loss": 0.4247,
      "num_input_tokens_seen": 21041920,
      "step": 32150
    },
    {
      "epoch": 16.85272536687631,
      "grad_norm": 0.454893559217453,
      "learning_rate": 3.678978970558286e-06,
      "loss": 0.3151,
      "num_input_tokens_seen": 21045344,
      "step": 32155
    },
    {
      "epoch": 16.855345911949687,
      "grad_norm": 0.5231283903121948,
      "learning_rate": 3.673010549413583e-06,
      "loss": 0.385,
      "num_input_tokens_seen": 21047936,
      "step": 32160
    },
    {
      "epoch": 16.85796645702306,
      "grad_norm": 0.48156681656837463,
      "learning_rate": 3.6670465896317592e-06,
      "loss": 0.5271,
      "num_input_tokens_seen": 21050752,
      "step": 32165
    },
    {
      "epoch": 16.860587002096437,
      "grad_norm": 0.5831772685050964,
      "learning_rate": 3.661087092460419e-06,
      "loss": 0.3733,
      "num_input_tokens_seen": 21054112,
      "step": 32170
    },
    {
      "epoch": 16.86320754716981,
      "grad_norm": 0.46967813372612,
      "learning_rate": 3.655132059146202e-06,
      "loss": 0.5531,
      "num_input_tokens_seen": 21056800,
      "step": 32175
    },
    {
      "epoch": 16.865828092243188,
      "grad_norm": 0.5502031445503235,
      "learning_rate": 3.6491814909348526e-06,
      "loss": 0.3726,
      "num_input_tokens_seen": 21059712,
      "step": 32180
    },
    {
      "epoch": 16.86844863731656,
      "grad_norm": 0.6878927946090698,
      "learning_rate": 3.6432353890711424e-06,
      "loss": 0.4911,
      "num_input_tokens_seen": 21062304,
      "step": 32185
    },
    {
      "epoch": 16.871069182389938,
      "grad_norm": 0.531534731388092,
      "learning_rate": 3.6372937547989425e-06,
      "loss": 0.4186,
      "num_input_tokens_seen": 21065408,
      "step": 32190
    },
    {
      "epoch": 16.87368972746331,
      "grad_norm": 0.6209862232208252,
      "learning_rate": 3.631356589361173e-06,
      "loss": 0.587,
      "num_input_tokens_seen": 21069216,
      "step": 32195
    },
    {
      "epoch": 16.87631027253669,
      "grad_norm": 0.6256406903266907,
      "learning_rate": 3.625423893999813e-06,
      "loss": 0.4161,
      "num_input_tokens_seen": 21071872,
      "step": 32200
    },
    {
      "epoch": 16.878930817610062,
      "grad_norm": 0.35580167174339294,
      "learning_rate": 3.6194956699559217e-06,
      "loss": 0.3865,
      "num_input_tokens_seen": 21075552,
      "step": 32205
    },
    {
      "epoch": 16.88155136268344,
      "grad_norm": 0.5250661969184875,
      "learning_rate": 3.6135719184696188e-06,
      "loss": 0.3497,
      "num_input_tokens_seen": 21078656,
      "step": 32210
    },
    {
      "epoch": 16.884171907756812,
      "grad_norm": 0.5841913819313049,
      "learning_rate": 3.6076526407800766e-06,
      "loss": 0.3719,
      "num_input_tokens_seen": 21081216,
      "step": 32215
    },
    {
      "epoch": 16.88679245283019,
      "grad_norm": 0.414320707321167,
      "learning_rate": 3.601737838125552e-06,
      "loss": 0.4107,
      "num_input_tokens_seen": 21084096,
      "step": 32220
    },
    {
      "epoch": 16.889412997903563,
      "grad_norm": 0.43545961380004883,
      "learning_rate": 3.595827511743341e-06,
      "loss": 0.3559,
      "num_input_tokens_seen": 21087232,
      "step": 32225
    },
    {
      "epoch": 16.89203354297694,
      "grad_norm": 0.3661132752895355,
      "learning_rate": 3.5899216628698217e-06,
      "loss": 0.4565,
      "num_input_tokens_seen": 21090272,
      "step": 32230
    },
    {
      "epoch": 16.894654088050313,
      "grad_norm": 0.3780471086502075,
      "learning_rate": 3.584020292740431e-06,
      "loss": 0.436,
      "num_input_tokens_seen": 21093152,
      "step": 32235
    },
    {
      "epoch": 16.89727463312369,
      "grad_norm": 0.7671961784362793,
      "learning_rate": 3.578123402589675e-06,
      "loss": 0.5441,
      "num_input_tokens_seen": 21096224,
      "step": 32240
    },
    {
      "epoch": 16.899895178197063,
      "grad_norm": 0.5047379732131958,
      "learning_rate": 3.5722309936511044e-06,
      "loss": 0.4639,
      "num_input_tokens_seen": 21099072,
      "step": 32245
    },
    {
      "epoch": 16.90251572327044,
      "grad_norm": 0.40828385949134827,
      "learning_rate": 3.5663430671573572e-06,
      "loss": 0.437,
      "num_input_tokens_seen": 21101888,
      "step": 32250
    },
    {
      "epoch": 16.905136268343817,
      "grad_norm": 0.2961065471172333,
      "learning_rate": 3.5604596243401038e-06,
      "loss": 0.4498,
      "num_input_tokens_seen": 21105152,
      "step": 32255
    },
    {
      "epoch": 16.90775681341719,
      "grad_norm": 0.42540937662124634,
      "learning_rate": 3.554580666430102e-06,
      "loss": 0.5262,
      "num_input_tokens_seen": 21108576,
      "step": 32260
    },
    {
      "epoch": 16.910377358490567,
      "grad_norm": 0.25047096610069275,
      "learning_rate": 3.5487061946571614e-06,
      "loss": 0.4908,
      "num_input_tokens_seen": 21112192,
      "step": 32265
    },
    {
      "epoch": 16.91299790356394,
      "grad_norm": 0.5377659797668457,
      "learning_rate": 3.542836210250161e-06,
      "loss": 0.3893,
      "num_input_tokens_seen": 21115104,
      "step": 32270
    },
    {
      "epoch": 16.915618448637318,
      "grad_norm": 0.7207348942756653,
      "learning_rate": 3.536970714437032e-06,
      "loss": 0.4473,
      "num_input_tokens_seen": 21118624,
      "step": 32275
    },
    {
      "epoch": 16.91823899371069,
      "grad_norm": 0.270657479763031,
      "learning_rate": 3.5311097084447575e-06,
      "loss": 0.3161,
      "num_input_tokens_seen": 21122784,
      "step": 32280
    },
    {
      "epoch": 16.920859538784068,
      "grad_norm": 0.9862434267997742,
      "learning_rate": 3.5252531934994033e-06,
      "loss": 0.2642,
      "num_input_tokens_seen": 21125920,
      "step": 32285
    },
    {
      "epoch": 16.92348008385744,
      "grad_norm": 0.35338878631591797,
      "learning_rate": 3.5194011708260866e-06,
      "loss": 0.4692,
      "num_input_tokens_seen": 21131488,
      "step": 32290
    },
    {
      "epoch": 16.92610062893082,
      "grad_norm": 0.46745336055755615,
      "learning_rate": 3.513553641648984e-06,
      "loss": 0.5032,
      "num_input_tokens_seen": 21134688,
      "step": 32295
    },
    {
      "epoch": 16.928721174004192,
      "grad_norm": 0.47729843854904175,
      "learning_rate": 3.5077106071913307e-06,
      "loss": 0.4046,
      "num_input_tokens_seen": 21137248,
      "step": 32300
    },
    {
      "epoch": 16.93134171907757,
      "grad_norm": 0.39625850319862366,
      "learning_rate": 3.5018720686754133e-06,
      "loss": 0.5325,
      "num_input_tokens_seen": 21140800,
      "step": 32305
    },
    {
      "epoch": 16.933962264150942,
      "grad_norm": 0.6207007169723511,
      "learning_rate": 3.4960380273225985e-06,
      "loss": 0.5075,
      "num_input_tokens_seen": 21144032,
      "step": 32310
    },
    {
      "epoch": 16.93658280922432,
      "grad_norm": 0.37603694200515747,
      "learning_rate": 3.490208484353297e-06,
      "loss": 0.3708,
      "num_input_tokens_seen": 21146528,
      "step": 32315
    },
    {
      "epoch": 16.939203354297693,
      "grad_norm": 0.963897705078125,
      "learning_rate": 3.4843834409869824e-06,
      "loss": 0.4514,
      "num_input_tokens_seen": 21149728,
      "step": 32320
    },
    {
      "epoch": 16.94182389937107,
      "grad_norm": 0.659470796585083,
      "learning_rate": 3.4785628984421977e-06,
      "loss": 0.4952,
      "num_input_tokens_seen": 21153568,
      "step": 32325
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 0.45368295907974243,
      "learning_rate": 3.4727468579365235e-06,
      "loss": 0.4746,
      "num_input_tokens_seen": 21158144,
      "step": 32330
    },
    {
      "epoch": 16.94706498951782,
      "grad_norm": 0.4564393162727356,
      "learning_rate": 3.466935320686604e-06,
      "loss": 0.4879,
      "num_input_tokens_seen": 21161344,
      "step": 32335
    },
    {
      "epoch": 16.949685534591197,
      "grad_norm": 0.29880866408348083,
      "learning_rate": 3.4611282879081553e-06,
      "loss": 0.3968,
      "num_input_tokens_seen": 21166368,
      "step": 32340
    },
    {
      "epoch": 16.95230607966457,
      "grad_norm": 0.342331200838089,
      "learning_rate": 3.455325760815939e-06,
      "loss": 0.6118,
      "num_input_tokens_seen": 21169440,
      "step": 32345
    },
    {
      "epoch": 16.954926624737947,
      "grad_norm": 0.4390330910682678,
      "learning_rate": 3.4495277406237843e-06,
      "loss": 0.4305,
      "num_input_tokens_seen": 21172544,
      "step": 32350
    },
    {
      "epoch": 16.95754716981132,
      "grad_norm": 0.3524971604347229,
      "learning_rate": 3.443734228544565e-06,
      "loss": 0.358,
      "num_input_tokens_seen": 21176192,
      "step": 32355
    },
    {
      "epoch": 16.960167714884697,
      "grad_norm": 0.5181728601455688,
      "learning_rate": 3.437945225790215e-06,
      "loss": 0.4628,
      "num_input_tokens_seen": 21179424,
      "step": 32360
    },
    {
      "epoch": 16.96278825995807,
      "grad_norm": 0.4308306872844696,
      "learning_rate": 3.432160733571729e-06,
      "loss": 0.3335,
      "num_input_tokens_seen": 21182272,
      "step": 32365
    },
    {
      "epoch": 16.965408805031448,
      "grad_norm": 0.5533532500267029,
      "learning_rate": 3.4263807530991595e-06,
      "loss": 0.5102,
      "num_input_tokens_seen": 21186048,
      "step": 32370
    },
    {
      "epoch": 16.96802935010482,
      "grad_norm": 0.40043431520462036,
      "learning_rate": 3.4206052855816155e-06,
      "loss": 0.3944,
      "num_input_tokens_seen": 21189024,
      "step": 32375
    },
    {
      "epoch": 16.970649895178198,
      "grad_norm": 0.4411877691745758,
      "learning_rate": 3.4148343322272567e-06,
      "loss": 0.4611,
      "num_input_tokens_seen": 21192576,
      "step": 32380
    },
    {
      "epoch": 16.97327044025157,
      "grad_norm": 0.48949113488197327,
      "learning_rate": 3.4090678942432923e-06,
      "loss": 0.4724,
      "num_input_tokens_seen": 21195616,
      "step": 32385
    },
    {
      "epoch": 16.97589098532495,
      "grad_norm": 0.31310877203941345,
      "learning_rate": 3.4033059728360018e-06,
      "loss": 0.4021,
      "num_input_tokens_seen": 21199584,
      "step": 32390
    },
    {
      "epoch": 16.978511530398322,
      "grad_norm": 0.2097359299659729,
      "learning_rate": 3.397548569210715e-06,
      "loss": 0.4712,
      "num_input_tokens_seen": 21203008,
      "step": 32395
    },
    {
      "epoch": 16.9811320754717,
      "grad_norm": 0.6086761951446533,
      "learning_rate": 3.3917956845718136e-06,
      "loss": 0.4679,
      "num_input_tokens_seen": 21206080,
      "step": 32400
    },
    {
      "epoch": 16.983752620545072,
      "grad_norm": 0.7328060865402222,
      "learning_rate": 3.38604732012274e-06,
      "loss": 0.5348,
      "num_input_tokens_seen": 21209088,
      "step": 32405
    },
    {
      "epoch": 16.98637316561845,
      "grad_norm": 0.5537043809890747,
      "learning_rate": 3.3803034770659825e-06,
      "loss": 0.4175,
      "num_input_tokens_seen": 21212288,
      "step": 32410
    },
    {
      "epoch": 16.988993710691823,
      "grad_norm": 0.6845318078994751,
      "learning_rate": 3.3745641566030826e-06,
      "loss": 0.4674,
      "num_input_tokens_seen": 21214560,
      "step": 32415
    },
    {
      "epoch": 16.9916142557652,
      "grad_norm": 0.851909339427948,
      "learning_rate": 3.3688293599346442e-06,
      "loss": 0.7622,
      "num_input_tokens_seen": 21217056,
      "step": 32420
    },
    {
      "epoch": 16.994234800838573,
      "grad_norm": 0.36063042283058167,
      "learning_rate": 3.3630990882603213e-06,
      "loss": 0.4581,
      "num_input_tokens_seen": 21220224,
      "step": 32425
    },
    {
      "epoch": 16.99685534591195,
      "grad_norm": 0.42470911145210266,
      "learning_rate": 3.35737334277883e-06,
      "loss": 0.5133,
      "num_input_tokens_seen": 21223360,
      "step": 32430
    },
    {
      "epoch": 16.999475890985323,
      "grad_norm": 0.47530892491340637,
      "learning_rate": 3.351652124687918e-06,
      "loss": 0.3909,
      "num_input_tokens_seen": 21226496,
      "step": 32435
    },
    {
      "epoch": 17.0,
      "eval_loss": 0.447039395570755,
      "eval_runtime": 9.6852,
      "eval_samples_per_second": 87.556,
      "eval_steps_per_second": 21.889,
      "num_input_tokens_seen": 21226656,
      "step": 32436
    },
    {
      "epoch": 17.0020964360587,
      "grad_norm": 0.5019704103469849,
      "learning_rate": 3.345935435184408e-06,
      "loss": 0.4093,
      "num_input_tokens_seen": 21228800,
      "step": 32440
    },
    {
      "epoch": 17.004716981132077,
      "grad_norm": 0.40689224004745483,
      "learning_rate": 3.3402232754641576e-06,
      "loss": 0.6353,
      "num_input_tokens_seen": 21232416,
      "step": 32445
    },
    {
      "epoch": 17.00733752620545,
      "grad_norm": 0.433848112821579,
      "learning_rate": 3.334515646722089e-06,
      "loss": 0.4058,
      "num_input_tokens_seen": 21235168,
      "step": 32450
    },
    {
      "epoch": 17.009958071278827,
      "grad_norm": 0.4280632734298706,
      "learning_rate": 3.32881255015218e-06,
      "loss": 0.3805,
      "num_input_tokens_seen": 21237952,
      "step": 32455
    },
    {
      "epoch": 17.0125786163522,
      "grad_norm": 0.6122974157333374,
      "learning_rate": 3.3231139869474433e-06,
      "loss": 0.4084,
      "num_input_tokens_seen": 21241056,
      "step": 32460
    },
    {
      "epoch": 17.015199161425578,
      "grad_norm": 0.5183130502700806,
      "learning_rate": 3.317419958299961e-06,
      "loss": 0.3614,
      "num_input_tokens_seen": 21244960,
      "step": 32465
    },
    {
      "epoch": 17.01781970649895,
      "grad_norm": 0.42902159690856934,
      "learning_rate": 3.31173046540085e-06,
      "loss": 0.53,
      "num_input_tokens_seen": 21247936,
      "step": 32470
    },
    {
      "epoch": 17.020440251572328,
      "grad_norm": 0.4426177442073822,
      "learning_rate": 3.306045509440295e-06,
      "loss": 0.4402,
      "num_input_tokens_seen": 21251104,
      "step": 32475
    },
    {
      "epoch": 17.0230607966457,
      "grad_norm": 0.4009924829006195,
      "learning_rate": 3.300365091607521e-06,
      "loss": 0.4905,
      "num_input_tokens_seen": 21255040,
      "step": 32480
    },
    {
      "epoch": 17.02568134171908,
      "grad_norm": 0.5355899333953857,
      "learning_rate": 3.2946892130908127e-06,
      "loss": 0.4717,
      "num_input_tokens_seen": 21258336,
      "step": 32485
    },
    {
      "epoch": 17.028301886792452,
      "grad_norm": 1.380940556526184,
      "learning_rate": 3.289017875077488e-06,
      "loss": 0.5407,
      "num_input_tokens_seen": 21261760,
      "step": 32490
    },
    {
      "epoch": 17.03092243186583,
      "grad_norm": 0.3584223985671997,
      "learning_rate": 3.283351078753938e-06,
      "loss": 0.4132,
      "num_input_tokens_seen": 21265056,
      "step": 32495
    },
    {
      "epoch": 17.033542976939202,
      "grad_norm": 0.8243399262428284,
      "learning_rate": 3.2776888253055777e-06,
      "loss": 0.5544,
      "num_input_tokens_seen": 21271744,
      "step": 32500
    },
    {
      "epoch": 17.03616352201258,
      "grad_norm": 0.43807554244995117,
      "learning_rate": 3.272031115916896e-06,
      "loss": 0.4739,
      "num_input_tokens_seen": 21274368,
      "step": 32505
    },
    {
      "epoch": 17.038784067085953,
      "grad_norm": 0.743483304977417,
      "learning_rate": 3.2663779517714238e-06,
      "loss": 0.4154,
      "num_input_tokens_seen": 21277120,
      "step": 32510
    },
    {
      "epoch": 17.04140461215933,
      "grad_norm": 0.5496719479560852,
      "learning_rate": 3.260729334051729e-06,
      "loss": 0.4304,
      "num_input_tokens_seen": 21281632,
      "step": 32515
    },
    {
      "epoch": 17.044025157232703,
      "grad_norm": 0.2623957097530365,
      "learning_rate": 3.2550852639394446e-06,
      "loss": 0.4502,
      "num_input_tokens_seen": 21284160,
      "step": 32520
    },
    {
      "epoch": 17.04664570230608,
      "grad_norm": 0.5719408988952637,
      "learning_rate": 3.2494457426152452e-06,
      "loss": 0.5375,
      "num_input_tokens_seen": 21286752,
      "step": 32525
    },
    {
      "epoch": 17.049266247379453,
      "grad_norm": 0.4844493269920349,
      "learning_rate": 3.243810771258851e-06,
      "loss": 0.5715,
      "num_input_tokens_seen": 21290016,
      "step": 32530
    },
    {
      "epoch": 17.05188679245283,
      "grad_norm": 0.46455878019332886,
      "learning_rate": 3.2381803510490387e-06,
      "loss": 0.4112,
      "num_input_tokens_seen": 21292832,
      "step": 32535
    },
    {
      "epoch": 17.054507337526207,
      "grad_norm": 0.4452871084213257,
      "learning_rate": 3.2325544831636217e-06,
      "loss": 0.3622,
      "num_input_tokens_seen": 21296032,
      "step": 32540
    },
    {
      "epoch": 17.05712788259958,
      "grad_norm": 0.7087856531143188,
      "learning_rate": 3.22693316877947e-06,
      "loss": 0.4834,
      "num_input_tokens_seen": 21298624,
      "step": 32545
    },
    {
      "epoch": 17.059748427672957,
      "grad_norm": 0.6034220457077026,
      "learning_rate": 3.2213164090725057e-06,
      "loss": 0.3271,
      "num_input_tokens_seen": 21301472,
      "step": 32550
    },
    {
      "epoch": 17.06236897274633,
      "grad_norm": 0.6762192249298096,
      "learning_rate": 3.2157042052176763e-06,
      "loss": 0.5773,
      "num_input_tokens_seen": 21305600,
      "step": 32555
    },
    {
      "epoch": 17.064989517819708,
      "grad_norm": 0.49233075976371765,
      "learning_rate": 3.2100965583890003e-06,
      "loss": 0.4354,
      "num_input_tokens_seen": 21308064,
      "step": 32560
    },
    {
      "epoch": 17.06761006289308,
      "grad_norm": 0.42070525884628296,
      "learning_rate": 3.204493469759537e-06,
      "loss": 0.4371,
      "num_input_tokens_seen": 21311200,
      "step": 32565
    },
    {
      "epoch": 17.070230607966458,
      "grad_norm": 0.750119686126709,
      "learning_rate": 3.1988949405013764e-06,
      "loss": 0.4559,
      "num_input_tokens_seen": 21314528,
      "step": 32570
    },
    {
      "epoch": 17.07285115303983,
      "grad_norm": 0.545648992061615,
      "learning_rate": 3.193300971785676e-06,
      "loss": 0.4552,
      "num_input_tokens_seen": 21317952,
      "step": 32575
    },
    {
      "epoch": 17.07547169811321,
      "grad_norm": 1.1753641366958618,
      "learning_rate": 3.1877115647826335e-06,
      "loss": 0.3899,
      "num_input_tokens_seen": 21321440,
      "step": 32580
    },
    {
      "epoch": 17.078092243186582,
      "grad_norm": 0.4370664358139038,
      "learning_rate": 3.1821267206614797e-06,
      "loss": 0.4704,
      "num_input_tokens_seen": 21324704,
      "step": 32585
    },
    {
      "epoch": 17.08071278825996,
      "grad_norm": 0.49513164162635803,
      "learning_rate": 3.176546440590508e-06,
      "loss": 0.3364,
      "num_input_tokens_seen": 21327040,
      "step": 32590
    },
    {
      "epoch": 17.083333333333332,
      "grad_norm": 0.5447585582733154,
      "learning_rate": 3.170970725737041e-06,
      "loss": 0.4104,
      "num_input_tokens_seen": 21329824,
      "step": 32595
    },
    {
      "epoch": 17.08595387840671,
      "grad_norm": 0.3169194161891937,
      "learning_rate": 3.165399577267461e-06,
      "loss": 0.4162,
      "num_input_tokens_seen": 21333024,
      "step": 32600
    },
    {
      "epoch": 17.088574423480082,
      "grad_norm": 0.4338512122631073,
      "learning_rate": 3.159832996347184e-06,
      "loss": 0.425,
      "num_input_tokens_seen": 21336480,
      "step": 32605
    },
    {
      "epoch": 17.09119496855346,
      "grad_norm": 0.5654332637786865,
      "learning_rate": 3.1542709841406865e-06,
      "loss": 0.4035,
      "num_input_tokens_seen": 21340160,
      "step": 32610
    },
    {
      "epoch": 17.093815513626833,
      "grad_norm": 0.42813533544540405,
      "learning_rate": 3.148713541811468e-06,
      "loss": 0.3902,
      "num_input_tokens_seen": 21344096,
      "step": 32615
    },
    {
      "epoch": 17.09643605870021,
      "grad_norm": 0.2969253659248352,
      "learning_rate": 3.1431606705220803e-06,
      "loss": 0.3819,
      "num_input_tokens_seen": 21347712,
      "step": 32620
    },
    {
      "epoch": 17.099056603773583,
      "grad_norm": 0.3470085561275482,
      "learning_rate": 3.1376123714341234e-06,
      "loss": 0.4547,
      "num_input_tokens_seen": 21351168,
      "step": 32625
    },
    {
      "epoch": 17.10167714884696,
      "grad_norm": 0.7615252137184143,
      "learning_rate": 3.1320686457082387e-06,
      "loss": 0.3782,
      "num_input_tokens_seen": 21354016,
      "step": 32630
    },
    {
      "epoch": 17.104297693920337,
      "grad_norm": 0.558835506439209,
      "learning_rate": 3.126529494504113e-06,
      "loss": 0.6858,
      "num_input_tokens_seen": 21357440,
      "step": 32635
    },
    {
      "epoch": 17.10691823899371,
      "grad_norm": 0.5224504470825195,
      "learning_rate": 3.1209949189804744e-06,
      "loss": 0.5409,
      "num_input_tokens_seen": 21360512,
      "step": 32640
    },
    {
      "epoch": 17.109538784067087,
      "grad_norm": 0.6796633005142212,
      "learning_rate": 3.115464920295089e-06,
      "loss": 0.3928,
      "num_input_tokens_seen": 21363040,
      "step": 32645
    },
    {
      "epoch": 17.11215932914046,
      "grad_norm": 0.4077489674091339,
      "learning_rate": 3.1099394996047666e-06,
      "loss": 0.4255,
      "num_input_tokens_seen": 21365568,
      "step": 32650
    },
    {
      "epoch": 17.114779874213838,
      "grad_norm": 0.5582059621810913,
      "learning_rate": 3.104418658065367e-06,
      "loss": 0.5303,
      "num_input_tokens_seen": 21368960,
      "step": 32655
    },
    {
      "epoch": 17.11740041928721,
      "grad_norm": 0.7966893911361694,
      "learning_rate": 3.098902396831782e-06,
      "loss": 0.5243,
      "num_input_tokens_seen": 21371712,
      "step": 32660
    },
    {
      "epoch": 17.120020964360588,
      "grad_norm": 0.5740167498588562,
      "learning_rate": 3.0933907170579646e-06,
      "loss": 0.4164,
      "num_input_tokens_seen": 21375040,
      "step": 32665
    },
    {
      "epoch": 17.12264150943396,
      "grad_norm": 0.36961671710014343,
      "learning_rate": 3.0878836198968835e-06,
      "loss": 0.4063,
      "num_input_tokens_seen": 21378880,
      "step": 32670
    },
    {
      "epoch": 17.12526205450734,
      "grad_norm": 0.5819514393806458,
      "learning_rate": 3.0823811065005564e-06,
      "loss": 0.4199,
      "num_input_tokens_seen": 21382048,
      "step": 32675
    },
    {
      "epoch": 17.127882599580712,
      "grad_norm": 0.3179068863391876,
      "learning_rate": 3.0768831780200542e-06,
      "loss": 0.3985,
      "num_input_tokens_seen": 21385888,
      "step": 32680
    },
    {
      "epoch": 17.13050314465409,
      "grad_norm": 0.3389877378940582,
      "learning_rate": 3.071389835605476e-06,
      "loss": 0.4149,
      "num_input_tokens_seen": 21390208,
      "step": 32685
    },
    {
      "epoch": 17.133123689727462,
      "grad_norm": 0.858227550983429,
      "learning_rate": 3.065901080405978e-06,
      "loss": 0.3666,
      "num_input_tokens_seen": 21393824,
      "step": 32690
    },
    {
      "epoch": 17.13574423480084,
      "grad_norm": 0.43540921807289124,
      "learning_rate": 3.060416913569733e-06,
      "loss": 0.3657,
      "num_input_tokens_seen": 21396960,
      "step": 32695
    },
    {
      "epoch": 17.138364779874212,
      "grad_norm": 0.5117927193641663,
      "learning_rate": 3.054937336243968e-06,
      "loss": 0.5416,
      "num_input_tokens_seen": 21400832,
      "step": 32700
    },
    {
      "epoch": 17.14098532494759,
      "grad_norm": 0.3180830776691437,
      "learning_rate": 3.049462349574947e-06,
      "loss": 0.3999,
      "num_input_tokens_seen": 21405120,
      "step": 32705
    },
    {
      "epoch": 17.143605870020963,
      "grad_norm": 1.010928750038147,
      "learning_rate": 3.043991954707981e-06,
      "loss": 0.4394,
      "num_input_tokens_seen": 21407840,
      "step": 32710
    },
    {
      "epoch": 17.14622641509434,
      "grad_norm": 0.3410692811012268,
      "learning_rate": 3.0385261527874075e-06,
      "loss": 0.4446,
      "num_input_tokens_seen": 21411616,
      "step": 32715
    },
    {
      "epoch": 17.148846960167713,
      "grad_norm": 0.5885851979255676,
      "learning_rate": 3.0330649449566205e-06,
      "loss": 0.3521,
      "num_input_tokens_seen": 21414528,
      "step": 32720
    },
    {
      "epoch": 17.15146750524109,
      "grad_norm": 0.47314485907554626,
      "learning_rate": 3.027608332358034e-06,
      "loss": 0.421,
      "num_input_tokens_seen": 21417344,
      "step": 32725
    },
    {
      "epoch": 17.154088050314467,
      "grad_norm": 0.9173288345336914,
      "learning_rate": 3.022156316133104e-06,
      "loss": 0.4843,
      "num_input_tokens_seen": 21420000,
      "step": 32730
    },
    {
      "epoch": 17.15670859538784,
      "grad_norm": 0.6383436322212219,
      "learning_rate": 3.016708897422335e-06,
      "loss": 0.4629,
      "num_input_tokens_seen": 21423104,
      "step": 32735
    },
    {
      "epoch": 17.159329140461217,
      "grad_norm": 0.5104482769966125,
      "learning_rate": 3.0112660773652685e-06,
      "loss": 0.4535,
      "num_input_tokens_seen": 21427744,
      "step": 32740
    },
    {
      "epoch": 17.16194968553459,
      "grad_norm": 0.3472250699996948,
      "learning_rate": 3.005827857100482e-06,
      "loss": 0.4122,
      "num_input_tokens_seen": 21430976,
      "step": 32745
    },
    {
      "epoch": 17.164570230607968,
      "grad_norm": 0.5796278119087219,
      "learning_rate": 3.000394237765586e-06,
      "loss": 0.5066,
      "num_input_tokens_seen": 21433664,
      "step": 32750
    },
    {
      "epoch": 17.16719077568134,
      "grad_norm": 0.7630361914634705,
      "learning_rate": 2.9949652204972254e-06,
      "loss": 0.527,
      "num_input_tokens_seen": 21436384,
      "step": 32755
    },
    {
      "epoch": 17.169811320754718,
      "grad_norm": 0.4317401349544525,
      "learning_rate": 2.989540806431093e-06,
      "loss": 0.4327,
      "num_input_tokens_seen": 21439520,
      "step": 32760
    },
    {
      "epoch": 17.17243186582809,
      "grad_norm": 0.4025495648384094,
      "learning_rate": 2.984120996701917e-06,
      "loss": 0.3438,
      "num_input_tokens_seen": 21445504,
      "step": 32765
    },
    {
      "epoch": 17.17505241090147,
      "grad_norm": 0.5187468528747559,
      "learning_rate": 2.9787057924434618e-06,
      "loss": 0.4299,
      "num_input_tokens_seen": 21451264,
      "step": 32770
    },
    {
      "epoch": 17.177672955974842,
      "grad_norm": 0.30707910656929016,
      "learning_rate": 2.9732951947885166e-06,
      "loss": 0.4489,
      "num_input_tokens_seen": 21454144,
      "step": 32775
    },
    {
      "epoch": 17.18029350104822,
      "grad_norm": 0.30943354964256287,
      "learning_rate": 2.967889204868929e-06,
      "loss": 0.3972,
      "num_input_tokens_seen": 21456928,
      "step": 32780
    },
    {
      "epoch": 17.182914046121592,
      "grad_norm": 0.38745948672294617,
      "learning_rate": 2.96248782381556e-06,
      "loss": 0.349,
      "num_input_tokens_seen": 21460032,
      "step": 32785
    },
    {
      "epoch": 17.18553459119497,
      "grad_norm": 0.8635214567184448,
      "learning_rate": 2.957091052758321e-06,
      "loss": 0.5007,
      "num_input_tokens_seen": 21463040,
      "step": 32790
    },
    {
      "epoch": 17.188155136268342,
      "grad_norm": 0.560364305973053,
      "learning_rate": 2.9516988928261517e-06,
      "loss": 0.8009,
      "num_input_tokens_seen": 21466336,
      "step": 32795
    },
    {
      "epoch": 17.19077568134172,
      "grad_norm": 0.6375240683555603,
      "learning_rate": 2.9463113451470403e-06,
      "loss": 0.4452,
      "num_input_tokens_seen": 21469440,
      "step": 32800
    },
    {
      "epoch": 17.193396226415093,
      "grad_norm": 0.6453613638877869,
      "learning_rate": 2.94092841084799e-06,
      "loss": 0.4555,
      "num_input_tokens_seen": 21472288,
      "step": 32805
    },
    {
      "epoch": 17.19601677148847,
      "grad_norm": 0.580665111541748,
      "learning_rate": 2.9355500910550586e-06,
      "loss": 0.4224,
      "num_input_tokens_seen": 21475136,
      "step": 32810
    },
    {
      "epoch": 17.198637316561843,
      "grad_norm": 0.5486705303192139,
      "learning_rate": 2.9301763868933157e-06,
      "loss": 0.4467,
      "num_input_tokens_seen": 21477632,
      "step": 32815
    },
    {
      "epoch": 17.20125786163522,
      "grad_norm": 0.3360903859138489,
      "learning_rate": 2.9248072994868896e-06,
      "loss": 0.4543,
      "num_input_tokens_seen": 21481120,
      "step": 32820
    },
    {
      "epoch": 17.203878406708597,
      "grad_norm": 1.1874891519546509,
      "learning_rate": 2.919442829958932e-06,
      "loss": 0.3627,
      "num_input_tokens_seen": 21483872,
      "step": 32825
    },
    {
      "epoch": 17.20649895178197,
      "grad_norm": 0.2750365138053894,
      "learning_rate": 2.9140829794316234e-06,
      "loss": 0.4173,
      "num_input_tokens_seen": 21487840,
      "step": 32830
    },
    {
      "epoch": 17.209119496855347,
      "grad_norm": 0.4401134252548218,
      "learning_rate": 2.908727749026191e-06,
      "loss": 0.3733,
      "num_input_tokens_seen": 21490880,
      "step": 32835
    },
    {
      "epoch": 17.21174004192872,
      "grad_norm": 0.4366152286529541,
      "learning_rate": 2.9033771398628756e-06,
      "loss": 0.3978,
      "num_input_tokens_seen": 21494656,
      "step": 32840
    },
    {
      "epoch": 17.214360587002098,
      "grad_norm": 0.4531736373901367,
      "learning_rate": 2.898031153060973e-06,
      "loss": 0.355,
      "num_input_tokens_seen": 21497696,
      "step": 32845
    },
    {
      "epoch": 17.21698113207547,
      "grad_norm": 0.4809952676296234,
      "learning_rate": 2.8926897897388033e-06,
      "loss": 0.4864,
      "num_input_tokens_seen": 21500640,
      "step": 32850
    },
    {
      "epoch": 17.219601677148848,
      "grad_norm": 0.4675936698913574,
      "learning_rate": 2.8873530510137136e-06,
      "loss": 0.4909,
      "num_input_tokens_seen": 21503808,
      "step": 32855
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 0.5558352470397949,
      "learning_rate": 2.8820209380020892e-06,
      "loss": 0.6433,
      "num_input_tokens_seen": 21506944,
      "step": 32860
    },
    {
      "epoch": 17.2248427672956,
      "grad_norm": 0.8826240301132202,
      "learning_rate": 2.8766934518193566e-06,
      "loss": 0.5091,
      "num_input_tokens_seen": 21509824,
      "step": 32865
    },
    {
      "epoch": 17.22746331236897,
      "grad_norm": 0.8654409646987915,
      "learning_rate": 2.87137059357995e-06,
      "loss": 0.359,
      "num_input_tokens_seen": 21512864,
      "step": 32870
    },
    {
      "epoch": 17.23008385744235,
      "grad_norm": 0.5465738773345947,
      "learning_rate": 2.8660523643973618e-06,
      "loss": 0.3901,
      "num_input_tokens_seen": 21516320,
      "step": 32875
    },
    {
      "epoch": 17.232704402515722,
      "grad_norm": 0.7214975357055664,
      "learning_rate": 2.860738765384105e-06,
      "loss": 0.4547,
      "num_input_tokens_seen": 21522240,
      "step": 32880
    },
    {
      "epoch": 17.2353249475891,
      "grad_norm": 0.5900150537490845,
      "learning_rate": 2.8554297976517154e-06,
      "loss": 0.5055,
      "num_input_tokens_seen": 21525952,
      "step": 32885
    },
    {
      "epoch": 17.237945492662472,
      "grad_norm": 0.32187139987945557,
      "learning_rate": 2.8501254623107707e-06,
      "loss": 0.338,
      "num_input_tokens_seen": 21530528,
      "step": 32890
    },
    {
      "epoch": 17.24056603773585,
      "grad_norm": 0.8199909925460815,
      "learning_rate": 2.84482576047089e-06,
      "loss": 0.6563,
      "num_input_tokens_seen": 21534400,
      "step": 32895
    },
    {
      "epoch": 17.243186582809223,
      "grad_norm": 0.38558366894721985,
      "learning_rate": 2.839530693240694e-06,
      "loss": 0.3552,
      "num_input_tokens_seen": 21537248,
      "step": 32900
    },
    {
      "epoch": 17.2458071278826,
      "grad_norm": 0.4131259024143219,
      "learning_rate": 2.8342402617278606e-06,
      "loss": 0.4731,
      "num_input_tokens_seen": 21539808,
      "step": 32905
    },
    {
      "epoch": 17.248427672955973,
      "grad_norm": 0.797429621219635,
      "learning_rate": 2.82895446703908e-06,
      "loss": 0.4086,
      "num_input_tokens_seen": 21546656,
      "step": 32910
    },
    {
      "epoch": 17.25104821802935,
      "grad_norm": 0.4478086233139038,
      "learning_rate": 2.823673310280084e-06,
      "loss": 0.463,
      "num_input_tokens_seen": 21549152,
      "step": 32915
    },
    {
      "epoch": 17.253668763102727,
      "grad_norm": 0.36984163522720337,
      "learning_rate": 2.8183967925556305e-06,
      "loss": 0.4523,
      "num_input_tokens_seen": 21552832,
      "step": 32920
    },
    {
      "epoch": 17.2562893081761,
      "grad_norm": 0.6594117879867554,
      "learning_rate": 2.8131249149695115e-06,
      "loss": 0.3737,
      "num_input_tokens_seen": 21555776,
      "step": 32925
    },
    {
      "epoch": 17.258909853249477,
      "grad_norm": 0.8259166479110718,
      "learning_rate": 2.8078576786245397e-06,
      "loss": 0.428,
      "num_input_tokens_seen": 21558624,
      "step": 32930
    },
    {
      "epoch": 17.26153039832285,
      "grad_norm": 0.27178698778152466,
      "learning_rate": 2.8025950846225557e-06,
      "loss": 0.2717,
      "num_input_tokens_seen": 21562432,
      "step": 32935
    },
    {
      "epoch": 17.264150943396228,
      "grad_norm": 0.501035213470459,
      "learning_rate": 2.7973371340644373e-06,
      "loss": 0.3966,
      "num_input_tokens_seen": 21565856,
      "step": 32940
    },
    {
      "epoch": 17.2667714884696,
      "grad_norm": 0.4439050555229187,
      "learning_rate": 2.7920838280500907e-06,
      "loss": 0.3578,
      "num_input_tokens_seen": 21569312,
      "step": 32945
    },
    {
      "epoch": 17.269392033542978,
      "grad_norm": 0.5390201807022095,
      "learning_rate": 2.7868351676784504e-06,
      "loss": 0.4738,
      "num_input_tokens_seen": 21573152,
      "step": 32950
    },
    {
      "epoch": 17.27201257861635,
      "grad_norm": 0.49923911690711975,
      "learning_rate": 2.7815911540474665e-06,
      "loss": 0.5315,
      "num_input_tokens_seen": 21575840,
      "step": 32955
    },
    {
      "epoch": 17.27463312368973,
      "grad_norm": 0.38264185190200806,
      "learning_rate": 2.7763517882541364e-06,
      "loss": 0.3945,
      "num_input_tokens_seen": 21578304,
      "step": 32960
    },
    {
      "epoch": 17.2772536687631,
      "grad_norm": 0.6327530145645142,
      "learning_rate": 2.771117071394466e-06,
      "loss": 0.5865,
      "num_input_tokens_seen": 21580832,
      "step": 32965
    },
    {
      "epoch": 17.27987421383648,
      "grad_norm": 0.44543537497520447,
      "learning_rate": 2.7658870045635055e-06,
      "loss": 0.5031,
      "num_input_tokens_seen": 21583456,
      "step": 32970
    },
    {
      "epoch": 17.282494758909852,
      "grad_norm": 0.8955784440040588,
      "learning_rate": 2.7606615888553234e-06,
      "loss": 0.4151,
      "num_input_tokens_seen": 21585856,
      "step": 32975
    },
    {
      "epoch": 17.28511530398323,
      "grad_norm": 0.41472357511520386,
      "learning_rate": 2.755440825363023e-06,
      "loss": 0.4579,
      "num_input_tokens_seen": 21588960,
      "step": 32980
    },
    {
      "epoch": 17.287735849056602,
      "grad_norm": 0.6966187357902527,
      "learning_rate": 2.75022471517872e-06,
      "loss": 0.4527,
      "num_input_tokens_seen": 21591904,
      "step": 32985
    },
    {
      "epoch": 17.29035639412998,
      "grad_norm": 0.6087362766265869,
      "learning_rate": 2.745013259393564e-06,
      "loss": 0.4684,
      "num_input_tokens_seen": 21594848,
      "step": 32990
    },
    {
      "epoch": 17.292976939203353,
      "grad_norm": 0.5217923521995544,
      "learning_rate": 2.739806459097735e-06,
      "loss": 0.3934,
      "num_input_tokens_seen": 21598112,
      "step": 32995
    },
    {
      "epoch": 17.29559748427673,
      "grad_norm": 0.5266857147216797,
      "learning_rate": 2.7346043153804403e-06,
      "loss": 0.5437,
      "num_input_tokens_seen": 21601312,
      "step": 33000
    },
    {
      "epoch": 17.298218029350103,
      "grad_norm": 0.5973474979400635,
      "learning_rate": 2.729406829329903e-06,
      "loss": 0.4836,
      "num_input_tokens_seen": 21604160,
      "step": 33005
    },
    {
      "epoch": 17.30083857442348,
      "grad_norm": 0.41584232449531555,
      "learning_rate": 2.72421400203339e-06,
      "loss": 0.4235,
      "num_input_tokens_seen": 21607008,
      "step": 33010
    },
    {
      "epoch": 17.303459119496857,
      "grad_norm": 0.5807555913925171,
      "learning_rate": 2.719025834577163e-06,
      "loss": 0.3314,
      "num_input_tokens_seen": 21610144,
      "step": 33015
    },
    {
      "epoch": 17.30607966457023,
      "grad_norm": 0.377699613571167,
      "learning_rate": 2.713842328046537e-06,
      "loss": 0.5894,
      "num_input_tokens_seen": 21613312,
      "step": 33020
    },
    {
      "epoch": 17.308700209643607,
      "grad_norm": 0.9521290063858032,
      "learning_rate": 2.7086634835258416e-06,
      "loss": 0.4927,
      "num_input_tokens_seen": 21615936,
      "step": 33025
    },
    {
      "epoch": 17.31132075471698,
      "grad_norm": 0.5332695245742798,
      "learning_rate": 2.7034893020984333e-06,
      "loss": 0.5507,
      "num_input_tokens_seen": 21619200,
      "step": 33030
    },
    {
      "epoch": 17.313941299790358,
      "grad_norm": 0.4523714780807495,
      "learning_rate": 2.6983197848466952e-06,
      "loss": 0.4292,
      "num_input_tokens_seen": 21621792,
      "step": 33035
    },
    {
      "epoch": 17.31656184486373,
      "grad_norm": 0.3692931532859802,
      "learning_rate": 2.6931549328520265e-06,
      "loss": 0.3908,
      "num_input_tokens_seen": 21624448,
      "step": 33040
    },
    {
      "epoch": 17.319182389937108,
      "grad_norm": 0.46639513969421387,
      "learning_rate": 2.6879947471948523e-06,
      "loss": 0.3428,
      "num_input_tokens_seen": 21628576,
      "step": 33045
    },
    {
      "epoch": 17.32180293501048,
      "grad_norm": 0.34795087575912476,
      "learning_rate": 2.6828392289546285e-06,
      "loss": 0.4947,
      "num_input_tokens_seen": 21632128,
      "step": 33050
    },
    {
      "epoch": 17.32442348008386,
      "grad_norm": 0.4385092556476593,
      "learning_rate": 2.677688379209828e-06,
      "loss": 0.4568,
      "num_input_tokens_seen": 21635616,
      "step": 33055
    },
    {
      "epoch": 17.32704402515723,
      "grad_norm": 0.403150349855423,
      "learning_rate": 2.6725421990379586e-06,
      "loss": 0.3203,
      "num_input_tokens_seen": 21638592,
      "step": 33060
    },
    {
      "epoch": 17.32966457023061,
      "grad_norm": 0.4237409830093384,
      "learning_rate": 2.667400689515537e-06,
      "loss": 0.4022,
      "num_input_tokens_seen": 21641920,
      "step": 33065
    },
    {
      "epoch": 17.332285115303982,
      "grad_norm": 0.42828235030174255,
      "learning_rate": 2.6622638517181e-06,
      "loss": 0.5717,
      "num_input_tokens_seen": 21645024,
      "step": 33070
    },
    {
      "epoch": 17.33490566037736,
      "grad_norm": 0.9592474699020386,
      "learning_rate": 2.657131686720224e-06,
      "loss": 0.6297,
      "num_input_tokens_seen": 21647904,
      "step": 33075
    },
    {
      "epoch": 17.337526205450732,
      "grad_norm": 0.6049651503562927,
      "learning_rate": 2.652004195595498e-06,
      "loss": 0.56,
      "num_input_tokens_seen": 21650528,
      "step": 33080
    },
    {
      "epoch": 17.34014675052411,
      "grad_norm": 0.5173696279525757,
      "learning_rate": 2.6468813794165356e-06,
      "loss": 0.5306,
      "num_input_tokens_seen": 21653664,
      "step": 33085
    },
    {
      "epoch": 17.342767295597483,
      "grad_norm": 0.6979979872703552,
      "learning_rate": 2.641763239254977e-06,
      "loss": 0.5503,
      "num_input_tokens_seen": 21656640,
      "step": 33090
    },
    {
      "epoch": 17.34538784067086,
      "grad_norm": 0.8582804799079895,
      "learning_rate": 2.6366497761814713e-06,
      "loss": 0.4133,
      "num_input_tokens_seen": 21659616,
      "step": 33095
    },
    {
      "epoch": 17.348008385744233,
      "grad_norm": 0.49031034111976624,
      "learning_rate": 2.631540991265691e-06,
      "loss": 0.485,
      "num_input_tokens_seen": 21663040,
      "step": 33100
    },
    {
      "epoch": 17.35062893081761,
      "grad_norm": 0.5728985071182251,
      "learning_rate": 2.6264368855763443e-06,
      "loss": 0.5027,
      "num_input_tokens_seen": 21666688,
      "step": 33105
    },
    {
      "epoch": 17.353249475890987,
      "grad_norm": 0.9124631285667419,
      "learning_rate": 2.6213374601811504e-06,
      "loss": 0.5305,
      "num_input_tokens_seen": 21669440,
      "step": 33110
    },
    {
      "epoch": 17.35587002096436,
      "grad_norm": 0.4306730628013611,
      "learning_rate": 2.616242716146855e-06,
      "loss": 0.4311,
      "num_input_tokens_seen": 21672448,
      "step": 33115
    },
    {
      "epoch": 17.358490566037737,
      "grad_norm": 0.3922063708305359,
      "learning_rate": 2.6111526545392145e-06,
      "loss": 0.5206,
      "num_input_tokens_seen": 21675488,
      "step": 33120
    },
    {
      "epoch": 17.36111111111111,
      "grad_norm": 0.4256207346916199,
      "learning_rate": 2.6060672764230106e-06,
      "loss": 0.4485,
      "num_input_tokens_seen": 21678272,
      "step": 33125
    },
    {
      "epoch": 17.363731656184488,
      "grad_norm": 0.5395570993423462,
      "learning_rate": 2.600986582862047e-06,
      "loss": 0.3952,
      "num_input_tokens_seen": 21681088,
      "step": 33130
    },
    {
      "epoch": 17.36635220125786,
      "grad_norm": 0.5080147981643677,
      "learning_rate": 2.5959105749191498e-06,
      "loss": 0.4859,
      "num_input_tokens_seen": 21683872,
      "step": 33135
    },
    {
      "epoch": 17.368972746331238,
      "grad_norm": 0.6545677185058594,
      "learning_rate": 2.590839253656166e-06,
      "loss": 0.5059,
      "num_input_tokens_seen": 21687104,
      "step": 33140
    },
    {
      "epoch": 17.37159329140461,
      "grad_norm": 0.35070404410362244,
      "learning_rate": 2.5857726201339473e-06,
      "loss": 0.4157,
      "num_input_tokens_seen": 21689952,
      "step": 33145
    },
    {
      "epoch": 17.37421383647799,
      "grad_norm": 0.40355125069618225,
      "learning_rate": 2.5807106754123873e-06,
      "loss": 0.5455,
      "num_input_tokens_seen": 21693088,
      "step": 33150
    },
    {
      "epoch": 17.37683438155136,
      "grad_norm": 0.5266485214233398,
      "learning_rate": 2.575653420550378e-06,
      "loss": 0.4615,
      "num_input_tokens_seen": 21695872,
      "step": 33155
    },
    {
      "epoch": 17.37945492662474,
      "grad_norm": 0.5887631773948669,
      "learning_rate": 2.57060085660584e-06,
      "loss": 0.5682,
      "num_input_tokens_seen": 21698272,
      "step": 33160
    },
    {
      "epoch": 17.382075471698112,
      "grad_norm": 0.4866378903388977,
      "learning_rate": 2.565552984635716e-06,
      "loss": 0.4329,
      "num_input_tokens_seen": 21701984,
      "step": 33165
    },
    {
      "epoch": 17.38469601677149,
      "grad_norm": 0.34377121925354004,
      "learning_rate": 2.5605098056959652e-06,
      "loss": 0.4362,
      "num_input_tokens_seen": 21705856,
      "step": 33170
    },
    {
      "epoch": 17.387316561844862,
      "grad_norm": 0.8538755178451538,
      "learning_rate": 2.555471320841557e-06,
      "loss": 0.3419,
      "num_input_tokens_seen": 21708160,
      "step": 33175
    },
    {
      "epoch": 17.38993710691824,
      "grad_norm": 0.4210233986377716,
      "learning_rate": 2.55043753112649e-06,
      "loss": 0.4691,
      "num_input_tokens_seen": 21713760,
      "step": 33180
    },
    {
      "epoch": 17.392557651991613,
      "grad_norm": 0.5067949891090393,
      "learning_rate": 2.5454084376037717e-06,
      "loss": 0.4747,
      "num_input_tokens_seen": 21716768,
      "step": 33185
    },
    {
      "epoch": 17.39517819706499,
      "grad_norm": 0.45925724506378174,
      "learning_rate": 2.5403840413254304e-06,
      "loss": 0.491,
      "num_input_tokens_seen": 21720128,
      "step": 33190
    },
    {
      "epoch": 17.397798742138363,
      "grad_norm": 0.4982985854148865,
      "learning_rate": 2.5353643433425196e-06,
      "loss": 0.4106,
      "num_input_tokens_seen": 21724384,
      "step": 33195
    },
    {
      "epoch": 17.40041928721174,
      "grad_norm": 0.7326173782348633,
      "learning_rate": 2.5303493447050965e-06,
      "loss": 0.4428,
      "num_input_tokens_seen": 21727136,
      "step": 33200
    },
    {
      "epoch": 17.403039832285117,
      "grad_norm": 0.34189626574516296,
      "learning_rate": 2.525339046462244e-06,
      "loss": 0.3772,
      "num_input_tokens_seen": 21731136,
      "step": 33205
    },
    {
      "epoch": 17.40566037735849,
      "grad_norm": 0.7094179391860962,
      "learning_rate": 2.5203334496620544e-06,
      "loss": 0.4603,
      "num_input_tokens_seen": 21734272,
      "step": 33210
    },
    {
      "epoch": 17.408280922431867,
      "grad_norm": 0.6853969693183899,
      "learning_rate": 2.515332555351646e-06,
      "loss": 0.4187,
      "num_input_tokens_seen": 21738112,
      "step": 33215
    },
    {
      "epoch": 17.41090146750524,
      "grad_norm": 0.33736541867256165,
      "learning_rate": 2.510336364577154e-06,
      "loss": 0.5742,
      "num_input_tokens_seen": 21740896,
      "step": 33220
    },
    {
      "epoch": 17.413522012578618,
      "grad_norm": 0.5534178614616394,
      "learning_rate": 2.5053448783837143e-06,
      "loss": 0.4967,
      "num_input_tokens_seen": 21744448,
      "step": 33225
    },
    {
      "epoch": 17.41614255765199,
      "grad_norm": 0.5589147210121155,
      "learning_rate": 2.5003580978154917e-06,
      "loss": 0.3788,
      "num_input_tokens_seen": 21747200,
      "step": 33230
    },
    {
      "epoch": 17.418763102725368,
      "grad_norm": 0.6440497040748596,
      "learning_rate": 2.4953760239156716e-06,
      "loss": 0.4778,
      "num_input_tokens_seen": 21750432,
      "step": 33235
    },
    {
      "epoch": 17.42138364779874,
      "grad_norm": 0.45061540603637695,
      "learning_rate": 2.4903986577264366e-06,
      "loss": 0.5036,
      "num_input_tokens_seen": 21753856,
      "step": 33240
    },
    {
      "epoch": 17.42400419287212,
      "grad_norm": 0.6369563341140747,
      "learning_rate": 2.4854260002889986e-06,
      "loss": 0.3865,
      "num_input_tokens_seen": 21756960,
      "step": 33245
    },
    {
      "epoch": 17.42662473794549,
      "grad_norm": 0.4119724929332733,
      "learning_rate": 2.480458052643586e-06,
      "loss": 0.3915,
      "num_input_tokens_seen": 21759968,
      "step": 33250
    },
    {
      "epoch": 17.42924528301887,
      "grad_norm": 0.42356714606285095,
      "learning_rate": 2.4754948158294293e-06,
      "loss": 0.4147,
      "num_input_tokens_seen": 21762592,
      "step": 33255
    },
    {
      "epoch": 17.431865828092242,
      "grad_norm": 0.6916909217834473,
      "learning_rate": 2.4705362908847814e-06,
      "loss": 0.327,
      "num_input_tokens_seen": 21765440,
      "step": 33260
    },
    {
      "epoch": 17.43448637316562,
      "grad_norm": 0.7173689007759094,
      "learning_rate": 2.4655824788469172e-06,
      "loss": 0.499,
      "num_input_tokens_seen": 21768736,
      "step": 33265
    },
    {
      "epoch": 17.437106918238992,
      "grad_norm": 0.48000985383987427,
      "learning_rate": 2.4606333807521087e-06,
      "loss": 0.4173,
      "num_input_tokens_seen": 21771392,
      "step": 33270
    },
    {
      "epoch": 17.43972746331237,
      "grad_norm": 0.5741320848464966,
      "learning_rate": 2.455688997635658e-06,
      "loss": 0.4343,
      "num_input_tokens_seen": 21774176,
      "step": 33275
    },
    {
      "epoch": 17.442348008385743,
      "grad_norm": 0.3952997922897339,
      "learning_rate": 2.4507493305318652e-06,
      "loss": 0.388,
      "num_input_tokens_seen": 21776736,
      "step": 33280
    },
    {
      "epoch": 17.44496855345912,
      "grad_norm": 0.5590067505836487,
      "learning_rate": 2.445814380474057e-06,
      "loss": 0.4188,
      "num_input_tokens_seen": 21780480,
      "step": 33285
    },
    {
      "epoch": 17.447589098532493,
      "grad_norm": 0.3224560022354126,
      "learning_rate": 2.4408841484945684e-06,
      "loss": 0.4504,
      "num_input_tokens_seen": 21784032,
      "step": 33290
    },
    {
      "epoch": 17.45020964360587,
      "grad_norm": 0.3982502222061157,
      "learning_rate": 2.435958635624755e-06,
      "loss": 0.3882,
      "num_input_tokens_seen": 21787584,
      "step": 33295
    },
    {
      "epoch": 17.452830188679247,
      "grad_norm": 0.5192642211914062,
      "learning_rate": 2.4310378428949678e-06,
      "loss": 0.4017,
      "num_input_tokens_seen": 21790336,
      "step": 33300
    },
    {
      "epoch": 17.45545073375262,
      "grad_norm": 0.8627792000770569,
      "learning_rate": 2.4261217713345804e-06,
      "loss": 0.5516,
      "num_input_tokens_seen": 21793216,
      "step": 33305
    },
    {
      "epoch": 17.458071278825997,
      "grad_norm": 0.7528723478317261,
      "learning_rate": 2.421210421971981e-06,
      "loss": 0.4793,
      "num_input_tokens_seen": 21796160,
      "step": 33310
    },
    {
      "epoch": 17.46069182389937,
      "grad_norm": 0.5449042320251465,
      "learning_rate": 2.416303795834571e-06,
      "loss": 0.4682,
      "num_input_tokens_seen": 21798656,
      "step": 33315
    },
    {
      "epoch": 17.463312368972748,
      "grad_norm": 0.7827796936035156,
      "learning_rate": 2.4114018939487615e-06,
      "loss": 0.3481,
      "num_input_tokens_seen": 21802912,
      "step": 33320
    },
    {
      "epoch": 17.46593291404612,
      "grad_norm": 0.6374076008796692,
      "learning_rate": 2.4065047173399662e-06,
      "loss": 0.3769,
      "num_input_tokens_seen": 21806304,
      "step": 33325
    },
    {
      "epoch": 17.468553459119498,
      "grad_norm": 0.401988685131073,
      "learning_rate": 2.4016122670326295e-06,
      "loss": 0.5524,
      "num_input_tokens_seen": 21809760,
      "step": 33330
    },
    {
      "epoch": 17.47117400419287,
      "grad_norm": 0.3970303237438202,
      "learning_rate": 2.396724544050186e-06,
      "loss": 0.3456,
      "num_input_tokens_seen": 21812576,
      "step": 33335
    },
    {
      "epoch": 17.47379454926625,
      "grad_norm": 0.8157826066017151,
      "learning_rate": 2.3918415494150944e-06,
      "loss": 0.3946,
      "num_input_tokens_seen": 21815392,
      "step": 33340
    },
    {
      "epoch": 17.47641509433962,
      "grad_norm": 0.6784214973449707,
      "learning_rate": 2.3869632841488245e-06,
      "loss": 0.4129,
      "num_input_tokens_seen": 21817920,
      "step": 33345
    },
    {
      "epoch": 17.479035639413,
      "grad_norm": 0.6277274489402771,
      "learning_rate": 2.382089749271857e-06,
      "loss": 0.4434,
      "num_input_tokens_seen": 21820704,
      "step": 33350
    },
    {
      "epoch": 17.481656184486372,
      "grad_norm": 0.5848276615142822,
      "learning_rate": 2.3772209458036737e-06,
      "loss": 0.3537,
      "num_input_tokens_seen": 21823872,
      "step": 33355
    },
    {
      "epoch": 17.48427672955975,
      "grad_norm": 0.7656492590904236,
      "learning_rate": 2.372356874762771e-06,
      "loss": 0.4292,
      "num_input_tokens_seen": 21827584,
      "step": 33360
    },
    {
      "epoch": 17.486897274633122,
      "grad_norm": 0.5253178477287292,
      "learning_rate": 2.36749753716666e-06,
      "loss": 0.3877,
      "num_input_tokens_seen": 21830624,
      "step": 33365
    },
    {
      "epoch": 17.4895178197065,
      "grad_norm": 0.2509162724018097,
      "learning_rate": 2.3626429340318613e-06,
      "loss": 0.4176,
      "num_input_tokens_seen": 21834208,
      "step": 33370
    },
    {
      "epoch": 17.492138364779873,
      "grad_norm": 0.44745469093322754,
      "learning_rate": 2.357793066373901e-06,
      "loss": 0.6212,
      "num_input_tokens_seen": 21837408,
      "step": 33375
    },
    {
      "epoch": 17.49475890985325,
      "grad_norm": 0.4942464530467987,
      "learning_rate": 2.352947935207317e-06,
      "loss": 0.3725,
      "num_input_tokens_seen": 21840160,
      "step": 33380
    },
    {
      "epoch": 17.497379454926623,
      "grad_norm": 1.1585922241210938,
      "learning_rate": 2.3481075415456495e-06,
      "loss": 0.4095,
      "num_input_tokens_seen": 21844288,
      "step": 33385
    },
    {
      "epoch": 17.5,
      "grad_norm": 1.2268109321594238,
      "learning_rate": 2.3432718864014576e-06,
      "loss": 0.602,
      "num_input_tokens_seen": 21846880,
      "step": 33390
    },
    {
      "epoch": 17.502620545073377,
      "grad_norm": 0.481096088886261,
      "learning_rate": 2.338440970786304e-06,
      "loss": 0.4744,
      "num_input_tokens_seen": 21850048,
      "step": 33395
    },
    {
      "epoch": 17.50524109014675,
      "grad_norm": 0.35129544138908386,
      "learning_rate": 2.3336147957107644e-06,
      "loss": 0.3432,
      "num_input_tokens_seen": 21853088,
      "step": 33400
    },
    {
      "epoch": 17.507861635220127,
      "grad_norm": 0.8899948000907898,
      "learning_rate": 2.3287933621844227e-06,
      "loss": 0.3933,
      "num_input_tokens_seen": 21856000,
      "step": 33405
    },
    {
      "epoch": 17.5104821802935,
      "grad_norm": 0.4876570403575897,
      "learning_rate": 2.3239766712158633e-06,
      "loss": 0.364,
      "num_input_tokens_seen": 21859392,
      "step": 33410
    },
    {
      "epoch": 17.513102725366878,
      "grad_norm": 0.26935824751853943,
      "learning_rate": 2.319164723812675e-06,
      "loss": 0.4049,
      "num_input_tokens_seen": 21862912,
      "step": 33415
    },
    {
      "epoch": 17.51572327044025,
      "grad_norm": 0.4572947919368744,
      "learning_rate": 2.314357520981472e-06,
      "loss": 0.5067,
      "num_input_tokens_seen": 21866016,
      "step": 33420
    },
    {
      "epoch": 17.518343815513628,
      "grad_norm": 0.459669828414917,
      "learning_rate": 2.309555063727864e-06,
      "loss": 0.5091,
      "num_input_tokens_seen": 21869248,
      "step": 33425
    },
    {
      "epoch": 17.520964360587,
      "grad_norm": 0.6093915104866028,
      "learning_rate": 2.304757353056475e-06,
      "loss": 0.4362,
      "num_input_tokens_seen": 21871968,
      "step": 33430
    },
    {
      "epoch": 17.52358490566038,
      "grad_norm": 0.47679403424263,
      "learning_rate": 2.2999643899709267e-06,
      "loss": 0.5432,
      "num_input_tokens_seen": 21875456,
      "step": 33435
    },
    {
      "epoch": 17.52620545073375,
      "grad_norm": 0.9928296208381653,
      "learning_rate": 2.2951761754738456e-06,
      "loss": 0.342,
      "num_input_tokens_seen": 21877792,
      "step": 33440
    },
    {
      "epoch": 17.52882599580713,
      "grad_norm": 0.47804147005081177,
      "learning_rate": 2.29039271056688e-06,
      "loss": 0.479,
      "num_input_tokens_seen": 21881440,
      "step": 33445
    },
    {
      "epoch": 17.531446540880502,
      "grad_norm": 0.6744857430458069,
      "learning_rate": 2.285613996250674e-06,
      "loss": 0.359,
      "num_input_tokens_seen": 21884256,
      "step": 33450
    },
    {
      "epoch": 17.53406708595388,
      "grad_norm": 0.5458667278289795,
      "learning_rate": 2.2808400335248865e-06,
      "loss": 0.4473,
      "num_input_tokens_seen": 21887136,
      "step": 33455
    },
    {
      "epoch": 17.536687631027252,
      "grad_norm": 0.5542755126953125,
      "learning_rate": 2.2760708233881654e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 21889568,
      "step": 33460
    },
    {
      "epoch": 17.53930817610063,
      "grad_norm": 0.5737400054931641,
      "learning_rate": 2.271306366838183e-06,
      "loss": 0.5899,
      "num_input_tokens_seen": 21892864,
      "step": 33465
    },
    {
      "epoch": 17.541928721174003,
      "grad_norm": 0.9453790783882141,
      "learning_rate": 2.266546664871602e-06,
      "loss": 0.4988,
      "num_input_tokens_seen": 21895968,
      "step": 33470
    },
    {
      "epoch": 17.54454926624738,
      "grad_norm": 0.45392748713493347,
      "learning_rate": 2.261791718484102e-06,
      "loss": 0.5294,
      "num_input_tokens_seen": 21898912,
      "step": 33475
    },
    {
      "epoch": 17.547169811320753,
      "grad_norm": 0.4470554292201996,
      "learning_rate": 2.257041528670362e-06,
      "loss": 0.3712,
      "num_input_tokens_seen": 21903456,
      "step": 33480
    },
    {
      "epoch": 17.54979035639413,
      "grad_norm": 0.4192371070384979,
      "learning_rate": 2.2522960964240753e-06,
      "loss": 0.47,
      "num_input_tokens_seen": 21906336,
      "step": 33485
    },
    {
      "epoch": 17.552410901467507,
      "grad_norm": 0.4475235641002655,
      "learning_rate": 2.247555422737924e-06,
      "loss": 0.372,
      "num_input_tokens_seen": 21910624,
      "step": 33490
    },
    {
      "epoch": 17.55503144654088,
      "grad_norm": 0.6824585795402527,
      "learning_rate": 2.2428195086036024e-06,
      "loss": 0.3857,
      "num_input_tokens_seen": 21914816,
      "step": 33495
    },
    {
      "epoch": 17.557651991614257,
      "grad_norm": 0.41296711564064026,
      "learning_rate": 2.238088355011811e-06,
      "loss": 0.3455,
      "num_input_tokens_seen": 21919456,
      "step": 33500
    },
    {
      "epoch": 17.56027253668763,
      "grad_norm": 0.35511624813079834,
      "learning_rate": 2.2333619629522518e-06,
      "loss": 0.4986,
      "num_input_tokens_seen": 21922880,
      "step": 33505
    },
    {
      "epoch": 17.562893081761008,
      "grad_norm": 0.4005453884601593,
      "learning_rate": 2.2286403334136424e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 21926240,
      "step": 33510
    },
    {
      "epoch": 17.56551362683438,
      "grad_norm": 1.4441721439361572,
      "learning_rate": 2.223923467383679e-06,
      "loss": 0.4398,
      "num_input_tokens_seen": 21929024,
      "step": 33515
    },
    {
      "epoch": 17.568134171907758,
      "grad_norm": 0.5363181233406067,
      "learning_rate": 2.2192113658490886e-06,
      "loss": 0.3503,
      "num_input_tokens_seen": 21931904,
      "step": 33520
    },
    {
      "epoch": 17.57075471698113,
      "grad_norm": 0.6127523183822632,
      "learning_rate": 2.2145040297955776e-06,
      "loss": 0.6673,
      "num_input_tokens_seen": 21934784,
      "step": 33525
    },
    {
      "epoch": 17.57337526205451,
      "grad_norm": 0.5403136610984802,
      "learning_rate": 2.209801460207875e-06,
      "loss": 0.3718,
      "num_input_tokens_seen": 21937376,
      "step": 33530
    },
    {
      "epoch": 17.57599580712788,
      "grad_norm": 0.35742971301078796,
      "learning_rate": 2.2051036580697042e-06,
      "loss": 0.4789,
      "num_input_tokens_seen": 21941568,
      "step": 33535
    },
    {
      "epoch": 17.57861635220126,
      "grad_norm": 0.40500912070274353,
      "learning_rate": 2.2004106243637885e-06,
      "loss": 0.3652,
      "num_input_tokens_seen": 21944800,
      "step": 33540
    },
    {
      "epoch": 17.581236897274632,
      "grad_norm": 0.8263679146766663,
      "learning_rate": 2.1957223600718587e-06,
      "loss": 0.4533,
      "num_input_tokens_seen": 21947680,
      "step": 33545
    },
    {
      "epoch": 17.58385744234801,
      "grad_norm": 0.4009869396686554,
      "learning_rate": 2.1910388661746493e-06,
      "loss": 0.4825,
      "num_input_tokens_seen": 21951392,
      "step": 33550
    },
    {
      "epoch": 17.586477987421382,
      "grad_norm": 0.43840494751930237,
      "learning_rate": 2.186360143651886e-06,
      "loss": 0.5197,
      "num_input_tokens_seen": 21954464,
      "step": 33555
    },
    {
      "epoch": 17.58909853249476,
      "grad_norm": 0.9966752529144287,
      "learning_rate": 2.1816861934823103e-06,
      "loss": 0.4061,
      "num_input_tokens_seen": 21957504,
      "step": 33560
    },
    {
      "epoch": 17.591719077568133,
      "grad_norm": 0.42262402176856995,
      "learning_rate": 2.1770170166436595e-06,
      "loss": 0.4319,
      "num_input_tokens_seen": 21960320,
      "step": 33565
    },
    {
      "epoch": 17.59433962264151,
      "grad_norm": 0.36486735939979553,
      "learning_rate": 2.172352614112669e-06,
      "loss": 0.4847,
      "num_input_tokens_seen": 21963648,
      "step": 33570
    },
    {
      "epoch": 17.596960167714883,
      "grad_norm": 0.444052129983902,
      "learning_rate": 2.167692986865083e-06,
      "loss": 0.463,
      "num_input_tokens_seen": 21967264,
      "step": 33575
    },
    {
      "epoch": 17.59958071278826,
      "grad_norm": 0.520409107208252,
      "learning_rate": 2.1630381358756362e-06,
      "loss": 0.4484,
      "num_input_tokens_seen": 21970464,
      "step": 33580
    },
    {
      "epoch": 17.602201257861637,
      "grad_norm": 0.5675212740898132,
      "learning_rate": 2.158388062118072e-06,
      "loss": 0.4443,
      "num_input_tokens_seen": 21973536,
      "step": 33585
    },
    {
      "epoch": 17.60482180293501,
      "grad_norm": 0.35870733857154846,
      "learning_rate": 2.1537427665651393e-06,
      "loss": 0.4125,
      "num_input_tokens_seen": 21976864,
      "step": 33590
    },
    {
      "epoch": 17.607442348008387,
      "grad_norm": 0.6480590105056763,
      "learning_rate": 2.1491022501885726e-06,
      "loss": 0.4716,
      "num_input_tokens_seen": 21980512,
      "step": 33595
    },
    {
      "epoch": 17.61006289308176,
      "grad_norm": 0.7211045622825623,
      "learning_rate": 2.1444665139591174e-06,
      "loss": 0.5108,
      "num_input_tokens_seen": 21983744,
      "step": 33600
    },
    {
      "epoch": 17.612683438155138,
      "grad_norm": 0.4154103100299835,
      "learning_rate": 2.139835558846523e-06,
      "loss": 0.4102,
      "num_input_tokens_seen": 21986336,
      "step": 33605
    },
    {
      "epoch": 17.61530398322851,
      "grad_norm": 0.3560143709182739,
      "learning_rate": 2.1352093858195225e-06,
      "loss": 0.3482,
      "num_input_tokens_seen": 21989280,
      "step": 33610
    },
    {
      "epoch": 17.617924528301888,
      "grad_norm": 0.48950424790382385,
      "learning_rate": 2.1305879958458676e-06,
      "loss": 0.4177,
      "num_input_tokens_seen": 21993312,
      "step": 33615
    },
    {
      "epoch": 17.62054507337526,
      "grad_norm": 0.4879930913448334,
      "learning_rate": 2.1259713898922894e-06,
      "loss": 0.4806,
      "num_input_tokens_seen": 21996160,
      "step": 33620
    },
    {
      "epoch": 17.62316561844864,
      "grad_norm": 0.3480820655822754,
      "learning_rate": 2.1213595689245386e-06,
      "loss": 0.3392,
      "num_input_tokens_seen": 21999264,
      "step": 33625
    },
    {
      "epoch": 17.62578616352201,
      "grad_norm": 0.43836531043052673,
      "learning_rate": 2.1167525339073513e-06,
      "loss": 0.4759,
      "num_input_tokens_seen": 22001984,
      "step": 33630
    },
    {
      "epoch": 17.62840670859539,
      "grad_norm": 0.4161217212677002,
      "learning_rate": 2.112150285804473e-06,
      "loss": 0.4221,
      "num_input_tokens_seen": 22007296,
      "step": 33635
    },
    {
      "epoch": 17.631027253668762,
      "grad_norm": 0.7542493343353271,
      "learning_rate": 2.107552825578632e-06,
      "loss": 0.3343,
      "num_input_tokens_seen": 22010272,
      "step": 33640
    },
    {
      "epoch": 17.63364779874214,
      "grad_norm": 0.47942081093788147,
      "learning_rate": 2.102960154191572e-06,
      "loss": 0.4801,
      "num_input_tokens_seen": 22013984,
      "step": 33645
    },
    {
      "epoch": 17.636268343815512,
      "grad_norm": 0.4097810983657837,
      "learning_rate": 2.098372272604021e-06,
      "loss": 0.4086,
      "num_input_tokens_seen": 22017216,
      "step": 33650
    },
    {
      "epoch": 17.63888888888889,
      "grad_norm": 0.4911932349205017,
      "learning_rate": 2.0937891817757155e-06,
      "loss": 0.4733,
      "num_input_tokens_seen": 22019872,
      "step": 33655
    },
    {
      "epoch": 17.641509433962263,
      "grad_norm": 0.5144511461257935,
      "learning_rate": 2.089210882665385e-06,
      "loss": 0.4011,
      "num_input_tokens_seen": 22022432,
      "step": 33660
    },
    {
      "epoch": 17.64412997903564,
      "grad_norm": 1.1089540719985962,
      "learning_rate": 2.0846373762307635e-06,
      "loss": 0.5014,
      "num_input_tokens_seen": 22025120,
      "step": 33665
    },
    {
      "epoch": 17.646750524109013,
      "grad_norm": 0.4620349705219269,
      "learning_rate": 2.080068663428569e-06,
      "loss": 0.4324,
      "num_input_tokens_seen": 22028896,
      "step": 33670
    },
    {
      "epoch": 17.64937106918239,
      "grad_norm": 1.593665599822998,
      "learning_rate": 2.075504745214521e-06,
      "loss": 0.4468,
      "num_input_tokens_seen": 22031168,
      "step": 33675
    },
    {
      "epoch": 17.651991614255767,
      "grad_norm": 0.49994200468063354,
      "learning_rate": 2.0709456225433437e-06,
      "loss": 0.4013,
      "num_input_tokens_seen": 22034400,
      "step": 33680
    },
    {
      "epoch": 17.65461215932914,
      "grad_norm": 0.311405211687088,
      "learning_rate": 2.0663912963687554e-06,
      "loss": 0.3379,
      "num_input_tokens_seen": 22037984,
      "step": 33685
    },
    {
      "epoch": 17.657232704402517,
      "grad_norm": 0.4041183888912201,
      "learning_rate": 2.0618417676434713e-06,
      "loss": 0.4271,
      "num_input_tokens_seen": 22040992,
      "step": 33690
    },
    {
      "epoch": 17.65985324947589,
      "grad_norm": 0.6496570706367493,
      "learning_rate": 2.0572970373191975e-06,
      "loss": 0.4569,
      "num_input_tokens_seen": 22043616,
      "step": 33695
    },
    {
      "epoch": 17.662473794549268,
      "grad_norm": 0.35187894105911255,
      "learning_rate": 2.052757106346634e-06,
      "loss": 0.3617,
      "num_input_tokens_seen": 22045920,
      "step": 33700
    },
    {
      "epoch": 17.66509433962264,
      "grad_norm": 0.421191930770874,
      "learning_rate": 2.048221975675488e-06,
      "loss": 0.6023,
      "num_input_tokens_seen": 22050144,
      "step": 33705
    },
    {
      "epoch": 17.667714884696018,
      "grad_norm": 0.45861154794692993,
      "learning_rate": 2.043691646254456e-06,
      "loss": 0.3452,
      "num_input_tokens_seen": 22054432,
      "step": 33710
    },
    {
      "epoch": 17.67033542976939,
      "grad_norm": 0.4381851851940155,
      "learning_rate": 2.039166119031233e-06,
      "loss": 0.4767,
      "num_input_tokens_seen": 22058272,
      "step": 33715
    },
    {
      "epoch": 17.67295597484277,
      "grad_norm": 0.38017934560775757,
      "learning_rate": 2.034645394952511e-06,
      "loss": 0.477,
      "num_input_tokens_seen": 22061088,
      "step": 33720
    },
    {
      "epoch": 17.67557651991614,
      "grad_norm": 0.7228896021842957,
      "learning_rate": 2.030129474963971e-06,
      "loss": 0.3584,
      "num_input_tokens_seen": 22064160,
      "step": 33725
    },
    {
      "epoch": 17.67819706498952,
      "grad_norm": 0.5019980669021606,
      "learning_rate": 2.025618360010284e-06,
      "loss": 0.4703,
      "num_input_tokens_seen": 22066720,
      "step": 33730
    },
    {
      "epoch": 17.680817610062892,
      "grad_norm": 0.30690330266952515,
      "learning_rate": 2.0211120510351317e-06,
      "loss": 0.5272,
      "num_input_tokens_seen": 22069952,
      "step": 33735
    },
    {
      "epoch": 17.68343815513627,
      "grad_norm": 0.733264148235321,
      "learning_rate": 2.0166105489811822e-06,
      "loss": 0.4012,
      "num_input_tokens_seen": 22074144,
      "step": 33740
    },
    {
      "epoch": 17.686058700209642,
      "grad_norm": 0.33281752467155457,
      "learning_rate": 2.0121138547901016e-06,
      "loss": 0.3585,
      "num_input_tokens_seen": 22077632,
      "step": 33745
    },
    {
      "epoch": 17.68867924528302,
      "grad_norm": 0.9637420177459717,
      "learning_rate": 2.0076219694025427e-06,
      "loss": 0.3787,
      "num_input_tokens_seen": 22081536,
      "step": 33750
    },
    {
      "epoch": 17.691299790356393,
      "grad_norm": 0.38096359372138977,
      "learning_rate": 2.003134893758152e-06,
      "loss": 0.5023,
      "num_input_tokens_seen": 22085600,
      "step": 33755
    },
    {
      "epoch": 17.69392033542977,
      "grad_norm": 0.6959258317947388,
      "learning_rate": 1.9986526287955805e-06,
      "loss": 0.4936,
      "num_input_tokens_seen": 22089088,
      "step": 33760
    },
    {
      "epoch": 17.696540880503143,
      "grad_norm": 0.4421130120754242,
      "learning_rate": 1.9941751754524644e-06,
      "loss": 0.3763,
      "num_input_tokens_seen": 22091584,
      "step": 33765
    },
    {
      "epoch": 17.69916142557652,
      "grad_norm": 0.6158466339111328,
      "learning_rate": 1.9897025346654412e-06,
      "loss": 0.3112,
      "num_input_tokens_seen": 22094496,
      "step": 33770
    },
    {
      "epoch": 17.701781970649897,
      "grad_norm": 0.394535094499588,
      "learning_rate": 1.985234707370134e-06,
      "loss": 0.3964,
      "num_input_tokens_seen": 22097728,
      "step": 33775
    },
    {
      "epoch": 17.70440251572327,
      "grad_norm": 0.5153741240501404,
      "learning_rate": 1.9807716945011544e-06,
      "loss": 0.5904,
      "num_input_tokens_seen": 22101696,
      "step": 33780
    },
    {
      "epoch": 17.707023060796647,
      "grad_norm": 0.6756526827812195,
      "learning_rate": 1.9763134969921215e-06,
      "loss": 0.5097,
      "num_input_tokens_seen": 22104256,
      "step": 33785
    },
    {
      "epoch": 17.70964360587002,
      "grad_norm": 0.5394156575202942,
      "learning_rate": 1.971860115775634e-06,
      "loss": 0.4334,
      "num_input_tokens_seen": 22107968,
      "step": 33790
    },
    {
      "epoch": 17.712264150943398,
      "grad_norm": 0.6161132454872131,
      "learning_rate": 1.967411551783294e-06,
      "loss": 0.4658,
      "num_input_tokens_seen": 22110464,
      "step": 33795
    },
    {
      "epoch": 17.71488469601677,
      "grad_norm": 0.5325239896774292,
      "learning_rate": 1.9629678059456895e-06,
      "loss": 0.4092,
      "num_input_tokens_seen": 22113344,
      "step": 33800
    },
    {
      "epoch": 17.717505241090148,
      "grad_norm": 0.46526896953582764,
      "learning_rate": 1.9585288791924004e-06,
      "loss": 0.3396,
      "num_input_tokens_seen": 22116640,
      "step": 33805
    },
    {
      "epoch": 17.72012578616352,
      "grad_norm": 0.5091137886047363,
      "learning_rate": 1.9540947724519947e-06,
      "loss": 0.3857,
      "num_input_tokens_seen": 22119808,
      "step": 33810
    },
    {
      "epoch": 17.7227463312369,
      "grad_norm": 0.6472842693328857,
      "learning_rate": 1.9496654866520414e-06,
      "loss": 0.5349,
      "num_input_tokens_seen": 22123328,
      "step": 33815
    },
    {
      "epoch": 17.72536687631027,
      "grad_norm": 0.5460161566734314,
      "learning_rate": 1.9452410227190937e-06,
      "loss": 0.4079,
      "num_input_tokens_seen": 22127648,
      "step": 33820
    },
    {
      "epoch": 17.72798742138365,
      "grad_norm": 0.4086003601551056,
      "learning_rate": 1.9408213815787087e-06,
      "loss": 0.4242,
      "num_input_tokens_seen": 22130816,
      "step": 33825
    },
    {
      "epoch": 17.730607966457022,
      "grad_norm": 0.4905826449394226,
      "learning_rate": 1.936406564155413e-06,
      "loss": 0.3698,
      "num_input_tokens_seen": 22133568,
      "step": 33830
    },
    {
      "epoch": 17.7332285115304,
      "grad_norm": 0.43264028429985046,
      "learning_rate": 1.9319965713727435e-06,
      "loss": 0.4625,
      "num_input_tokens_seen": 22137888,
      "step": 33835
    },
    {
      "epoch": 17.735849056603772,
      "grad_norm": 0.3714067041873932,
      "learning_rate": 1.9275914041532144e-06,
      "loss": 0.3403,
      "num_input_tokens_seen": 22140448,
      "step": 33840
    },
    {
      "epoch": 17.73846960167715,
      "grad_norm": 0.7030137777328491,
      "learning_rate": 1.9231910634183394e-06,
      "loss": 0.4978,
      "num_input_tokens_seen": 22142400,
      "step": 33845
    },
    {
      "epoch": 17.741090146750523,
      "grad_norm": 0.59859699010849,
      "learning_rate": 1.918795550088623e-06,
      "loss": 0.4805,
      "num_input_tokens_seen": 22145824,
      "step": 33850
    },
    {
      "epoch": 17.7437106918239,
      "grad_norm": 0.39346060156822205,
      "learning_rate": 1.9144048650835496e-06,
      "loss": 0.5204,
      "num_input_tokens_seen": 22149536,
      "step": 33855
    },
    {
      "epoch": 17.746331236897273,
      "grad_norm": 0.689742386341095,
      "learning_rate": 1.9100190093216093e-06,
      "loss": 0.479,
      "num_input_tokens_seen": 22152384,
      "step": 33860
    },
    {
      "epoch": 17.74895178197065,
      "grad_norm": 0.42709869146347046,
      "learning_rate": 1.9056379837202627e-06,
      "loss": 0.4737,
      "num_input_tokens_seen": 22156480,
      "step": 33865
    },
    {
      "epoch": 17.751572327044027,
      "grad_norm": 0.3020852208137512,
      "learning_rate": 1.9012617891959766e-06,
      "loss": 0.3456,
      "num_input_tokens_seen": 22160928,
      "step": 33870
    },
    {
      "epoch": 17.7541928721174,
      "grad_norm": 0.5114675164222717,
      "learning_rate": 1.8968904266641996e-06,
      "loss": 0.4813,
      "num_input_tokens_seen": 22164576,
      "step": 33875
    },
    {
      "epoch": 17.756813417190777,
      "grad_norm": 0.26917290687561035,
      "learning_rate": 1.892523897039372e-06,
      "loss": 0.3997,
      "num_input_tokens_seen": 22168704,
      "step": 33880
    },
    {
      "epoch": 17.75943396226415,
      "grad_norm": 0.41975918412208557,
      "learning_rate": 1.888162201234922e-06,
      "loss": 0.4566,
      "num_input_tokens_seen": 22172896,
      "step": 33885
    },
    {
      "epoch": 17.762054507337528,
      "grad_norm": 0.7012583017349243,
      "learning_rate": 1.8838053401632666e-06,
      "loss": 0.4728,
      "num_input_tokens_seen": 22175552,
      "step": 33890
    },
    {
      "epoch": 17.7646750524109,
      "grad_norm": 0.4812372028827667,
      "learning_rate": 1.8794533147358074e-06,
      "loss": 0.4161,
      "num_input_tokens_seen": 22179200,
      "step": 33895
    },
    {
      "epoch": 17.767295597484278,
      "grad_norm": 0.29841262102127075,
      "learning_rate": 1.875106125862941e-06,
      "loss": 0.4294,
      "num_input_tokens_seen": 22183456,
      "step": 33900
    },
    {
      "epoch": 17.76991614255765,
      "grad_norm": 0.4537148177623749,
      "learning_rate": 1.870763774454057e-06,
      "loss": 0.5194,
      "num_input_tokens_seen": 22187584,
      "step": 33905
    },
    {
      "epoch": 17.77253668763103,
      "grad_norm": 0.4271456301212311,
      "learning_rate": 1.8664262614175115e-06,
      "loss": 0.4212,
      "num_input_tokens_seen": 22191296,
      "step": 33910
    },
    {
      "epoch": 17.7751572327044,
      "grad_norm": 0.4359722435474396,
      "learning_rate": 1.862093587660671e-06,
      "loss": 0.5071,
      "num_input_tokens_seen": 22195552,
      "step": 33915
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.39648279547691345,
      "learning_rate": 1.8577657540898851e-06,
      "loss": 0.412,
      "num_input_tokens_seen": 22199296,
      "step": 33920
    },
    {
      "epoch": 17.780398322851152,
      "grad_norm": 0.504641056060791,
      "learning_rate": 1.8534427616104798e-06,
      "loss": 0.4619,
      "num_input_tokens_seen": 22202144,
      "step": 33925
    },
    {
      "epoch": 17.78301886792453,
      "grad_norm": 0.4783768951892853,
      "learning_rate": 1.8491246111267846e-06,
      "loss": 0.5155,
      "num_input_tokens_seen": 22205440,
      "step": 33930
    },
    {
      "epoch": 17.785639412997902,
      "grad_norm": 0.5295785069465637,
      "learning_rate": 1.8448113035420934e-06,
      "loss": 0.4466,
      "num_input_tokens_seen": 22209120,
      "step": 33935
    },
    {
      "epoch": 17.78825995807128,
      "grad_norm": 0.39061516523361206,
      "learning_rate": 1.840502839758712e-06,
      "loss": 0.3697,
      "num_input_tokens_seen": 22213120,
      "step": 33940
    },
    {
      "epoch": 17.790880503144653,
      "grad_norm": 0.5111088156700134,
      "learning_rate": 1.8361992206779226e-06,
      "loss": 0.3971,
      "num_input_tokens_seen": 22216768,
      "step": 33945
    },
    {
      "epoch": 17.79350104821803,
      "grad_norm": 0.44417113065719604,
      "learning_rate": 1.8319004471999856e-06,
      "loss": 0.3194,
      "num_input_tokens_seen": 22219968,
      "step": 33950
    },
    {
      "epoch": 17.796121593291403,
      "grad_norm": 0.6598866581916809,
      "learning_rate": 1.8276065202241592e-06,
      "loss": 0.4369,
      "num_input_tokens_seen": 22222368,
      "step": 33955
    },
    {
      "epoch": 17.79874213836478,
      "grad_norm": 0.33406466245651245,
      "learning_rate": 1.8233174406486863e-06,
      "loss": 0.4061,
      "num_input_tokens_seen": 22226112,
      "step": 33960
    },
    {
      "epoch": 17.801362683438157,
      "grad_norm": 0.3076362609863281,
      "learning_rate": 1.8190332093707907e-06,
      "loss": 0.4284,
      "num_input_tokens_seen": 22229728,
      "step": 33965
    },
    {
      "epoch": 17.80398322851153,
      "grad_norm": 0.34489890933036804,
      "learning_rate": 1.8147538272866833e-06,
      "loss": 0.4214,
      "num_input_tokens_seen": 22232928,
      "step": 33970
    },
    {
      "epoch": 17.806603773584907,
      "grad_norm": 0.5181482434272766,
      "learning_rate": 1.8104792952915672e-06,
      "loss": 0.4673,
      "num_input_tokens_seen": 22235776,
      "step": 33975
    },
    {
      "epoch": 17.80922431865828,
      "grad_norm": 0.40745317935943604,
      "learning_rate": 1.8062096142796192e-06,
      "loss": 0.3767,
      "num_input_tokens_seen": 22239136,
      "step": 33980
    },
    {
      "epoch": 17.811844863731658,
      "grad_norm": 0.75086510181427,
      "learning_rate": 1.8019447851440163e-06,
      "loss": 0.3654,
      "num_input_tokens_seen": 22243872,
      "step": 33985
    },
    {
      "epoch": 17.81446540880503,
      "grad_norm": 0.8892475366592407,
      "learning_rate": 1.7976848087769032e-06,
      "loss": 0.5053,
      "num_input_tokens_seen": 22247168,
      "step": 33990
    },
    {
      "epoch": 17.817085953878408,
      "grad_norm": 0.4069085717201233,
      "learning_rate": 1.79342968606942e-06,
      "loss": 0.552,
      "num_input_tokens_seen": 22251648,
      "step": 33995
    },
    {
      "epoch": 17.81970649895178,
      "grad_norm": 0.598483145236969,
      "learning_rate": 1.7891794179116933e-06,
      "loss": 0.4956,
      "num_input_tokens_seen": 22254240,
      "step": 34000
    },
    {
      "epoch": 17.822327044025158,
      "grad_norm": 0.3543824851512909,
      "learning_rate": 1.7849340051928316e-06,
      "loss": 0.5486,
      "num_input_tokens_seen": 22258176,
      "step": 34005
    },
    {
      "epoch": 17.82494758909853,
      "grad_norm": 0.5074234008789062,
      "learning_rate": 1.780693448800927e-06,
      "loss": 0.4385,
      "num_input_tokens_seen": 22261120,
      "step": 34010
    },
    {
      "epoch": 17.82756813417191,
      "grad_norm": 0.4885628819465637,
      "learning_rate": 1.7764577496230478e-06,
      "loss": 0.4466,
      "num_input_tokens_seen": 22264480,
      "step": 34015
    },
    {
      "epoch": 17.830188679245282,
      "grad_norm": 0.43687963485717773,
      "learning_rate": 1.7722269085452602e-06,
      "loss": 0.3318,
      "num_input_tokens_seen": 22267392,
      "step": 34020
    },
    {
      "epoch": 17.83280922431866,
      "grad_norm": 0.5888258814811707,
      "learning_rate": 1.7680009264526038e-06,
      "loss": 0.5041,
      "num_input_tokens_seen": 22270336,
      "step": 34025
    },
    {
      "epoch": 17.835429769392032,
      "grad_norm": 0.5514424443244934,
      "learning_rate": 1.7637798042291126e-06,
      "loss": 0.5031,
      "num_input_tokens_seen": 22273472,
      "step": 34030
    },
    {
      "epoch": 17.83805031446541,
      "grad_norm": 0.23883044719696045,
      "learning_rate": 1.7595635427577944e-06,
      "loss": 0.2994,
      "num_input_tokens_seen": 22276320,
      "step": 34035
    },
    {
      "epoch": 17.840670859538783,
      "grad_norm": 0.3669430613517761,
      "learning_rate": 1.7553521429206465e-06,
      "loss": 0.3755,
      "num_input_tokens_seen": 22280224,
      "step": 34040
    },
    {
      "epoch": 17.84329140461216,
      "grad_norm": 0.5179800987243652,
      "learning_rate": 1.7511456055986335e-06,
      "loss": 0.4529,
      "num_input_tokens_seen": 22283584,
      "step": 34045
    },
    {
      "epoch": 17.845911949685533,
      "grad_norm": 0.5873843431472778,
      "learning_rate": 1.746943931671724e-06,
      "loss": 0.3882,
      "num_input_tokens_seen": 22285952,
      "step": 34050
    },
    {
      "epoch": 17.84853249475891,
      "grad_norm": 0.651024341583252,
      "learning_rate": 1.742747122018859e-06,
      "loss": 0.4189,
      "num_input_tokens_seen": 22288608,
      "step": 34055
    },
    {
      "epoch": 17.851153039832283,
      "grad_norm": 0.5742501020431519,
      "learning_rate": 1.738555177517967e-06,
      "loss": 0.3726,
      "num_input_tokens_seen": 22292416,
      "step": 34060
    },
    {
      "epoch": 17.85377358490566,
      "grad_norm": 0.4558820426464081,
      "learning_rate": 1.734368099045952e-06,
      "loss": 0.3994,
      "num_input_tokens_seen": 22295776,
      "step": 34065
    },
    {
      "epoch": 17.856394129979037,
      "grad_norm": 0.5663416981697083,
      "learning_rate": 1.730185887478697e-06,
      "loss": 0.5602,
      "num_input_tokens_seen": 22298272,
      "step": 34070
    },
    {
      "epoch": 17.85901467505241,
      "grad_norm": 0.6937442421913147,
      "learning_rate": 1.7260085436910794e-06,
      "loss": 0.3203,
      "num_input_tokens_seen": 22300960,
      "step": 34075
    },
    {
      "epoch": 17.861635220125788,
      "grad_norm": 0.670898973941803,
      "learning_rate": 1.7218360685569506e-06,
      "loss": 0.3985,
      "num_input_tokens_seen": 22303904,
      "step": 34080
    },
    {
      "epoch": 17.86425576519916,
      "grad_norm": 0.4646751582622528,
      "learning_rate": 1.7176684629491513e-06,
      "loss": 0.3572,
      "num_input_tokens_seen": 22307584,
      "step": 34085
    },
    {
      "epoch": 17.866876310272538,
      "grad_norm": 0.39819827675819397,
      "learning_rate": 1.7135057277394868e-06,
      "loss": 0.412,
      "num_input_tokens_seen": 22311008,
      "step": 34090
    },
    {
      "epoch": 17.86949685534591,
      "grad_norm": 0.3163166046142578,
      "learning_rate": 1.7093478637987548e-06,
      "loss": 0.3528,
      "num_input_tokens_seen": 22313888,
      "step": 34095
    },
    {
      "epoch": 17.872117400419288,
      "grad_norm": 0.3193244934082031,
      "learning_rate": 1.7051948719967375e-06,
      "loss": 0.4217,
      "num_input_tokens_seen": 22317376,
      "step": 34100
    },
    {
      "epoch": 17.87473794549266,
      "grad_norm": 0.49394601583480835,
      "learning_rate": 1.7010467532021901e-06,
      "loss": 0.5297,
      "num_input_tokens_seen": 22320320,
      "step": 34105
    },
    {
      "epoch": 17.87735849056604,
      "grad_norm": 0.5105839371681213,
      "learning_rate": 1.6969035082828544e-06,
      "loss": 0.4455,
      "num_input_tokens_seen": 22323136,
      "step": 34110
    },
    {
      "epoch": 17.879979035639412,
      "grad_norm": 0.3432508111000061,
      "learning_rate": 1.692765138105451e-06,
      "loss": 0.4366,
      "num_input_tokens_seen": 22326848,
      "step": 34115
    },
    {
      "epoch": 17.88259958071279,
      "grad_norm": 0.4226905405521393,
      "learning_rate": 1.6886316435356792e-06,
      "loss": 0.354,
      "num_input_tokens_seen": 22329856,
      "step": 34120
    },
    {
      "epoch": 17.885220125786162,
      "grad_norm": 1.3422647714614868,
      "learning_rate": 1.6845030254382138e-06,
      "loss": 0.579,
      "num_input_tokens_seen": 22332032,
      "step": 34125
    },
    {
      "epoch": 17.88784067085954,
      "grad_norm": 0.4527592360973358,
      "learning_rate": 1.6803792846767197e-06,
      "loss": 0.5394,
      "num_input_tokens_seen": 22334656,
      "step": 34130
    },
    {
      "epoch": 17.890461215932913,
      "grad_norm": 0.2951538860797882,
      "learning_rate": 1.6762604221138316e-06,
      "loss": 0.4098,
      "num_input_tokens_seen": 22337632,
      "step": 34135
    },
    {
      "epoch": 17.89308176100629,
      "grad_norm": 0.624562680721283,
      "learning_rate": 1.6721464386111774e-06,
      "loss": 0.5567,
      "num_input_tokens_seen": 22340288,
      "step": 34140
    },
    {
      "epoch": 17.895702306079663,
      "grad_norm": 0.40480273962020874,
      "learning_rate": 1.6680373350293516e-06,
      "loss": 0.4114,
      "num_input_tokens_seen": 22342848,
      "step": 34145
    },
    {
      "epoch": 17.89832285115304,
      "grad_norm": 0.5152974724769592,
      "learning_rate": 1.6639331122279223e-06,
      "loss": 0.5121,
      "num_input_tokens_seen": 22346400,
      "step": 34150
    },
    {
      "epoch": 17.900943396226417,
      "grad_norm": 0.42015376687049866,
      "learning_rate": 1.6598337710654583e-06,
      "loss": 0.387,
      "num_input_tokens_seen": 22350592,
      "step": 34155
    },
    {
      "epoch": 17.90356394129979,
      "grad_norm": 0.2852230370044708,
      "learning_rate": 1.6557393123994874e-06,
      "loss": 0.3723,
      "num_input_tokens_seen": 22354560,
      "step": 34160
    },
    {
      "epoch": 17.906184486373167,
      "grad_norm": 0.43330878019332886,
      "learning_rate": 1.651649737086533e-06,
      "loss": 0.4655,
      "num_input_tokens_seen": 22357888,
      "step": 34165
    },
    {
      "epoch": 17.90880503144654,
      "grad_norm": 0.9999493956565857,
      "learning_rate": 1.6475650459820774e-06,
      "loss": 0.4526,
      "num_input_tokens_seen": 22360224,
      "step": 34170
    },
    {
      "epoch": 17.911425576519918,
      "grad_norm": 0.44895368814468384,
      "learning_rate": 1.643485239940598e-06,
      "loss": 0.4377,
      "num_input_tokens_seen": 22364480,
      "step": 34175
    },
    {
      "epoch": 17.91404612159329,
      "grad_norm": 1.5591427087783813,
      "learning_rate": 1.6394103198155375e-06,
      "loss": 0.417,
      "num_input_tokens_seen": 22367808,
      "step": 34180
    },
    {
      "epoch": 17.916666666666668,
      "grad_norm": 0.4646463692188263,
      "learning_rate": 1.6353402864593276e-06,
      "loss": 0.3921,
      "num_input_tokens_seen": 22371328,
      "step": 34185
    },
    {
      "epoch": 17.91928721174004,
      "grad_norm": 0.285546213388443,
      "learning_rate": 1.6312751407233706e-06,
      "loss": 0.4801,
      "num_input_tokens_seen": 22374080,
      "step": 34190
    },
    {
      "epoch": 17.921907756813418,
      "grad_norm": 0.41708919405937195,
      "learning_rate": 1.6272148834580559e-06,
      "loss": 0.4833,
      "num_input_tokens_seen": 22377600,
      "step": 34195
    },
    {
      "epoch": 17.92452830188679,
      "grad_norm": 0.5371168851852417,
      "learning_rate": 1.6231595155127317e-06,
      "loss": 0.3593,
      "num_input_tokens_seen": 22380608,
      "step": 34200
    },
    {
      "epoch": 17.92714884696017,
      "grad_norm": 0.6765771508216858,
      "learning_rate": 1.6191090377357414e-06,
      "loss": 0.432,
      "num_input_tokens_seen": 22383328,
      "step": 34205
    },
    {
      "epoch": 17.929769392033542,
      "grad_norm": 0.5724443793296814,
      "learning_rate": 1.6150634509743939e-06,
      "loss": 0.3417,
      "num_input_tokens_seen": 22387520,
      "step": 34210
    },
    {
      "epoch": 17.93238993710692,
      "grad_norm": 0.393126517534256,
      "learning_rate": 1.611022756074984e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 22390368,
      "step": 34215
    },
    {
      "epoch": 17.935010482180292,
      "grad_norm": 0.6122647523880005,
      "learning_rate": 1.606986953882783e-06,
      "loss": 0.4443,
      "num_input_tokens_seen": 22392864,
      "step": 34220
    },
    {
      "epoch": 17.93763102725367,
      "grad_norm": 0.5615323781967163,
      "learning_rate": 1.6029560452420238e-06,
      "loss": 0.368,
      "num_input_tokens_seen": 22395616,
      "step": 34225
    },
    {
      "epoch": 17.940251572327043,
      "grad_norm": 0.6999538540840149,
      "learning_rate": 1.5989300309959377e-06,
      "loss": 0.3617,
      "num_input_tokens_seen": 22399040,
      "step": 34230
    },
    {
      "epoch": 17.94287211740042,
      "grad_norm": 0.45034950971603394,
      "learning_rate": 1.5949089119867116e-06,
      "loss": 0.5093,
      "num_input_tokens_seen": 22402016,
      "step": 34235
    },
    {
      "epoch": 17.945492662473793,
      "grad_norm": 0.39439985156059265,
      "learning_rate": 1.5908926890555204e-06,
      "loss": 0.4432,
      "num_input_tokens_seen": 22405440,
      "step": 34240
    },
    {
      "epoch": 17.94811320754717,
      "grad_norm": 0.9096153974533081,
      "learning_rate": 1.5868813630425167e-06,
      "loss": 0.4061,
      "num_input_tokens_seen": 22408640,
      "step": 34245
    },
    {
      "epoch": 17.950733752620543,
      "grad_norm": 0.5457698106765747,
      "learning_rate": 1.5828749347868183e-06,
      "loss": 0.4531,
      "num_input_tokens_seen": 22411808,
      "step": 34250
    },
    {
      "epoch": 17.95335429769392,
      "grad_norm": 0.5130938291549683,
      "learning_rate": 1.5788734051265268e-06,
      "loss": 0.5026,
      "num_input_tokens_seen": 22415168,
      "step": 34255
    },
    {
      "epoch": 17.955974842767297,
      "grad_norm": 0.6428270936012268,
      "learning_rate": 1.5748767748987231e-06,
      "loss": 0.3721,
      "num_input_tokens_seen": 22419264,
      "step": 34260
    },
    {
      "epoch": 17.95859538784067,
      "grad_norm": 0.6258736252784729,
      "learning_rate": 1.5708850449394436e-06,
      "loss": 0.4378,
      "num_input_tokens_seen": 22422688,
      "step": 34265
    },
    {
      "epoch": 17.961215932914047,
      "grad_norm": 0.4688395857810974,
      "learning_rate": 1.5668982160837176e-06,
      "loss": 0.4599,
      "num_input_tokens_seen": 22425600,
      "step": 34270
    },
    {
      "epoch": 17.96383647798742,
      "grad_norm": 0.7230631113052368,
      "learning_rate": 1.562916289165553e-06,
      "loss": 0.3502,
      "num_input_tokens_seen": 22430496,
      "step": 34275
    },
    {
      "epoch": 17.966457023060798,
      "grad_norm": 0.3914540708065033,
      "learning_rate": 1.5589392650179114e-06,
      "loss": 0.4147,
      "num_input_tokens_seen": 22433504,
      "step": 34280
    },
    {
      "epoch": 17.96907756813417,
      "grad_norm": 0.3556866943836212,
      "learning_rate": 1.5549671444727437e-06,
      "loss": 0.5441,
      "num_input_tokens_seen": 22436640,
      "step": 34285
    },
    {
      "epoch": 17.971698113207548,
      "grad_norm": 1.0250269174575806,
      "learning_rate": 1.55099992836098e-06,
      "loss": 0.3593,
      "num_input_tokens_seen": 22439648,
      "step": 34290
    },
    {
      "epoch": 17.97431865828092,
      "grad_norm": 0.4061185419559479,
      "learning_rate": 1.5470376175125061e-06,
      "loss": 0.4514,
      "num_input_tokens_seen": 22443360,
      "step": 34295
    },
    {
      "epoch": 17.9769392033543,
      "grad_norm": 0.7671093344688416,
      "learning_rate": 1.5430802127562038e-06,
      "loss": 0.479,
      "num_input_tokens_seen": 22445888,
      "step": 34300
    },
    {
      "epoch": 17.979559748427672,
      "grad_norm": 0.7475358247756958,
      "learning_rate": 1.5391277149199023e-06,
      "loss": 0.3732,
      "num_input_tokens_seen": 22449536,
      "step": 34305
    },
    {
      "epoch": 17.98218029350105,
      "grad_norm": 0.6981600522994995,
      "learning_rate": 1.5351801248304292e-06,
      "loss": 0.4999,
      "num_input_tokens_seen": 22453504,
      "step": 34310
    },
    {
      "epoch": 17.984800838574422,
      "grad_norm": 0.6190467476844788,
      "learning_rate": 1.5312374433135767e-06,
      "loss": 0.5998,
      "num_input_tokens_seen": 22457312,
      "step": 34315
    },
    {
      "epoch": 17.9874213836478,
      "grad_norm": 0.6242013573646545,
      "learning_rate": 1.5272996711941018e-06,
      "loss": 0.4237,
      "num_input_tokens_seen": 22459840,
      "step": 34320
    },
    {
      "epoch": 17.990041928721173,
      "grad_norm": 0.40717393159866333,
      "learning_rate": 1.5233668092957482e-06,
      "loss": 0.4236,
      "num_input_tokens_seen": 22462528,
      "step": 34325
    },
    {
      "epoch": 17.99266247379455,
      "grad_norm": 0.7452895641326904,
      "learning_rate": 1.5194388584412217e-06,
      "loss": 0.4661,
      "num_input_tokens_seen": 22465696,
      "step": 34330
    },
    {
      "epoch": 17.995283018867923,
      "grad_norm": 0.43182921409606934,
      "learning_rate": 1.5155158194522013e-06,
      "loss": 0.531,
      "num_input_tokens_seen": 22468448,
      "step": 34335
    },
    {
      "epoch": 17.9979035639413,
      "grad_norm": 1.0065010786056519,
      "learning_rate": 1.5115976931493498e-06,
      "loss": 0.4489,
      "num_input_tokens_seen": 22471168,
      "step": 34340
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.44693058729171753,
      "eval_runtime": 9.7385,
      "eval_samples_per_second": 87.077,
      "eval_steps_per_second": 21.769,
      "num_input_tokens_seen": 22472696,
      "step": 34344
    },
    {
      "epoch": 18.000524109014677,
      "grad_norm": 0.6502271294593811,
      "learning_rate": 1.5076844803522922e-06,
      "loss": 0.5118,
      "num_input_tokens_seen": 22473112,
      "step": 34345
    },
    {
      "epoch": 18.00314465408805,
      "grad_norm": 0.4003787934780121,
      "learning_rate": 1.5037761818796236e-06,
      "loss": 0.5387,
      "num_input_tokens_seen": 22475928,
      "step": 34350
    },
    {
      "epoch": 18.005765199161427,
      "grad_norm": 0.3942863941192627,
      "learning_rate": 1.4998727985489235e-06,
      "loss": 0.4264,
      "num_input_tokens_seen": 22478424,
      "step": 34355
    },
    {
      "epoch": 18.0083857442348,
      "grad_norm": 0.29683050513267517,
      "learning_rate": 1.4959743311767249e-06,
      "loss": 0.3787,
      "num_input_tokens_seen": 22481496,
      "step": 34360
    },
    {
      "epoch": 18.011006289308177,
      "grad_norm": 0.2905215620994568,
      "learning_rate": 1.4920807805785502e-06,
      "loss": 0.5309,
      "num_input_tokens_seen": 22488472,
      "step": 34365
    },
    {
      "epoch": 18.01362683438155,
      "grad_norm": 0.6801394820213318,
      "learning_rate": 1.4881921475688815e-06,
      "loss": 0.6181,
      "num_input_tokens_seen": 22491384,
      "step": 34370
    },
    {
      "epoch": 18.016247379454928,
      "grad_norm": 0.29824939370155334,
      "learning_rate": 1.4843084329611844e-06,
      "loss": 0.3399,
      "num_input_tokens_seen": 22494104,
      "step": 34375
    },
    {
      "epoch": 18.0188679245283,
      "grad_norm": 0.3262975811958313,
      "learning_rate": 1.480429637567879e-06,
      "loss": 0.5158,
      "num_input_tokens_seen": 22497656,
      "step": 34380
    },
    {
      "epoch": 18.021488469601678,
      "grad_norm": 0.601962149143219,
      "learning_rate": 1.4765557622003684e-06,
      "loss": 0.4147,
      "num_input_tokens_seen": 22500920,
      "step": 34385
    },
    {
      "epoch": 18.02410901467505,
      "grad_norm": 0.992867648601532,
      "learning_rate": 1.4726868076690186e-06,
      "loss": 0.3791,
      "num_input_tokens_seen": 22504408,
      "step": 34390
    },
    {
      "epoch": 18.02672955974843,
      "grad_norm": 0.5001320838928223,
      "learning_rate": 1.4688227747831767e-06,
      "loss": 0.4067,
      "num_input_tokens_seen": 22506904,
      "step": 34395
    },
    {
      "epoch": 18.029350104821802,
      "grad_norm": 0.5476073026657104,
      "learning_rate": 1.4649636643511573e-06,
      "loss": 0.5196,
      "num_input_tokens_seen": 22510200,
      "step": 34400
    },
    {
      "epoch": 18.03197064989518,
      "grad_norm": 0.41890034079551697,
      "learning_rate": 1.4611094771802364e-06,
      "loss": 0.3482,
      "num_input_tokens_seen": 22513112,
      "step": 34405
    },
    {
      "epoch": 18.034591194968552,
      "grad_norm": 0.583307683467865,
      "learning_rate": 1.457260214076664e-06,
      "loss": 0.4377,
      "num_input_tokens_seen": 22516824,
      "step": 34410
    },
    {
      "epoch": 18.03721174004193,
      "grad_norm": 0.9491830468177795,
      "learning_rate": 1.4534158758456623e-06,
      "loss": 0.398,
      "num_input_tokens_seen": 22520184,
      "step": 34415
    },
    {
      "epoch": 18.039832285115303,
      "grad_norm": 0.502398669719696,
      "learning_rate": 1.4495764632914271e-06,
      "loss": 0.3777,
      "num_input_tokens_seen": 22523000,
      "step": 34420
    },
    {
      "epoch": 18.04245283018868,
      "grad_norm": 0.5628058910369873,
      "learning_rate": 1.4457419772171215e-06,
      "loss": 0.3423,
      "num_input_tokens_seen": 22525400,
      "step": 34425
    },
    {
      "epoch": 18.045073375262053,
      "grad_norm": 0.7064323425292969,
      "learning_rate": 1.4419124184248733e-06,
      "loss": 0.4499,
      "num_input_tokens_seen": 22528344,
      "step": 34430
    },
    {
      "epoch": 18.04769392033543,
      "grad_norm": 0.5416465401649475,
      "learning_rate": 1.4380877877157834e-06,
      "loss": 0.458,
      "num_input_tokens_seen": 22530776,
      "step": 34435
    },
    {
      "epoch": 18.050314465408807,
      "grad_norm": 0.8749720454216003,
      "learning_rate": 1.4342680858899143e-06,
      "loss": 0.3877,
      "num_input_tokens_seen": 22534584,
      "step": 34440
    },
    {
      "epoch": 18.05293501048218,
      "grad_norm": 0.33101972937583923,
      "learning_rate": 1.4304533137463134e-06,
      "loss": 0.4328,
      "num_input_tokens_seen": 22539192,
      "step": 34445
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.28886541724205017,
      "learning_rate": 1.4266434720829808e-06,
      "loss": 0.3348,
      "num_input_tokens_seen": 22541624,
      "step": 34450
    },
    {
      "epoch": 18.05817610062893,
      "grad_norm": 0.33920782804489136,
      "learning_rate": 1.4228385616968987e-06,
      "loss": 0.4318,
      "num_input_tokens_seen": 22544888,
      "step": 34455
    },
    {
      "epoch": 18.060796645702307,
      "grad_norm": 0.5265498161315918,
      "learning_rate": 1.419038583384008e-06,
      "loss": 0.4076,
      "num_input_tokens_seen": 22547704,
      "step": 34460
    },
    {
      "epoch": 18.06341719077568,
      "grad_norm": 0.3990752100944519,
      "learning_rate": 1.4152435379392148e-06,
      "loss": 0.3685,
      "num_input_tokens_seen": 22551352,
      "step": 34465
    },
    {
      "epoch": 18.066037735849058,
      "grad_norm": 0.5405150651931763,
      "learning_rate": 1.411453426156406e-06,
      "loss": 0.4936,
      "num_input_tokens_seen": 22554520,
      "step": 34470
    },
    {
      "epoch": 18.06865828092243,
      "grad_norm": 0.4765534996986389,
      "learning_rate": 1.4076682488284282e-06,
      "loss": 0.4288,
      "num_input_tokens_seen": 22557528,
      "step": 34475
    },
    {
      "epoch": 18.071278825995808,
      "grad_norm": 0.821807861328125,
      "learning_rate": 1.4038880067470978e-06,
      "loss": 0.3678,
      "num_input_tokens_seen": 22560632,
      "step": 34480
    },
    {
      "epoch": 18.07389937106918,
      "grad_norm": 0.6600490212440491,
      "learning_rate": 1.400112700703196e-06,
      "loss": 0.4504,
      "num_input_tokens_seen": 22563896,
      "step": 34485
    },
    {
      "epoch": 18.07651991614256,
      "grad_norm": 0.5480439066886902,
      "learning_rate": 1.3963423314864805e-06,
      "loss": 0.4463,
      "num_input_tokens_seen": 22566136,
      "step": 34490
    },
    {
      "epoch": 18.079140461215932,
      "grad_norm": 0.576748251914978,
      "learning_rate": 1.3925768998856587e-06,
      "loss": 0.3429,
      "num_input_tokens_seen": 22568824,
      "step": 34495
    },
    {
      "epoch": 18.08176100628931,
      "grad_norm": 0.6114988923072815,
      "learning_rate": 1.388816406688423e-06,
      "loss": 0.5611,
      "num_input_tokens_seen": 22573304,
      "step": 34500
    },
    {
      "epoch": 18.084381551362682,
      "grad_norm": 0.5573322772979736,
      "learning_rate": 1.3850608526814246e-06,
      "loss": 0.6834,
      "num_input_tokens_seen": 22577304,
      "step": 34505
    },
    {
      "epoch": 18.08700209643606,
      "grad_norm": 0.49441424012184143,
      "learning_rate": 1.3813102386502853e-06,
      "loss": 0.5134,
      "num_input_tokens_seen": 22580568,
      "step": 34510
    },
    {
      "epoch": 18.089622641509433,
      "grad_norm": 0.4253787398338318,
      "learning_rate": 1.3775645653795881e-06,
      "loss": 0.4965,
      "num_input_tokens_seen": 22583416,
      "step": 34515
    },
    {
      "epoch": 18.09224318658281,
      "grad_norm": 0.6032217144966125,
      "learning_rate": 1.373823833652882e-06,
      "loss": 0.329,
      "num_input_tokens_seen": 22587544,
      "step": 34520
    },
    {
      "epoch": 18.094863731656183,
      "grad_norm": 0.273321270942688,
      "learning_rate": 1.3700880442526876e-06,
      "loss": 0.5345,
      "num_input_tokens_seen": 22590712,
      "step": 34525
    },
    {
      "epoch": 18.09748427672956,
      "grad_norm": 0.4583373963832855,
      "learning_rate": 1.3663571979604884e-06,
      "loss": 0.4029,
      "num_input_tokens_seen": 22593528,
      "step": 34530
    },
    {
      "epoch": 18.100104821802937,
      "grad_norm": 0.5524290800094604,
      "learning_rate": 1.3626312955567406e-06,
      "loss": 0.5333,
      "num_input_tokens_seen": 22597496,
      "step": 34535
    },
    {
      "epoch": 18.10272536687631,
      "grad_norm": 0.5276644229888916,
      "learning_rate": 1.358910337820854e-06,
      "loss": 0.49,
      "num_input_tokens_seen": 22599960,
      "step": 34540
    },
    {
      "epoch": 18.105345911949687,
      "grad_norm": 0.41979557275772095,
      "learning_rate": 1.3551943255312172e-06,
      "loss": 0.6337,
      "num_input_tokens_seen": 22602872,
      "step": 34545
    },
    {
      "epoch": 18.10796645702306,
      "grad_norm": 0.7133653163909912,
      "learning_rate": 1.3514832594651694e-06,
      "loss": 0.4911,
      "num_input_tokens_seen": 22605592,
      "step": 34550
    },
    {
      "epoch": 18.110587002096437,
      "grad_norm": 0.41723477840423584,
      "learning_rate": 1.347777140399026e-06,
      "loss": 0.5924,
      "num_input_tokens_seen": 22609112,
      "step": 34555
    },
    {
      "epoch": 18.11320754716981,
      "grad_norm": 0.393171489238739,
      "learning_rate": 1.344075969108069e-06,
      "loss": 0.4816,
      "num_input_tokens_seen": 22611736,
      "step": 34560
    },
    {
      "epoch": 18.115828092243188,
      "grad_norm": 0.4740667939186096,
      "learning_rate": 1.3403797463665352e-06,
      "loss": 0.3603,
      "num_input_tokens_seen": 22615896,
      "step": 34565
    },
    {
      "epoch": 18.11844863731656,
      "grad_norm": 0.6092923879623413,
      "learning_rate": 1.3366884729476365e-06,
      "loss": 0.3549,
      "num_input_tokens_seen": 22618296,
      "step": 34570
    },
    {
      "epoch": 18.121069182389938,
      "grad_norm": 0.5716156959533691,
      "learning_rate": 1.3330021496235467e-06,
      "loss": 0.4978,
      "num_input_tokens_seen": 22621976,
      "step": 34575
    },
    {
      "epoch": 18.12368972746331,
      "grad_norm": 0.5122687816619873,
      "learning_rate": 1.3293207771653965e-06,
      "loss": 0.3193,
      "num_input_tokens_seen": 22625720,
      "step": 34580
    },
    {
      "epoch": 18.12631027253669,
      "grad_norm": 0.40838858485221863,
      "learning_rate": 1.3256443563432891e-06,
      "loss": 0.4398,
      "num_input_tokens_seen": 22629496,
      "step": 34585
    },
    {
      "epoch": 18.128930817610062,
      "grad_norm": 0.5607420802116394,
      "learning_rate": 1.3219728879262956e-06,
      "loss": 0.5061,
      "num_input_tokens_seen": 22632920,
      "step": 34590
    },
    {
      "epoch": 18.13155136268344,
      "grad_norm": 0.34628763794898987,
      "learning_rate": 1.3183063726824375e-06,
      "loss": 0.3753,
      "num_input_tokens_seen": 22636536,
      "step": 34595
    },
    {
      "epoch": 18.134171907756812,
      "grad_norm": 0.4917004704475403,
      "learning_rate": 1.3146448113787151e-06,
      "loss": 0.6382,
      "num_input_tokens_seen": 22639800,
      "step": 34600
    },
    {
      "epoch": 18.13679245283019,
      "grad_norm": 0.48886892199516296,
      "learning_rate": 1.3109882047810773e-06,
      "loss": 0.5246,
      "num_input_tokens_seen": 22642520,
      "step": 34605
    },
    {
      "epoch": 18.139412997903563,
      "grad_norm": 0.38999637961387634,
      "learning_rate": 1.307336553654448e-06,
      "loss": 0.3765,
      "num_input_tokens_seen": 22645784,
      "step": 34610
    },
    {
      "epoch": 18.14203354297694,
      "grad_norm": 0.9508981704711914,
      "learning_rate": 1.303689858762716e-06,
      "loss": 0.3768,
      "num_input_tokens_seen": 22648664,
      "step": 34615
    },
    {
      "epoch": 18.144654088050313,
      "grad_norm": 0.511347770690918,
      "learning_rate": 1.3000481208687188e-06,
      "loss": 0.5112,
      "num_input_tokens_seen": 22651832,
      "step": 34620
    },
    {
      "epoch": 18.14727463312369,
      "grad_norm": 0.30889570713043213,
      "learning_rate": 1.2964113407342743e-06,
      "loss": 0.4592,
      "num_input_tokens_seen": 22654808,
      "step": 34625
    },
    {
      "epoch": 18.149895178197063,
      "grad_norm": 0.39110246300697327,
      "learning_rate": 1.292779519120152e-06,
      "loss": 0.4465,
      "num_input_tokens_seen": 22657624,
      "step": 34630
    },
    {
      "epoch": 18.15251572327044,
      "grad_norm": 0.4283411204814911,
      "learning_rate": 1.2891526567860856e-06,
      "loss": 0.4212,
      "num_input_tokens_seen": 22659960,
      "step": 34635
    },
    {
      "epoch": 18.155136268343817,
      "grad_norm": 0.5787113904953003,
      "learning_rate": 1.2855307544907825e-06,
      "loss": 0.4654,
      "num_input_tokens_seen": 22665080,
      "step": 34640
    },
    {
      "epoch": 18.15775681341719,
      "grad_norm": 0.6169754862785339,
      "learning_rate": 1.281913812991889e-06,
      "loss": 0.5964,
      "num_input_tokens_seen": 22667544,
      "step": 34645
    },
    {
      "epoch": 18.160377358490567,
      "grad_norm": 0.45301735401153564,
      "learning_rate": 1.2783018330460389e-06,
      "loss": 0.3906,
      "num_input_tokens_seen": 22671032,
      "step": 34650
    },
    {
      "epoch": 18.16299790356394,
      "grad_norm": 0.2635698616504669,
      "learning_rate": 1.2746948154088111e-06,
      "loss": 0.3558,
      "num_input_tokens_seen": 22675416,
      "step": 34655
    },
    {
      "epoch": 18.165618448637318,
      "grad_norm": 0.4503406882286072,
      "learning_rate": 1.2710927608347605e-06,
      "loss": 0.4692,
      "num_input_tokens_seen": 22678008,
      "step": 34660
    },
    {
      "epoch": 18.16823899371069,
      "grad_norm": 0.7437065243721008,
      "learning_rate": 1.2674956700773872e-06,
      "loss": 0.3522,
      "num_input_tokens_seen": 22680408,
      "step": 34665
    },
    {
      "epoch": 18.170859538784068,
      "grad_norm": 0.3843350112438202,
      "learning_rate": 1.2639035438891666e-06,
      "loss": 0.4688,
      "num_input_tokens_seen": 22685176,
      "step": 34670
    },
    {
      "epoch": 18.17348008385744,
      "grad_norm": 0.8133977055549622,
      "learning_rate": 1.260316383021526e-06,
      "loss": 0.2967,
      "num_input_tokens_seen": 22687960,
      "step": 34675
    },
    {
      "epoch": 18.17610062893082,
      "grad_norm": 0.4607027769088745,
      "learning_rate": 1.2567341882248645e-06,
      "loss": 0.2888,
      "num_input_tokens_seen": 22690840,
      "step": 34680
    },
    {
      "epoch": 18.178721174004192,
      "grad_norm": 0.31223684549331665,
      "learning_rate": 1.2531569602485305e-06,
      "loss": 0.3996,
      "num_input_tokens_seen": 22694424,
      "step": 34685
    },
    {
      "epoch": 18.18134171907757,
      "grad_norm": 0.35666561126708984,
      "learning_rate": 1.2495846998408445e-06,
      "loss": 0.4265,
      "num_input_tokens_seen": 22697976,
      "step": 34690
    },
    {
      "epoch": 18.183962264150942,
      "grad_norm": 0.9951964616775513,
      "learning_rate": 1.246017407749084e-06,
      "loss": 0.5046,
      "num_input_tokens_seen": 22701176,
      "step": 34695
    },
    {
      "epoch": 18.18658280922432,
      "grad_norm": 0.44355520606040955,
      "learning_rate": 1.2424550847194739e-06,
      "loss": 0.4394,
      "num_input_tokens_seen": 22703864,
      "step": 34700
    },
    {
      "epoch": 18.189203354297693,
      "grad_norm": 0.6753301024436951,
      "learning_rate": 1.2388977314972238e-06,
      "loss": 0.3485,
      "num_input_tokens_seen": 22706616,
      "step": 34705
    },
    {
      "epoch": 18.19182389937107,
      "grad_norm": 0.6424007415771484,
      "learning_rate": 1.235345348826486e-06,
      "loss": 0.3958,
      "num_input_tokens_seen": 22709272,
      "step": 34710
    },
    {
      "epoch": 18.194444444444443,
      "grad_norm": 0.3975776433944702,
      "learning_rate": 1.2317979374503825e-06,
      "loss": 0.365,
      "num_input_tokens_seen": 22712568,
      "step": 34715
    },
    {
      "epoch": 18.19706498951782,
      "grad_norm": 0.5651477575302124,
      "learning_rate": 1.228255498110989e-06,
      "loss": 0.4084,
      "num_input_tokens_seen": 22715576,
      "step": 34720
    },
    {
      "epoch": 18.199685534591197,
      "grad_norm": 0.6065927743911743,
      "learning_rate": 1.224718031549338e-06,
      "loss": 0.4953,
      "num_input_tokens_seen": 22718584,
      "step": 34725
    },
    {
      "epoch": 18.20230607966457,
      "grad_norm": 0.49469193816185,
      "learning_rate": 1.2211855385054344e-06,
      "loss": 0.3806,
      "num_input_tokens_seen": 22722456,
      "step": 34730
    },
    {
      "epoch": 18.204926624737947,
      "grad_norm": 0.261313796043396,
      "learning_rate": 1.217658019718232e-06,
      "loss": 0.4921,
      "num_input_tokens_seen": 22726392,
      "step": 34735
    },
    {
      "epoch": 18.20754716981132,
      "grad_norm": 0.3792133927345276,
      "learning_rate": 1.2141354759256484e-06,
      "loss": 0.446,
      "num_input_tokens_seen": 22729560,
      "step": 34740
    },
    {
      "epoch": 18.210167714884697,
      "grad_norm": 0.46418628096580505,
      "learning_rate": 1.2106179078645641e-06,
      "loss": 0.4449,
      "num_input_tokens_seen": 22733496,
      "step": 34745
    },
    {
      "epoch": 18.21278825995807,
      "grad_norm": 0.6833747625350952,
      "learning_rate": 1.2071053162708096e-06,
      "loss": 0.3099,
      "num_input_tokens_seen": 22736120,
      "step": 34750
    },
    {
      "epoch": 18.215408805031448,
      "grad_norm": 0.5257778167724609,
      "learning_rate": 1.203597701879175e-06,
      "loss": 0.4647,
      "num_input_tokens_seen": 22738872,
      "step": 34755
    },
    {
      "epoch": 18.21802935010482,
      "grad_norm": 0.9043521881103516,
      "learning_rate": 1.2000950654234177e-06,
      "loss": 0.5077,
      "num_input_tokens_seen": 22741336,
      "step": 34760
    },
    {
      "epoch": 18.220649895178198,
      "grad_norm": 0.4667295217514038,
      "learning_rate": 1.1965974076362491e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 22744632,
      "step": 34765
    },
    {
      "epoch": 18.22327044025157,
      "grad_norm": 0.3940984010696411,
      "learning_rate": 1.193104729249342e-06,
      "loss": 0.5372,
      "num_input_tokens_seen": 22749048,
      "step": 34770
    },
    {
      "epoch": 18.22589098532495,
      "grad_norm": 0.4815024733543396,
      "learning_rate": 1.189617030993323e-06,
      "loss": 0.4029,
      "num_input_tokens_seen": 22753048,
      "step": 34775
    },
    {
      "epoch": 18.228511530398322,
      "grad_norm": 0.559604823589325,
      "learning_rate": 1.1861343135977753e-06,
      "loss": 0.4733,
      "num_input_tokens_seen": 22756440,
      "step": 34780
    },
    {
      "epoch": 18.2311320754717,
      "grad_norm": 0.7257733941078186,
      "learning_rate": 1.182656577791244e-06,
      "loss": 0.4738,
      "num_input_tokens_seen": 22759096,
      "step": 34785
    },
    {
      "epoch": 18.233752620545072,
      "grad_norm": 0.38668325543403625,
      "learning_rate": 1.1791838243012359e-06,
      "loss": 0.3594,
      "num_input_tokens_seen": 22762424,
      "step": 34790
    },
    {
      "epoch": 18.23637316561845,
      "grad_norm": 0.5797048807144165,
      "learning_rate": 1.1757160538542117e-06,
      "loss": 0.3997,
      "num_input_tokens_seen": 22764600,
      "step": 34795
    },
    {
      "epoch": 18.238993710691823,
      "grad_norm": 0.4492945969104767,
      "learning_rate": 1.1722532671755882e-06,
      "loss": 0.4721,
      "num_input_tokens_seen": 22767672,
      "step": 34800
    },
    {
      "epoch": 18.2416142557652,
      "grad_norm": 0.3983095586299896,
      "learning_rate": 1.168795464989736e-06,
      "loss": 0.4799,
      "num_input_tokens_seen": 22770776,
      "step": 34805
    },
    {
      "epoch": 18.244234800838573,
      "grad_norm": 0.35358279943466187,
      "learning_rate": 1.165342648019993e-06,
      "loss": 0.2839,
      "num_input_tokens_seen": 22774968,
      "step": 34810
    },
    {
      "epoch": 18.24685534591195,
      "grad_norm": 0.5263785719871521,
      "learning_rate": 1.1618948169886457e-06,
      "loss": 0.449,
      "num_input_tokens_seen": 22778040,
      "step": 34815
    },
    {
      "epoch": 18.249475890985323,
      "grad_norm": 0.4838366210460663,
      "learning_rate": 1.1584519726169446e-06,
      "loss": 0.5861,
      "num_input_tokens_seen": 22780760,
      "step": 34820
    },
    {
      "epoch": 18.2520964360587,
      "grad_norm": 0.4786967933177948,
      "learning_rate": 1.1550141156250942e-06,
      "loss": 0.4031,
      "num_input_tokens_seen": 22783480,
      "step": 34825
    },
    {
      "epoch": 18.254716981132077,
      "grad_norm": 0.3274240493774414,
      "learning_rate": 1.1515812467322556e-06,
      "loss": 0.4763,
      "num_input_tokens_seen": 22787704,
      "step": 34830
    },
    {
      "epoch": 18.25733752620545,
      "grad_norm": 0.3673456609249115,
      "learning_rate": 1.1481533666565375e-06,
      "loss": 0.3693,
      "num_input_tokens_seen": 22791384,
      "step": 34835
    },
    {
      "epoch": 18.259958071278827,
      "grad_norm": 0.6431542634963989,
      "learning_rate": 1.144730476115019e-06,
      "loss": 0.4622,
      "num_input_tokens_seen": 22795000,
      "step": 34840
    },
    {
      "epoch": 18.2625786163522,
      "grad_norm": 0.4035244584083557,
      "learning_rate": 1.1413125758237304e-06,
      "loss": 0.4056,
      "num_input_tokens_seen": 22798072,
      "step": 34845
    },
    {
      "epoch": 18.265199161425578,
      "grad_norm": 0.53762286901474,
      "learning_rate": 1.1378996664976577e-06,
      "loss": 0.432,
      "num_input_tokens_seen": 22801240,
      "step": 34850
    },
    {
      "epoch": 18.26781970649895,
      "grad_norm": 0.43020305037498474,
      "learning_rate": 1.134491748850741e-06,
      "loss": 0.4295,
      "num_input_tokens_seen": 22804056,
      "step": 34855
    },
    {
      "epoch": 18.270440251572328,
      "grad_norm": 0.6845558881759644,
      "learning_rate": 1.13108882359588e-06,
      "loss": 0.4523,
      "num_input_tokens_seen": 22807480,
      "step": 34860
    },
    {
      "epoch": 18.2730607966457,
      "grad_norm": 0.4212273061275482,
      "learning_rate": 1.127690891444924e-06,
      "loss": 0.4964,
      "num_input_tokens_seen": 22810648,
      "step": 34865
    },
    {
      "epoch": 18.27568134171908,
      "grad_norm": 0.5155633091926575,
      "learning_rate": 1.1242979531086829e-06,
      "loss": 0.4444,
      "num_input_tokens_seen": 22814520,
      "step": 34870
    },
    {
      "epoch": 18.278301886792452,
      "grad_norm": 0.44849032163619995,
      "learning_rate": 1.1209100092969244e-06,
      "loss": 0.4878,
      "num_input_tokens_seen": 22818264,
      "step": 34875
    },
    {
      "epoch": 18.28092243186583,
      "grad_norm": 0.5931501388549805,
      "learning_rate": 1.1175270607183625e-06,
      "loss": 0.4227,
      "num_input_tokens_seen": 22821208,
      "step": 34880
    },
    {
      "epoch": 18.283542976939202,
      "grad_norm": 0.2998751699924469,
      "learning_rate": 1.114149108080678e-06,
      "loss": 0.4104,
      "num_input_tokens_seen": 22824696,
      "step": 34885
    },
    {
      "epoch": 18.28616352201258,
      "grad_norm": 0.3720603287220001,
      "learning_rate": 1.1107761520904925e-06,
      "loss": 0.3701,
      "num_input_tokens_seen": 22828024,
      "step": 34890
    },
    {
      "epoch": 18.288784067085953,
      "grad_norm": 0.6121291518211365,
      "learning_rate": 1.1074081934533908e-06,
      "loss": 0.4299,
      "num_input_tokens_seen": 22830456,
      "step": 34895
    },
    {
      "epoch": 18.29140461215933,
      "grad_norm": 0.39110681414604187,
      "learning_rate": 1.1040452328739153e-06,
      "loss": 0.3751,
      "num_input_tokens_seen": 22833592,
      "step": 34900
    },
    {
      "epoch": 18.294025157232703,
      "grad_norm": 0.6529070734977722,
      "learning_rate": 1.1006872710555615e-06,
      "loss": 0.4885,
      "num_input_tokens_seen": 22836696,
      "step": 34905
    },
    {
      "epoch": 18.29664570230608,
      "grad_norm": 0.6053429841995239,
      "learning_rate": 1.0973343087007703e-06,
      "loss": 0.3641,
      "num_input_tokens_seen": 22839960,
      "step": 34910
    },
    {
      "epoch": 18.299266247379457,
      "grad_norm": 0.594173014163971,
      "learning_rate": 1.0939863465109473e-06,
      "loss": 0.4526,
      "num_input_tokens_seen": 22843000,
      "step": 34915
    },
    {
      "epoch": 18.30188679245283,
      "grad_norm": 0.46012625098228455,
      "learning_rate": 1.0906433851864433e-06,
      "loss": 0.3516,
      "num_input_tokens_seen": 22847032,
      "step": 34920
    },
    {
      "epoch": 18.304507337526207,
      "grad_norm": 0.5559630393981934,
      "learning_rate": 1.0873054254265713e-06,
      "loss": 0.4094,
      "num_input_tokens_seen": 22849976,
      "step": 34925
    },
    {
      "epoch": 18.30712788259958,
      "grad_norm": 0.726600170135498,
      "learning_rate": 1.0839724679295976e-06,
      "loss": 0.4162,
      "num_input_tokens_seen": 22852888,
      "step": 34930
    },
    {
      "epoch": 18.309748427672957,
      "grad_norm": 0.4154735207557678,
      "learning_rate": 1.0806445133927311e-06,
      "loss": 0.3957,
      "num_input_tokens_seen": 22855672,
      "step": 34935
    },
    {
      "epoch": 18.31236897274633,
      "grad_norm": 0.3177565932273865,
      "learning_rate": 1.0773215625121457e-06,
      "loss": 0.4416,
      "num_input_tokens_seen": 22859032,
      "step": 34940
    },
    {
      "epoch": 18.314989517819708,
      "grad_norm": 0.5718942284584045,
      "learning_rate": 1.0740036159829715e-06,
      "loss": 0.3566,
      "num_input_tokens_seen": 22863160,
      "step": 34945
    },
    {
      "epoch": 18.31761006289308,
      "grad_norm": 0.3727361261844635,
      "learning_rate": 1.0706906744992724e-06,
      "loss": 0.4337,
      "num_input_tokens_seen": 22866104,
      "step": 34950
    },
    {
      "epoch": 18.320230607966458,
      "grad_norm": 0.6831474304199219,
      "learning_rate": 1.0673827387540857e-06,
      "loss": 0.4532,
      "num_input_tokens_seen": 22869112,
      "step": 34955
    },
    {
      "epoch": 18.32285115303983,
      "grad_norm": 0.6009002327919006,
      "learning_rate": 1.0640798094393916e-06,
      "loss": 0.3681,
      "num_input_tokens_seen": 22872184,
      "step": 34960
    },
    {
      "epoch": 18.32547169811321,
      "grad_norm": 0.41765207052230835,
      "learning_rate": 1.060781887246126e-06,
      "loss": 0.3927,
      "num_input_tokens_seen": 22876376,
      "step": 34965
    },
    {
      "epoch": 18.328092243186582,
      "grad_norm": 0.6772778630256653,
      "learning_rate": 1.057488972864179e-06,
      "loss": 0.3884,
      "num_input_tokens_seen": 22879640,
      "step": 34970
    },
    {
      "epoch": 18.33071278825996,
      "grad_norm": 0.8635320663452148,
      "learning_rate": 1.0542010669823855e-06,
      "loss": 0.5015,
      "num_input_tokens_seen": 22883832,
      "step": 34975
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.6558769941329956,
      "learning_rate": 1.0509181702885395e-06,
      "loss": 0.4023,
      "num_input_tokens_seen": 22886328,
      "step": 34980
    },
    {
      "epoch": 18.33595387840671,
      "grad_norm": 0.44488564133644104,
      "learning_rate": 1.047640283469392e-06,
      "loss": 0.4594,
      "num_input_tokens_seen": 22889080,
      "step": 34985
    },
    {
      "epoch": 18.338574423480082,
      "grad_norm": 0.39428287744522095,
      "learning_rate": 1.0443674072106308e-06,
      "loss": 0.4589,
      "num_input_tokens_seen": 22891576,
      "step": 34990
    },
    {
      "epoch": 18.34119496855346,
      "grad_norm": 0.7258082628250122,
      "learning_rate": 1.041099542196905e-06,
      "loss": 0.6458,
      "num_input_tokens_seen": 22894808,
      "step": 34995
    },
    {
      "epoch": 18.343815513626833,
      "grad_norm": 0.786459743976593,
      "learning_rate": 1.0378366891118234e-06,
      "loss": 0.4294,
      "num_input_tokens_seen": 22897912,
      "step": 35000
    },
    {
      "epoch": 18.34643605870021,
      "grad_norm": 0.7264249324798584,
      "learning_rate": 1.0345788486379265e-06,
      "loss": 0.4709,
      "num_input_tokens_seen": 22900824,
      "step": 35005
    },
    {
      "epoch": 18.349056603773583,
      "grad_norm": 0.5027327537536621,
      "learning_rate": 1.0313260214567272e-06,
      "loss": 0.4792,
      "num_input_tokens_seen": 22903864,
      "step": 35010
    },
    {
      "epoch": 18.35167714884696,
      "grad_norm": 0.6577122211456299,
      "learning_rate": 1.0280782082486729e-06,
      "loss": 0.3539,
      "num_input_tokens_seen": 22906744,
      "step": 35015
    },
    {
      "epoch": 18.354297693920337,
      "grad_norm": 0.33701589703559875,
      "learning_rate": 1.02483540969317e-06,
      "loss": 0.4429,
      "num_input_tokens_seen": 22910904,
      "step": 35020
    },
    {
      "epoch": 18.35691823899371,
      "grad_norm": 0.43883973360061646,
      "learning_rate": 1.0215976264685763e-06,
      "loss": 0.443,
      "num_input_tokens_seen": 22914072,
      "step": 35025
    },
    {
      "epoch": 18.359538784067087,
      "grad_norm": 0.3413703143596649,
      "learning_rate": 1.0183648592522021e-06,
      "loss": 0.3926,
      "num_input_tokens_seen": 22916856,
      "step": 35030
    },
    {
      "epoch": 18.36215932914046,
      "grad_norm": 0.8266654014587402,
      "learning_rate": 1.0151371087203015e-06,
      "loss": 0.3706,
      "num_input_tokens_seen": 22922008,
      "step": 35035
    },
    {
      "epoch": 18.364779874213838,
      "grad_norm": 0.4525100588798523,
      "learning_rate": 1.0119143755480815e-06,
      "loss": 0.4488,
      "num_input_tokens_seen": 22924760,
      "step": 35040
    },
    {
      "epoch": 18.36740041928721,
      "grad_norm": 0.4410887360572815,
      "learning_rate": 1.0086966604097026e-06,
      "loss": 0.4954,
      "num_input_tokens_seen": 22927416,
      "step": 35045
    },
    {
      "epoch": 18.370020964360588,
      "grad_norm": 0.3396233916282654,
      "learning_rate": 1.0054839639782766e-06,
      "loss": 0.5084,
      "num_input_tokens_seen": 22930200,
      "step": 35050
    },
    {
      "epoch": 18.37264150943396,
      "grad_norm": 0.7847930788993835,
      "learning_rate": 1.00227628692586e-06,
      "loss": 0.4999,
      "num_input_tokens_seen": 22933592,
      "step": 35055
    },
    {
      "epoch": 18.37526205450734,
      "grad_norm": 0.5017517805099487,
      "learning_rate": 9.990736299234692e-07,
      "loss": 0.3657,
      "num_input_tokens_seen": 22936312,
      "step": 35060
    },
    {
      "epoch": 18.377882599580712,
      "grad_norm": 0.3864693343639374,
      "learning_rate": 9.958759936410573e-07,
      "loss": 0.4355,
      "num_input_tokens_seen": 22939352,
      "step": 35065
    },
    {
      "epoch": 18.38050314465409,
      "grad_norm": 0.5600377321243286,
      "learning_rate": 9.926833787475304e-07,
      "loss": 0.5305,
      "num_input_tokens_seen": 22942552,
      "step": 35070
    },
    {
      "epoch": 18.383123689727462,
      "grad_norm": 0.40713730454444885,
      "learning_rate": 9.89495785910749e-07,
      "loss": 0.3761,
      "num_input_tokens_seen": 22945592,
      "step": 35075
    },
    {
      "epoch": 18.38574423480084,
      "grad_norm": 0.580375075340271,
      "learning_rate": 9.86313215797524e-07,
      "loss": 0.401,
      "num_input_tokens_seen": 22948280,
      "step": 35080
    },
    {
      "epoch": 18.388364779874212,
      "grad_norm": 0.4196126163005829,
      "learning_rate": 9.831356690736149e-07,
      "loss": 0.4577,
      "num_input_tokens_seen": 22951512,
      "step": 35085
    },
    {
      "epoch": 18.39098532494759,
      "grad_norm": 0.37600114941596985,
      "learning_rate": 9.799631464037256e-07,
      "loss": 0.5174,
      "num_input_tokens_seen": 22954904,
      "step": 35090
    },
    {
      "epoch": 18.393605870020963,
      "grad_norm": 0.37146279215812683,
      "learning_rate": 9.76795648451509e-07,
      "loss": 0.3794,
      "num_input_tokens_seen": 22958136,
      "step": 35095
    },
    {
      "epoch": 18.39622641509434,
      "grad_norm": 0.8391894698143005,
      "learning_rate": 9.73633175879568e-07,
      "loss": 0.3579,
      "num_input_tokens_seen": 22961720,
      "step": 35100
    },
    {
      "epoch": 18.398846960167713,
      "grad_norm": 0.7663734555244446,
      "learning_rate": 9.704757293494627e-07,
      "loss": 0.4002,
      "num_input_tokens_seen": 22964856,
      "step": 35105
    },
    {
      "epoch": 18.40146750524109,
      "grad_norm": 0.4515080153942108,
      "learning_rate": 9.673233095216922e-07,
      "loss": 0.3884,
      "num_input_tokens_seen": 22967864,
      "step": 35110
    },
    {
      "epoch": 18.404088050314467,
      "grad_norm": 0.4241914451122284,
      "learning_rate": 9.641759170557068e-07,
      "loss": 0.4312,
      "num_input_tokens_seen": 22971704,
      "step": 35115
    },
    {
      "epoch": 18.40670859538784,
      "grad_norm": 0.6222153902053833,
      "learning_rate": 9.610335526099024e-07,
      "loss": 0.4685,
      "num_input_tokens_seen": 22975384,
      "step": 35120
    },
    {
      "epoch": 18.409329140461217,
      "grad_norm": 0.34409788250923157,
      "learning_rate": 9.578962168416278e-07,
      "loss": 0.3963,
      "num_input_tokens_seen": 22978488,
      "step": 35125
    },
    {
      "epoch": 18.41194968553459,
      "grad_norm": 0.45272108912467957,
      "learning_rate": 9.547639104071775e-07,
      "loss": 0.5307,
      "num_input_tokens_seen": 22982136,
      "step": 35130
    },
    {
      "epoch": 18.414570230607968,
      "grad_norm": 0.45663467049598694,
      "learning_rate": 9.516366339617944e-07,
      "loss": 0.5078,
      "num_input_tokens_seen": 22984440,
      "step": 35135
    },
    {
      "epoch": 18.41719077568134,
      "grad_norm": 0.31009984016418457,
      "learning_rate": 9.485143881596714e-07,
      "loss": 0.414,
      "num_input_tokens_seen": 22987800,
      "step": 35140
    },
    {
      "epoch": 18.419811320754718,
      "grad_norm": 0.4569532573223114,
      "learning_rate": 9.453971736539447e-07,
      "loss": 0.5192,
      "num_input_tokens_seen": 22990584,
      "step": 35145
    },
    {
      "epoch": 18.42243186582809,
      "grad_norm": 0.5967194437980652,
      "learning_rate": 9.42284991096698e-07,
      "loss": 0.4008,
      "num_input_tokens_seen": 22994008,
      "step": 35150
    },
    {
      "epoch": 18.42505241090147,
      "grad_norm": 0.8227834105491638,
      "learning_rate": 9.391778411389634e-07,
      "loss": 0.4817,
      "num_input_tokens_seen": 22998136,
      "step": 35155
    },
    {
      "epoch": 18.427672955974842,
      "grad_norm": 0.7176875472068787,
      "learning_rate": 9.360757244307267e-07,
      "loss": 0.366,
      "num_input_tokens_seen": 23000568,
      "step": 35160
    },
    {
      "epoch": 18.43029350104822,
      "grad_norm": 0.40778595209121704,
      "learning_rate": 9.329786416209157e-07,
      "loss": 0.3925,
      "num_input_tokens_seen": 23004056,
      "step": 35165
    },
    {
      "epoch": 18.432914046121592,
      "grad_norm": 0.4293075203895569,
      "learning_rate": 9.298865933574014e-07,
      "loss": 0.5492,
      "num_input_tokens_seen": 23007640,
      "step": 35170
    },
    {
      "epoch": 18.43553459119497,
      "grad_norm": 0.3995901942253113,
      "learning_rate": 9.267995802870022e-07,
      "loss": 0.3992,
      "num_input_tokens_seen": 23010488,
      "step": 35175
    },
    {
      "epoch": 18.438155136268342,
      "grad_norm": 0.5385372638702393,
      "learning_rate": 9.237176030554878e-07,
      "loss": 0.4114,
      "num_input_tokens_seen": 23014552,
      "step": 35180
    },
    {
      "epoch": 18.44077568134172,
      "grad_norm": 0.5418314337730408,
      "learning_rate": 9.206406623075758e-07,
      "loss": 0.5349,
      "num_input_tokens_seen": 23018840,
      "step": 35185
    },
    {
      "epoch": 18.443396226415093,
      "grad_norm": 0.4183289706707001,
      "learning_rate": 9.175687586869292e-07,
      "loss": 0.6375,
      "num_input_tokens_seen": 23022520,
      "step": 35190
    },
    {
      "epoch": 18.44601677148847,
      "grad_norm": 0.3233988583087921,
      "learning_rate": 9.145018928361476e-07,
      "loss": 0.575,
      "num_input_tokens_seen": 23026104,
      "step": 35195
    },
    {
      "epoch": 18.448637316561843,
      "grad_norm": 0.3432331681251526,
      "learning_rate": 9.114400653967903e-07,
      "loss": 0.5943,
      "num_input_tokens_seen": 23030296,
      "step": 35200
    },
    {
      "epoch": 18.45125786163522,
      "grad_norm": 0.6344202756881714,
      "learning_rate": 9.083832770093531e-07,
      "loss": 0.4562,
      "num_input_tokens_seen": 23033944,
      "step": 35205
    },
    {
      "epoch": 18.453878406708597,
      "grad_norm": 0.449180543422699,
      "learning_rate": 9.053315283132802e-07,
      "loss": 0.5731,
      "num_input_tokens_seen": 23037912,
      "step": 35210
    },
    {
      "epoch": 18.45649895178197,
      "grad_norm": 0.7942148447036743,
      "learning_rate": 9.022848199469663e-07,
      "loss": 0.6219,
      "num_input_tokens_seen": 23041048,
      "step": 35215
    },
    {
      "epoch": 18.459119496855347,
      "grad_norm": 0.6382949352264404,
      "learning_rate": 8.992431525477518e-07,
      "loss": 0.4171,
      "num_input_tokens_seen": 23043992,
      "step": 35220
    },
    {
      "epoch": 18.46174004192872,
      "grad_norm": 0.44820088148117065,
      "learning_rate": 8.962065267519082e-07,
      "loss": 0.5147,
      "num_input_tokens_seen": 23047352,
      "step": 35225
    },
    {
      "epoch": 18.464360587002098,
      "grad_norm": 0.32827404141426086,
      "learning_rate": 8.931749431946718e-07,
      "loss": 0.5947,
      "num_input_tokens_seen": 23050488,
      "step": 35230
    },
    {
      "epoch": 18.46698113207547,
      "grad_norm": 0.6609553694725037,
      "learning_rate": 8.901484025102103e-07,
      "loss": 0.5069,
      "num_input_tokens_seen": 23053496,
      "step": 35235
    },
    {
      "epoch": 18.469601677148848,
      "grad_norm": 0.4270005226135254,
      "learning_rate": 8.871269053316422e-07,
      "loss": 0.4267,
      "num_input_tokens_seen": 23056760,
      "step": 35240
    },
    {
      "epoch": 18.47222222222222,
      "grad_norm": 0.359977662563324,
      "learning_rate": 8.841104522910343e-07,
      "loss": 0.4775,
      "num_input_tokens_seen": 23060152,
      "step": 35245
    },
    {
      "epoch": 18.4748427672956,
      "grad_norm": 0.7449907660484314,
      "learning_rate": 8.8109904401939e-07,
      "loss": 0.4349,
      "num_input_tokens_seen": 23063448,
      "step": 35250
    },
    {
      "epoch": 18.47746331236897,
      "grad_norm": 0.5618864893913269,
      "learning_rate": 8.780926811466639e-07,
      "loss": 0.4615,
      "num_input_tokens_seen": 23066776,
      "step": 35255
    },
    {
      "epoch": 18.48008385744235,
      "grad_norm": 0.5553646087646484,
      "learning_rate": 8.7509136430175e-07,
      "loss": 0.538,
      "num_input_tokens_seen": 23069784,
      "step": 35260
    },
    {
      "epoch": 18.482704402515722,
      "grad_norm": 0.4730527400970459,
      "learning_rate": 8.720950941124877e-07,
      "loss": 0.4854,
      "num_input_tokens_seen": 23072920,
      "step": 35265
    },
    {
      "epoch": 18.4853249475891,
      "grad_norm": 0.36225834488868713,
      "learning_rate": 8.691038712056732e-07,
      "loss": 0.4727,
      "num_input_tokens_seen": 23076408,
      "step": 35270
    },
    {
      "epoch": 18.487945492662472,
      "grad_norm": 0.9732751250267029,
      "learning_rate": 8.66117696207025e-07,
      "loss": 0.4933,
      "num_input_tokens_seen": 23079256,
      "step": 35275
    },
    {
      "epoch": 18.49056603773585,
      "grad_norm": 0.5863161087036133,
      "learning_rate": 8.631365697412186e-07,
      "loss": 0.3896,
      "num_input_tokens_seen": 23082488,
      "step": 35280
    },
    {
      "epoch": 18.493186582809223,
      "grad_norm": 0.4459501802921295,
      "learning_rate": 8.601604924318802e-07,
      "loss": 0.4044,
      "num_input_tokens_seen": 23089176,
      "step": 35285
    },
    {
      "epoch": 18.4958071278826,
      "grad_norm": 0.7117477059364319,
      "learning_rate": 8.571894649015588e-07,
      "loss": 0.4179,
      "num_input_tokens_seen": 23091928,
      "step": 35290
    },
    {
      "epoch": 18.498427672955973,
      "grad_norm": 0.3791217803955078,
      "learning_rate": 8.542234877717686e-07,
      "loss": 0.3955,
      "num_input_tokens_seen": 23095064,
      "step": 35295
    },
    {
      "epoch": 18.50104821802935,
      "grad_norm": 0.7577242255210876,
      "learning_rate": 8.512625616629549e-07,
      "loss": 0.4077,
      "num_input_tokens_seen": 23098616,
      "step": 35300
    },
    {
      "epoch": 18.503668763102727,
      "grad_norm": 0.8431711792945862,
      "learning_rate": 8.483066871945083e-07,
      "loss": 0.439,
      "num_input_tokens_seen": 23101784,
      "step": 35305
    },
    {
      "epoch": 18.5062893081761,
      "grad_norm": 0.5407892465591431,
      "learning_rate": 8.453558649847676e-07,
      "loss": 0.3218,
      "num_input_tokens_seen": 23105816,
      "step": 35310
    },
    {
      "epoch": 18.508909853249477,
      "grad_norm": 0.4435403347015381,
      "learning_rate": 8.424100956510112e-07,
      "loss": 0.2714,
      "num_input_tokens_seen": 23110232,
      "step": 35315
    },
    {
      "epoch": 18.51153039832285,
      "grad_norm": 0.5866396427154541,
      "learning_rate": 8.394693798094544e-07,
      "loss": 0.3658,
      "num_input_tokens_seen": 23113624,
      "step": 35320
    },
    {
      "epoch": 18.514150943396228,
      "grad_norm": 1.1286261081695557,
      "learning_rate": 8.365337180752691e-07,
      "loss": 0.4104,
      "num_input_tokens_seen": 23117176,
      "step": 35325
    },
    {
      "epoch": 18.5167714884696,
      "grad_norm": 0.3883405029773712,
      "learning_rate": 8.336031110625558e-07,
      "loss": 0.4219,
      "num_input_tokens_seen": 23120472,
      "step": 35330
    },
    {
      "epoch": 18.519392033542978,
      "grad_norm": 0.6315858364105225,
      "learning_rate": 8.306775593843657e-07,
      "loss": 0.3823,
      "num_input_tokens_seen": 23124344,
      "step": 35335
    },
    {
      "epoch": 18.52201257861635,
      "grad_norm": 0.49803414940834045,
      "learning_rate": 8.277570636526955e-07,
      "loss": 0.4937,
      "num_input_tokens_seen": 23127608,
      "step": 35340
    },
    {
      "epoch": 18.52463312368973,
      "grad_norm": 0.39796483516693115,
      "learning_rate": 8.24841624478473e-07,
      "loss": 0.4468,
      "num_input_tokens_seen": 23130552,
      "step": 35345
    },
    {
      "epoch": 18.5272536687631,
      "grad_norm": 0.5725746750831604,
      "learning_rate": 8.219312424715825e-07,
      "loss": 0.5063,
      "num_input_tokens_seen": 23133784,
      "step": 35350
    },
    {
      "epoch": 18.52987421383648,
      "grad_norm": 0.5503124594688416,
      "learning_rate": 8.190259182408345e-07,
      "loss": 0.4776,
      "num_input_tokens_seen": 23138072,
      "step": 35355
    },
    {
      "epoch": 18.532494758909852,
      "grad_norm": 0.42898809909820557,
      "learning_rate": 8.161256523939925e-07,
      "loss": 0.4245,
      "num_input_tokens_seen": 23141080,
      "step": 35360
    },
    {
      "epoch": 18.53511530398323,
      "grad_norm": 0.7960025668144226,
      "learning_rate": 8.132304455377576e-07,
      "loss": 0.5679,
      "num_input_tokens_seen": 23143960,
      "step": 35365
    },
    {
      "epoch": 18.537735849056602,
      "grad_norm": 0.43494275212287903,
      "learning_rate": 8.103402982777813e-07,
      "loss": 0.5119,
      "num_input_tokens_seen": 23147608,
      "step": 35370
    },
    {
      "epoch": 18.54035639412998,
      "grad_norm": 0.3274831771850586,
      "learning_rate": 8.07455211218644e-07,
      "loss": 0.425,
      "num_input_tokens_seen": 23150872,
      "step": 35375
    },
    {
      "epoch": 18.542976939203353,
      "grad_norm": 0.4534812271595001,
      "learning_rate": 8.045751849638738e-07,
      "loss": 0.4668,
      "num_input_tokens_seen": 23153656,
      "step": 35380
    },
    {
      "epoch": 18.54559748427673,
      "grad_norm": 0.3194688856601715,
      "learning_rate": 8.017002201159363e-07,
      "loss": 0.343,
      "num_input_tokens_seen": 23157176,
      "step": 35385
    },
    {
      "epoch": 18.548218029350103,
      "grad_norm": 0.4070020914077759,
      "learning_rate": 7.988303172762418e-07,
      "loss": 0.3739,
      "num_input_tokens_seen": 23160280,
      "step": 35390
    },
    {
      "epoch": 18.55083857442348,
      "grad_norm": 0.37392038106918335,
      "learning_rate": 7.959654770451463e-07,
      "loss": 0.2931,
      "num_input_tokens_seen": 23166552,
      "step": 35395
    },
    {
      "epoch": 18.553459119496857,
      "grad_norm": 0.4718474745750427,
      "learning_rate": 7.9310570002194e-07,
      "loss": 0.3773,
      "num_input_tokens_seen": 23170008,
      "step": 35400
    },
    {
      "epoch": 18.55607966457023,
      "grad_norm": 0.5614501237869263,
      "learning_rate": 7.902509868048552e-07,
      "loss": 0.4668,
      "num_input_tokens_seen": 23173336,
      "step": 35405
    },
    {
      "epoch": 18.558700209643607,
      "grad_norm": 1.051094889640808,
      "learning_rate": 7.874013379910616e-07,
      "loss": 0.5145,
      "num_input_tokens_seen": 23177336,
      "step": 35410
    },
    {
      "epoch": 18.56132075471698,
      "grad_norm": 0.2973920702934265,
      "learning_rate": 7.845567541766769e-07,
      "loss": 0.476,
      "num_input_tokens_seen": 23180632,
      "step": 35415
    },
    {
      "epoch": 18.563941299790358,
      "grad_norm": 0.7699101567268372,
      "learning_rate": 7.817172359567526e-07,
      "loss": 0.4971,
      "num_input_tokens_seen": 23183192,
      "step": 35420
    },
    {
      "epoch": 18.56656184486373,
      "grad_norm": 0.303469717502594,
      "learning_rate": 7.788827839252888e-07,
      "loss": 0.3775,
      "num_input_tokens_seen": 23185912,
      "step": 35425
    },
    {
      "epoch": 18.569182389937108,
      "grad_norm": 0.6848767399787903,
      "learning_rate": 7.760533986752222e-07,
      "loss": 0.6211,
      "num_input_tokens_seen": 23189368,
      "step": 35430
    },
    {
      "epoch": 18.57180293501048,
      "grad_norm": 0.38690754771232605,
      "learning_rate": 7.732290807984155e-07,
      "loss": 0.4336,
      "num_input_tokens_seen": 23192728,
      "step": 35435
    },
    {
      "epoch": 18.57442348008386,
      "grad_norm": 0.37054142355918884,
      "learning_rate": 7.704098308856933e-07,
      "loss": 0.5132,
      "num_input_tokens_seen": 23195640,
      "step": 35440
    },
    {
      "epoch": 18.57704402515723,
      "grad_norm": 0.5992307662963867,
      "learning_rate": 7.675956495268088e-07,
      "loss": 0.5984,
      "num_input_tokens_seen": 23198200,
      "step": 35445
    },
    {
      "epoch": 18.57966457023061,
      "grad_norm": 0.5410493612289429,
      "learning_rate": 7.647865373104551e-07,
      "loss": 0.3863,
      "num_input_tokens_seen": 23200760,
      "step": 35450
    },
    {
      "epoch": 18.582285115303982,
      "grad_norm": 0.4863520562648773,
      "learning_rate": 7.619824948242732e-07,
      "loss": 0.4081,
      "num_input_tokens_seen": 23204280,
      "step": 35455
    },
    {
      "epoch": 18.58490566037736,
      "grad_norm": 0.8495361804962158,
      "learning_rate": 7.5918352265483e-07,
      "loss": 0.5586,
      "num_input_tokens_seen": 23207800,
      "step": 35460
    },
    {
      "epoch": 18.587526205450732,
      "grad_norm": 0.35675448179244995,
      "learning_rate": 7.563896213876348e-07,
      "loss": 0.3935,
      "num_input_tokens_seen": 23210648,
      "step": 35465
    },
    {
      "epoch": 18.59014675052411,
      "grad_norm": 0.484194278717041,
      "learning_rate": 7.53600791607148e-07,
      "loss": 0.3137,
      "num_input_tokens_seen": 23213752,
      "step": 35470
    },
    {
      "epoch": 18.592767295597483,
      "grad_norm": 0.3271220028400421,
      "learning_rate": 7.508170338967585e-07,
      "loss": 0.3457,
      "num_input_tokens_seen": 23216408,
      "step": 35475
    },
    {
      "epoch": 18.59538784067086,
      "grad_norm": 0.44615504145622253,
      "learning_rate": 7.480383488387976e-07,
      "loss": 0.4056,
      "num_input_tokens_seen": 23219480,
      "step": 35480
    },
    {
      "epoch": 18.598008385744233,
      "grad_norm": 0.6021988391876221,
      "learning_rate": 7.452647370145338e-07,
      "loss": 0.5387,
      "num_input_tokens_seen": 23222872,
      "step": 35485
    },
    {
      "epoch": 18.60062893081761,
      "grad_norm": 0.4693080484867096,
      "learning_rate": 7.424961990041696e-07,
      "loss": 0.5807,
      "num_input_tokens_seen": 23226584,
      "step": 35490
    },
    {
      "epoch": 18.603249475890987,
      "grad_norm": 0.4243025779724121,
      "learning_rate": 7.397327353868583e-07,
      "loss": 0.445,
      "num_input_tokens_seen": 23229112,
      "step": 35495
    },
    {
      "epoch": 18.60587002096436,
      "grad_norm": 0.3134455680847168,
      "learning_rate": 7.369743467406848e-07,
      "loss": 0.3234,
      "num_input_tokens_seen": 23231800,
      "step": 35500
    },
    {
      "epoch": 18.608490566037737,
      "grad_norm": 0.624154806137085,
      "learning_rate": 7.34221033642668e-07,
      "loss": 0.3985,
      "num_input_tokens_seen": 23235192,
      "step": 35505
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.9417574405670166,
      "learning_rate": 7.314727966687751e-07,
      "loss": 0.4381,
      "num_input_tokens_seen": 23238008,
      "step": 35510
    },
    {
      "epoch": 18.613731656184488,
      "grad_norm": 0.6083296537399292,
      "learning_rate": 7.287296363939045e-07,
      "loss": 0.4998,
      "num_input_tokens_seen": 23240536,
      "step": 35515
    },
    {
      "epoch": 18.61635220125786,
      "grad_norm": 0.3878071904182434,
      "learning_rate": 7.259915533918888e-07,
      "loss": 0.5748,
      "num_input_tokens_seen": 23243576,
      "step": 35520
    },
    {
      "epoch": 18.618972746331238,
      "grad_norm": 0.3856129050254822,
      "learning_rate": 7.232585482355086e-07,
      "loss": 0.4372,
      "num_input_tokens_seen": 23247256,
      "step": 35525
    },
    {
      "epoch": 18.62159329140461,
      "grad_norm": 0.2879435420036316,
      "learning_rate": 7.205306214964736e-07,
      "loss": 0.3352,
      "num_input_tokens_seen": 23250456,
      "step": 35530
    },
    {
      "epoch": 18.62421383647799,
      "grad_norm": 0.22817090153694153,
      "learning_rate": 7.178077737454436e-07,
      "loss": 0.4192,
      "num_input_tokens_seen": 23253272,
      "step": 35535
    },
    {
      "epoch": 18.62683438155136,
      "grad_norm": 0.39221805334091187,
      "learning_rate": 7.150900055519993e-07,
      "loss": 0.47,
      "num_input_tokens_seen": 23256824,
      "step": 35540
    },
    {
      "epoch": 18.62945492662474,
      "grad_norm": 0.6547248959541321,
      "learning_rate": 7.123773174846693e-07,
      "loss": 0.5624,
      "num_input_tokens_seen": 23261048,
      "step": 35545
    },
    {
      "epoch": 18.632075471698112,
      "grad_norm": 0.6976428627967834,
      "learning_rate": 7.096697101109134e-07,
      "loss": 0.4842,
      "num_input_tokens_seen": 23264216,
      "step": 35550
    },
    {
      "epoch": 18.63469601677149,
      "grad_norm": 0.37893539667129517,
      "learning_rate": 7.069671839971371e-07,
      "loss": 0.4323,
      "num_input_tokens_seen": 23267448,
      "step": 35555
    },
    {
      "epoch": 18.637316561844862,
      "grad_norm": 0.6034651398658752,
      "learning_rate": 7.042697397086795e-07,
      "loss": 0.4396,
      "num_input_tokens_seen": 23271448,
      "step": 35560
    },
    {
      "epoch": 18.63993710691824,
      "grad_norm": 0.5375216007232666,
      "learning_rate": 7.015773778098089e-07,
      "loss": 0.501,
      "num_input_tokens_seen": 23274296,
      "step": 35565
    },
    {
      "epoch": 18.642557651991613,
      "grad_norm": 1.059113621711731,
      "learning_rate": 6.988900988637415e-07,
      "loss": 0.3999,
      "num_input_tokens_seen": 23276664,
      "step": 35570
    },
    {
      "epoch": 18.64517819706499,
      "grad_norm": 0.38731250166893005,
      "learning_rate": 6.962079034326218e-07,
      "loss": 0.4041,
      "num_input_tokens_seen": 23280632,
      "step": 35575
    },
    {
      "epoch": 18.647798742138363,
      "grad_norm": 0.44187963008880615,
      "learning_rate": 6.935307920775346e-07,
      "loss": 0.307,
      "num_input_tokens_seen": 23285624,
      "step": 35580
    },
    {
      "epoch": 18.65041928721174,
      "grad_norm": 0.4346022307872772,
      "learning_rate": 6.90858765358507e-07,
      "loss": 0.3964,
      "num_input_tokens_seen": 23288632,
      "step": 35585
    },
    {
      "epoch": 18.653039832285117,
      "grad_norm": 0.4517505168914795,
      "learning_rate": 6.881918238344915e-07,
      "loss": 0.4267,
      "num_input_tokens_seen": 23291800,
      "step": 35590
    },
    {
      "epoch": 18.65566037735849,
      "grad_norm": 0.4957241714000702,
      "learning_rate": 6.855299680633808e-07,
      "loss": 0.4329,
      "num_input_tokens_seen": 23296088,
      "step": 35595
    },
    {
      "epoch": 18.658280922431867,
      "grad_norm": 0.43185675144195557,
      "learning_rate": 6.828731986020104e-07,
      "loss": 0.3788,
      "num_input_tokens_seen": 23299096,
      "step": 35600
    },
    {
      "epoch": 18.66090146750524,
      "grad_norm": 0.338627427816391,
      "learning_rate": 6.802215160061381e-07,
      "loss": 0.4889,
      "num_input_tokens_seen": 23301624,
      "step": 35605
    },
    {
      "epoch": 18.663522012578618,
      "grad_norm": 0.4357195496559143,
      "learning_rate": 6.775749208304704e-07,
      "loss": 0.5239,
      "num_input_tokens_seen": 23305240,
      "step": 35610
    },
    {
      "epoch": 18.66614255765199,
      "grad_norm": 0.5364294648170471,
      "learning_rate": 6.749334136286478e-07,
      "loss": 0.4099,
      "num_input_tokens_seen": 23308344,
      "step": 35615
    },
    {
      "epoch": 18.668763102725368,
      "grad_norm": 0.46074098348617554,
      "learning_rate": 6.722969949532365e-07,
      "loss": 0.4972,
      "num_input_tokens_seen": 23313272,
      "step": 35620
    },
    {
      "epoch": 18.67138364779874,
      "grad_norm": 0.4610159993171692,
      "learning_rate": 6.69665665355751e-07,
      "loss": 0.5377,
      "num_input_tokens_seen": 23316088,
      "step": 35625
    },
    {
      "epoch": 18.67400419287212,
      "grad_norm": 0.6124574542045593,
      "learning_rate": 6.670394253866314e-07,
      "loss": 0.4086,
      "num_input_tokens_seen": 23320312,
      "step": 35630
    },
    {
      "epoch": 18.67662473794549,
      "grad_norm": 1.4352399110794067,
      "learning_rate": 6.644182755952605e-07,
      "loss": 0.4022,
      "num_input_tokens_seen": 23322936,
      "step": 35635
    },
    {
      "epoch": 18.67924528301887,
      "grad_norm": 0.5031679272651672,
      "learning_rate": 6.618022165299526e-07,
      "loss": 0.5261,
      "num_input_tokens_seen": 23326968,
      "step": 35640
    },
    {
      "epoch": 18.681865828092242,
      "grad_norm": 0.5714063048362732,
      "learning_rate": 6.591912487379531e-07,
      "loss": 0.4589,
      "num_input_tokens_seen": 23330264,
      "step": 35645
    },
    {
      "epoch": 18.68448637316562,
      "grad_norm": 0.46981948614120483,
      "learning_rate": 6.565853727654503e-07,
      "loss": 0.3765,
      "num_input_tokens_seen": 23333912,
      "step": 35650
    },
    {
      "epoch": 18.687106918238992,
      "grad_norm": 0.7093966007232666,
      "learning_rate": 6.539845891575636e-07,
      "loss": 0.3713,
      "num_input_tokens_seen": 23336920,
      "step": 35655
    },
    {
      "epoch": 18.68972746331237,
      "grad_norm": 0.49234387278556824,
      "learning_rate": 6.513888984583466e-07,
      "loss": 0.4963,
      "num_input_tokens_seen": 23339800,
      "step": 35660
    },
    {
      "epoch": 18.692348008385743,
      "grad_norm": 0.5073952078819275,
      "learning_rate": 6.487983012107873e-07,
      "loss": 0.4064,
      "num_input_tokens_seen": 23343608,
      "step": 35665
    },
    {
      "epoch": 18.69496855345912,
      "grad_norm": 0.4996681213378906,
      "learning_rate": 6.462127979568133e-07,
      "loss": 0.488,
      "num_input_tokens_seen": 23346456,
      "step": 35670
    },
    {
      "epoch": 18.697589098532493,
      "grad_norm": 0.3995349705219269,
      "learning_rate": 6.436323892372781e-07,
      "loss": 0.4639,
      "num_input_tokens_seen": 23349336,
      "step": 35675
    },
    {
      "epoch": 18.70020964360587,
      "grad_norm": 0.5397994518280029,
      "learning_rate": 6.410570755919721e-07,
      "loss": 0.3661,
      "num_input_tokens_seen": 23352344,
      "step": 35680
    },
    {
      "epoch": 18.702830188679247,
      "grad_norm": 0.32077643275260925,
      "learning_rate": 6.384868575596281e-07,
      "loss": 0.3513,
      "num_input_tokens_seen": 23355064,
      "step": 35685
    },
    {
      "epoch": 18.70545073375262,
      "grad_norm": 0.5771808624267578,
      "learning_rate": 6.359217356779024e-07,
      "loss": 0.4394,
      "num_input_tokens_seen": 23358680,
      "step": 35690
    },
    {
      "epoch": 18.708071278825997,
      "grad_norm": 0.7368121147155762,
      "learning_rate": 6.333617104833878e-07,
      "loss": 0.3622,
      "num_input_tokens_seen": 23361496,
      "step": 35695
    },
    {
      "epoch": 18.71069182389937,
      "grad_norm": 0.5489485859870911,
      "learning_rate": 6.308067825116142e-07,
      "loss": 0.5159,
      "num_input_tokens_seen": 23366264,
      "step": 35700
    },
    {
      "epoch": 18.713312368972748,
      "grad_norm": 0.6524865627288818,
      "learning_rate": 6.282569522970433e-07,
      "loss": 0.4328,
      "num_input_tokens_seen": 23368888,
      "step": 35705
    },
    {
      "epoch": 18.71593291404612,
      "grad_norm": 0.29598468542099,
      "learning_rate": 6.257122203730731e-07,
      "loss": 0.3771,
      "num_input_tokens_seen": 23372216,
      "step": 35710
    },
    {
      "epoch": 18.718553459119498,
      "grad_norm": 0.40113696455955505,
      "learning_rate": 6.231725872720251e-07,
      "loss": 0.4463,
      "num_input_tokens_seen": 23376248,
      "step": 35715
    },
    {
      "epoch": 18.72117400419287,
      "grad_norm": 0.2819842994213104,
      "learning_rate": 6.206380535251716e-07,
      "loss": 0.4255,
      "num_input_tokens_seen": 23379544,
      "step": 35720
    },
    {
      "epoch": 18.72379454926625,
      "grad_norm": 0.3793524503707886,
      "learning_rate": 6.181086196626995e-07,
      "loss": 0.5164,
      "num_input_tokens_seen": 23383576,
      "step": 35725
    },
    {
      "epoch": 18.72641509433962,
      "grad_norm": 0.2265443205833435,
      "learning_rate": 6.155842862137384e-07,
      "loss": 0.5373,
      "num_input_tokens_seen": 23387352,
      "step": 35730
    },
    {
      "epoch": 18.729035639413,
      "grad_norm": 0.45103713870048523,
      "learning_rate": 6.130650537063548e-07,
      "loss": 0.2946,
      "num_input_tokens_seen": 23390808,
      "step": 35735
    },
    {
      "epoch": 18.731656184486372,
      "grad_norm": 0.4467643201351166,
      "learning_rate": 6.105509226675438e-07,
      "loss": 0.392,
      "num_input_tokens_seen": 23393432,
      "step": 35740
    },
    {
      "epoch": 18.73427672955975,
      "grad_norm": 0.298188716173172,
      "learning_rate": 6.08041893623229e-07,
      "loss": 0.3136,
      "num_input_tokens_seen": 23396536,
      "step": 35745
    },
    {
      "epoch": 18.736897274633122,
      "grad_norm": 0.40377333760261536,
      "learning_rate": 6.055379670982713e-07,
      "loss": 0.2998,
      "num_input_tokens_seen": 23399416,
      "step": 35750
    },
    {
      "epoch": 18.7395178197065,
      "grad_norm": 0.5363641977310181,
      "learning_rate": 6.030391436164628e-07,
      "loss": 0.3901,
      "num_input_tokens_seen": 23402392,
      "step": 35755
    },
    {
      "epoch": 18.742138364779873,
      "grad_norm": 0.5023297071456909,
      "learning_rate": 6.005454237005298e-07,
      "loss": 0.369,
      "num_input_tokens_seen": 23404792,
      "step": 35760
    },
    {
      "epoch": 18.74475890985325,
      "grad_norm": 0.7973053455352783,
      "learning_rate": 5.980568078721327e-07,
      "loss": 0.3559,
      "num_input_tokens_seen": 23408472,
      "step": 35765
    },
    {
      "epoch": 18.747379454926623,
      "grad_norm": 0.30086439847946167,
      "learning_rate": 5.955732966518606e-07,
      "loss": 0.4075,
      "num_input_tokens_seen": 23411768,
      "step": 35770
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.7724984884262085,
      "learning_rate": 5.930948905592343e-07,
      "loss": 0.4719,
      "num_input_tokens_seen": 23415128,
      "step": 35775
    },
    {
      "epoch": 18.752620545073377,
      "grad_norm": 0.3723592460155487,
      "learning_rate": 5.906215901127054e-07,
      "loss": 0.418,
      "num_input_tokens_seen": 23418008,
      "step": 35780
    },
    {
      "epoch": 18.75524109014675,
      "grad_norm": 0.7731172442436218,
      "learning_rate": 5.881533958296631e-07,
      "loss": 0.4349,
      "num_input_tokens_seen": 23420888,
      "step": 35785
    },
    {
      "epoch": 18.757861635220127,
      "grad_norm": 0.4520636200904846,
      "learning_rate": 5.856903082264248e-07,
      "loss": 0.4763,
      "num_input_tokens_seen": 23423128,
      "step": 35790
    },
    {
      "epoch": 18.7604821802935,
      "grad_norm": 0.43905726075172424,
      "learning_rate": 5.832323278182422e-07,
      "loss": 0.4603,
      "num_input_tokens_seen": 23426136,
      "step": 35795
    },
    {
      "epoch": 18.763102725366878,
      "grad_norm": 0.4433368146419525,
      "learning_rate": 5.807794551192986e-07,
      "loss": 0.4457,
      "num_input_tokens_seen": 23429144,
      "step": 35800
    },
    {
      "epoch": 18.76572327044025,
      "grad_norm": 0.5004038214683533,
      "learning_rate": 5.783316906427e-07,
      "loss": 0.4195,
      "num_input_tokens_seen": 23432216,
      "step": 35805
    },
    {
      "epoch": 18.768343815513628,
      "grad_norm": 0.6883478760719299,
      "learning_rate": 5.758890349004925e-07,
      "loss": 0.3945,
      "num_input_tokens_seen": 23435000,
      "step": 35810
    },
    {
      "epoch": 18.770964360587,
      "grad_norm": 0.614679217338562,
      "learning_rate": 5.734514884036563e-07,
      "loss": 0.3441,
      "num_input_tokens_seen": 23438008,
      "step": 35815
    },
    {
      "epoch": 18.77358490566038,
      "grad_norm": 0.3790077865123749,
      "learning_rate": 5.710190516620945e-07,
      "loss": 0.5925,
      "num_input_tokens_seen": 23440504,
      "step": 35820
    },
    {
      "epoch": 18.77620545073375,
      "grad_norm": 0.6259176135063171,
      "learning_rate": 5.685917251846501e-07,
      "loss": 0.4645,
      "num_input_tokens_seen": 23443096,
      "step": 35825
    },
    {
      "epoch": 18.77882599580713,
      "grad_norm": 0.4880164563655853,
      "learning_rate": 5.661695094790892e-07,
      "loss": 0.3967,
      "num_input_tokens_seen": 23446392,
      "step": 35830
    },
    {
      "epoch": 18.781446540880502,
      "grad_norm": 0.49517393112182617,
      "learning_rate": 5.637524050521064e-07,
      "loss": 0.3512,
      "num_input_tokens_seen": 23449208,
      "step": 35835
    },
    {
      "epoch": 18.78406708595388,
      "grad_norm": 0.4539058208465576,
      "learning_rate": 5.613404124093391e-07,
      "loss": 0.4228,
      "num_input_tokens_seen": 23452376,
      "step": 35840
    },
    {
      "epoch": 18.786687631027252,
      "grad_norm": 0.7424442172050476,
      "learning_rate": 5.589335320553446e-07,
      "loss": 0.3403,
      "num_input_tokens_seen": 23455768,
      "step": 35845
    },
    {
      "epoch": 18.78930817610063,
      "grad_norm": 0.5466732978820801,
      "learning_rate": 5.565317644936202e-07,
      "loss": 0.4138,
      "num_input_tokens_seen": 23458680,
      "step": 35850
    },
    {
      "epoch": 18.791928721174003,
      "grad_norm": 0.5386579632759094,
      "learning_rate": 5.541351102265863e-07,
      "loss": 0.4506,
      "num_input_tokens_seen": 23462456,
      "step": 35855
    },
    {
      "epoch": 18.79454926624738,
      "grad_norm": 0.5081287622451782,
      "learning_rate": 5.51743569755589e-07,
      "loss": 0.5138,
      "num_input_tokens_seen": 23465656,
      "step": 35860
    },
    {
      "epoch": 18.797169811320753,
      "grad_norm": 0.7112874388694763,
      "learning_rate": 5.493571435809197e-07,
      "loss": 0.5377,
      "num_input_tokens_seen": 23468472,
      "step": 35865
    },
    {
      "epoch": 18.79979035639413,
      "grad_norm": 0.3830147385597229,
      "learning_rate": 5.469758322017849e-07,
      "loss": 0.4005,
      "num_input_tokens_seen": 23472248,
      "step": 35870
    },
    {
      "epoch": 18.802410901467507,
      "grad_norm": 0.4741698205471039,
      "learning_rate": 5.445996361163358e-07,
      "loss": 0.3645,
      "num_input_tokens_seen": 23474680,
      "step": 35875
    },
    {
      "epoch": 18.80503144654088,
      "grad_norm": 0.454013854265213,
      "learning_rate": 5.422285558216389e-07,
      "loss": 0.523,
      "num_input_tokens_seen": 23477368,
      "step": 35880
    },
    {
      "epoch": 18.807651991614257,
      "grad_norm": 0.6282473802566528,
      "learning_rate": 5.398625918137001e-07,
      "loss": 0.4943,
      "num_input_tokens_seen": 23481112,
      "step": 35885
    },
    {
      "epoch": 18.81027253668763,
      "grad_norm": 0.6314863562583923,
      "learning_rate": 5.375017445874486e-07,
      "loss": 0.4554,
      "num_input_tokens_seen": 23484088,
      "step": 35890
    },
    {
      "epoch": 18.812893081761008,
      "grad_norm": 0.433280348777771,
      "learning_rate": 5.351460146367504e-07,
      "loss": 0.4117,
      "num_input_tokens_seen": 23487288,
      "step": 35895
    },
    {
      "epoch": 18.81551362683438,
      "grad_norm": 0.6854654550552368,
      "learning_rate": 5.327954024543946e-07,
      "loss": 0.4713,
      "num_input_tokens_seen": 23490680,
      "step": 35900
    },
    {
      "epoch": 18.818134171907758,
      "grad_norm": 0.47358405590057373,
      "learning_rate": 5.304499085321074e-07,
      "loss": 0.4636,
      "num_input_tokens_seen": 23493144,
      "step": 35905
    },
    {
      "epoch": 18.82075471698113,
      "grad_norm": 0.5666239857673645,
      "learning_rate": 5.281095333605352e-07,
      "loss": 0.4938,
      "num_input_tokens_seen": 23495608,
      "step": 35910
    },
    {
      "epoch": 18.82337526205451,
      "grad_norm": 0.44136935472488403,
      "learning_rate": 5.257742774292556e-07,
      "loss": 0.3774,
      "num_input_tokens_seen": 23499544,
      "step": 35915
    },
    {
      "epoch": 18.82599580712788,
      "grad_norm": 0.7241557240486145,
      "learning_rate": 5.234441412267837e-07,
      "loss": 0.4009,
      "num_input_tokens_seen": 23502168,
      "step": 35920
    },
    {
      "epoch": 18.82861635220126,
      "grad_norm": 0.3265298306941986,
      "learning_rate": 5.211191252405517e-07,
      "loss": 0.4489,
      "num_input_tokens_seen": 23505816,
      "step": 35925
    },
    {
      "epoch": 18.831236897274632,
      "grad_norm": 0.5754569172859192,
      "learning_rate": 5.187992299569316e-07,
      "loss": 0.4625,
      "num_input_tokens_seen": 23509176,
      "step": 35930
    },
    {
      "epoch": 18.83385744234801,
      "grad_norm": 0.7571565508842468,
      "learning_rate": 5.164844558612131e-07,
      "loss": 0.4875,
      "num_input_tokens_seen": 23512024,
      "step": 35935
    },
    {
      "epoch": 18.836477987421382,
      "grad_norm": 0.47765469551086426,
      "learning_rate": 5.141748034376254e-07,
      "loss": 0.5067,
      "num_input_tokens_seen": 23514552,
      "step": 35940
    },
    {
      "epoch": 18.83909853249476,
      "grad_norm": 0.5405924916267395,
      "learning_rate": 5.118702731693209e-07,
      "loss": 0.5094,
      "num_input_tokens_seen": 23517528,
      "step": 35945
    },
    {
      "epoch": 18.841719077568133,
      "grad_norm": 0.3762573301792145,
      "learning_rate": 5.095708655383751e-07,
      "loss": 0.4128,
      "num_input_tokens_seen": 23521272,
      "step": 35950
    },
    {
      "epoch": 18.84433962264151,
      "grad_norm": 0.43900057673454285,
      "learning_rate": 5.07276581025809e-07,
      "loss": 0.4743,
      "num_input_tokens_seen": 23524632,
      "step": 35955
    },
    {
      "epoch": 18.846960167714883,
      "grad_norm": 0.32590025663375854,
      "learning_rate": 5.049874201115495e-07,
      "loss": 0.4021,
      "num_input_tokens_seen": 23531160,
      "step": 35960
    },
    {
      "epoch": 18.84958071278826,
      "grad_norm": 0.5335837602615356,
      "learning_rate": 5.027033832744693e-07,
      "loss": 0.5249,
      "num_input_tokens_seen": 23534104,
      "step": 35965
    },
    {
      "epoch": 18.852201257861637,
      "grad_norm": 0.42005011439323425,
      "learning_rate": 5.004244709923639e-07,
      "loss": 0.4564,
      "num_input_tokens_seen": 23536856,
      "step": 35970
    },
    {
      "epoch": 18.85482180293501,
      "grad_norm": 0.39906302094459534,
      "learning_rate": 4.981506837419519e-07,
      "loss": 0.5829,
      "num_input_tokens_seen": 23540152,
      "step": 35975
    },
    {
      "epoch": 18.857442348008387,
      "grad_norm": 0.4319671392440796,
      "learning_rate": 4.958820219988836e-07,
      "loss": 0.4906,
      "num_input_tokens_seen": 23543640,
      "step": 35980
    },
    {
      "epoch": 18.86006289308176,
      "grad_norm": 0.5816067457199097,
      "learning_rate": 4.936184862377402e-07,
      "loss": 0.4443,
      "num_input_tokens_seen": 23547064,
      "step": 35985
    },
    {
      "epoch": 18.862683438155138,
      "grad_norm": 0.5799731016159058,
      "learning_rate": 4.913600769320237e-07,
      "loss": 0.5028,
      "num_input_tokens_seen": 23549944,
      "step": 35990
    },
    {
      "epoch": 18.86530398322851,
      "grad_norm": 0.36982986330986023,
      "learning_rate": 4.891067945541728e-07,
      "loss": 0.4863,
      "num_input_tokens_seen": 23553560,
      "step": 35995
    },
    {
      "epoch": 18.867924528301888,
      "grad_norm": 0.6896655559539795,
      "learning_rate": 4.868586395755437e-07,
      "loss": 0.4641,
      "num_input_tokens_seen": 23556504,
      "step": 36000
    },
    {
      "epoch": 18.87054507337526,
      "grad_norm": 0.41878873109817505,
      "learning_rate": 4.846156124664242e-07,
      "loss": 0.5846,
      "num_input_tokens_seen": 23560376,
      "step": 36005
    },
    {
      "epoch": 18.87316561844864,
      "grad_norm": 1.0097169876098633,
      "learning_rate": 4.823777136960362e-07,
      "loss": 0.3293,
      "num_input_tokens_seen": 23563064,
      "step": 36010
    },
    {
      "epoch": 18.87578616352201,
      "grad_norm": 0.33046290278434753,
      "learning_rate": 4.801449437325162e-07,
      "loss": 0.3912,
      "num_input_tokens_seen": 23566456,
      "step": 36015
    },
    {
      "epoch": 18.87840670859539,
      "grad_norm": 0.36989185214042664,
      "learning_rate": 4.779173030429379e-07,
      "loss": 0.3739,
      "num_input_tokens_seen": 23569976,
      "step": 36020
    },
    {
      "epoch": 18.881027253668762,
      "grad_norm": 0.45481517910957336,
      "learning_rate": 4.7569479209329804e-07,
      "loss": 0.5555,
      "num_input_tokens_seen": 23573368,
      "step": 36025
    },
    {
      "epoch": 18.88364779874214,
      "grad_norm": 0.713714063167572,
      "learning_rate": 4.734774113485163e-07,
      "loss": 0.3739,
      "num_input_tokens_seen": 23576952,
      "step": 36030
    },
    {
      "epoch": 18.886268343815512,
      "grad_norm": 1.0188286304473877,
      "learning_rate": 4.7126516127244945e-07,
      "loss": 0.4276,
      "num_input_tokens_seen": 23580248,
      "step": 36035
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.43538278341293335,
      "learning_rate": 4.690580423278718e-07,
      "loss": 0.3557,
      "num_input_tokens_seen": 23583128,
      "step": 36040
    },
    {
      "epoch": 18.891509433962263,
      "grad_norm": 0.6534988284111023,
      "learning_rate": 4.668560549764861e-07,
      "loss": 0.4854,
      "num_input_tokens_seen": 23586104,
      "step": 36045
    },
    {
      "epoch": 18.89412997903564,
      "grad_norm": 0.4504760801792145,
      "learning_rate": 4.646591996789268e-07,
      "loss": 0.5109,
      "num_input_tokens_seen": 23589400,
      "step": 36050
    },
    {
      "epoch": 18.896750524109013,
      "grad_norm": 0.4979248344898224,
      "learning_rate": 4.6246747689474847e-07,
      "loss": 0.5031,
      "num_input_tokens_seen": 23592184,
      "step": 36055
    },
    {
      "epoch": 18.89937106918239,
      "grad_norm": 0.37942036986351013,
      "learning_rate": 4.6028088708243434e-07,
      "loss": 0.5069,
      "num_input_tokens_seen": 23595064,
      "step": 36060
    },
    {
      "epoch": 18.901991614255767,
      "grad_norm": 0.4947958290576935,
      "learning_rate": 4.580994306993991e-07,
      "loss": 0.3783,
      "num_input_tokens_seen": 23597432,
      "step": 36065
    },
    {
      "epoch": 18.90461215932914,
      "grad_norm": 0.5511537194252014,
      "learning_rate": 4.5592310820196936e-07,
      "loss": 0.4672,
      "num_input_tokens_seen": 23600600,
      "step": 36070
    },
    {
      "epoch": 18.907232704402517,
      "grad_norm": 0.4043079912662506,
      "learning_rate": 4.537519200454116e-07,
      "loss": 0.4487,
      "num_input_tokens_seen": 23604440,
      "step": 36075
    },
    {
      "epoch": 18.90985324947589,
      "grad_norm": 0.37681862711906433,
      "learning_rate": 4.515858666839151e-07,
      "loss": 0.5593,
      "num_input_tokens_seen": 23607736,
      "step": 36080
    },
    {
      "epoch": 18.912473794549268,
      "grad_norm": 0.6451094150543213,
      "learning_rate": 4.494249485705926e-07,
      "loss": 0.5197,
      "num_input_tokens_seen": 23615096,
      "step": 36085
    },
    {
      "epoch": 18.91509433962264,
      "grad_norm": 0.4948718845844269,
      "learning_rate": 4.472691661574824e-07,
      "loss": 0.3672,
      "num_input_tokens_seen": 23618584,
      "step": 36090
    },
    {
      "epoch": 18.917714884696018,
      "grad_norm": 0.682930052280426,
      "learning_rate": 4.4511851989555156e-07,
      "loss": 0.5941,
      "num_input_tokens_seen": 23622808,
      "step": 36095
    },
    {
      "epoch": 18.92033542976939,
      "grad_norm": 0.7741916179656982,
      "learning_rate": 4.429730102346874e-07,
      "loss": 0.4686,
      "num_input_tokens_seen": 23625752,
      "step": 36100
    },
    {
      "epoch": 18.92295597484277,
      "grad_norm": 0.3313983082771301,
      "learning_rate": 4.4083263762370873e-07,
      "loss": 0.4813,
      "num_input_tokens_seen": 23629144,
      "step": 36105
    },
    {
      "epoch": 18.92557651991614,
      "grad_norm": 0.6418562531471252,
      "learning_rate": 4.386974025103574e-07,
      "loss": 0.3718,
      "num_input_tokens_seen": 23632440,
      "step": 36110
    },
    {
      "epoch": 18.92819706498952,
      "grad_norm": 0.26245996356010437,
      "learning_rate": 4.365673053413011e-07,
      "loss": 0.3823,
      "num_input_tokens_seen": 23635448,
      "step": 36115
    },
    {
      "epoch": 18.930817610062892,
      "grad_norm": 0.4022831320762634,
      "learning_rate": 4.34442346562125e-07,
      "loss": 0.5813,
      "num_input_tokens_seen": 23639160,
      "step": 36120
    },
    {
      "epoch": 18.93343815513627,
      "grad_norm": 0.3759991526603699,
      "learning_rate": 4.323225266173542e-07,
      "loss": 0.3631,
      "num_input_tokens_seen": 23642904,
      "step": 36125
    },
    {
      "epoch": 18.936058700209642,
      "grad_norm": 0.3448411226272583,
      "learning_rate": 4.302078459504283e-07,
      "loss": 0.2888,
      "num_input_tokens_seen": 23646200,
      "step": 36130
    },
    {
      "epoch": 18.93867924528302,
      "grad_norm": 0.5008806586265564,
      "learning_rate": 4.2809830500371294e-07,
      "loss": 0.4155,
      "num_input_tokens_seen": 23649304,
      "step": 36135
    },
    {
      "epoch": 18.941299790356393,
      "grad_norm": 0.41877615451812744,
      "learning_rate": 4.2599390421850236e-07,
      "loss": 0.5367,
      "num_input_tokens_seen": 23652024,
      "step": 36140
    },
    {
      "epoch": 18.94392033542977,
      "grad_norm": 0.7294613718986511,
      "learning_rate": 4.2389464403501113e-07,
      "loss": 0.419,
      "num_input_tokens_seen": 23655608,
      "step": 36145
    },
    {
      "epoch": 18.946540880503143,
      "grad_norm": 0.4861803352832794,
      "learning_rate": 4.218005248923823e-07,
      "loss": 0.435,
      "num_input_tokens_seen": 23658616,
      "step": 36150
    },
    {
      "epoch": 18.94916142557652,
      "grad_norm": 0.65232914686203,
      "learning_rate": 4.1971154722867946e-07,
      "loss": 0.3863,
      "num_input_tokens_seen": 23661976,
      "step": 36155
    },
    {
      "epoch": 18.951781970649897,
      "grad_norm": 0.6666560769081116,
      "learning_rate": 4.176277114808946e-07,
      "loss": 0.4078,
      "num_input_tokens_seen": 23665080,
      "step": 36160
    },
    {
      "epoch": 18.95440251572327,
      "grad_norm": 0.49495208263397217,
      "learning_rate": 4.155490180849431e-07,
      "loss": 0.5342,
      "num_input_tokens_seen": 23667736,
      "step": 36165
    },
    {
      "epoch": 18.957023060796647,
      "grad_norm": 0.3067017197608948,
      "learning_rate": 4.1347546747566304e-07,
      "loss": 0.3093,
      "num_input_tokens_seen": 23672824,
      "step": 36170
    },
    {
      "epoch": 18.95964360587002,
      "grad_norm": 0.5225083827972412,
      "learning_rate": 4.114070600868131e-07,
      "loss": 0.4392,
      "num_input_tokens_seen": 23676024,
      "step": 36175
    },
    {
      "epoch": 18.962264150943398,
      "grad_norm": 0.5352888703346252,
      "learning_rate": 4.093437963510888e-07,
      "loss": 0.4801,
      "num_input_tokens_seen": 23678936,
      "step": 36180
    },
    {
      "epoch": 18.96488469601677,
      "grad_norm": 0.5749709010124207,
      "learning_rate": 4.072856767000949e-07,
      "loss": 0.5579,
      "num_input_tokens_seen": 23683160,
      "step": 36185
    },
    {
      "epoch": 18.967505241090148,
      "grad_norm": 0.4243354797363281,
      "learning_rate": 4.052327015643703e-07,
      "loss": 0.4122,
      "num_input_tokens_seen": 23685912,
      "step": 36190
    },
    {
      "epoch": 18.97012578616352,
      "grad_norm": 0.44860610365867615,
      "learning_rate": 4.031848713733744e-07,
      "loss": 0.4298,
      "num_input_tokens_seen": 23688536,
      "step": 36195
    },
    {
      "epoch": 18.9727463312369,
      "grad_norm": 0.491026371717453,
      "learning_rate": 4.011421865554837e-07,
      "loss": 0.4092,
      "num_input_tokens_seen": 23691608,
      "step": 36200
    },
    {
      "epoch": 18.97536687631027,
      "grad_norm": 0.37530946731567383,
      "learning_rate": 3.9910464753801224e-07,
      "loss": 0.4185,
      "num_input_tokens_seen": 23694936,
      "step": 36205
    },
    {
      "epoch": 18.97798742138365,
      "grad_norm": 0.4461648762226105,
      "learning_rate": 3.9707225474718557e-07,
      "loss": 0.4577,
      "num_input_tokens_seen": 23698168,
      "step": 36210
    },
    {
      "epoch": 18.980607966457022,
      "grad_norm": 0.5093175172805786,
      "learning_rate": 3.950450086081553e-07,
      "loss": 0.4518,
      "num_input_tokens_seen": 23701080,
      "step": 36215
    },
    {
      "epoch": 18.9832285115304,
      "grad_norm": 0.4171779751777649,
      "learning_rate": 3.930229095450072e-07,
      "loss": 0.3948,
      "num_input_tokens_seen": 23704088,
      "step": 36220
    },
    {
      "epoch": 18.985849056603772,
      "grad_norm": 0.5430624485015869,
      "learning_rate": 3.9100595798073345e-07,
      "loss": 0.5251,
      "num_input_tokens_seen": 23706520,
      "step": 36225
    },
    {
      "epoch": 18.98846960167715,
      "grad_norm": 0.4689739942550659,
      "learning_rate": 3.8899415433725493e-07,
      "loss": 0.4356,
      "num_input_tokens_seen": 23709656,
      "step": 36230
    },
    {
      "epoch": 18.991090146750523,
      "grad_norm": 0.7971360683441162,
      "learning_rate": 3.8698749903542666e-07,
      "loss": 0.4116,
      "num_input_tokens_seen": 23712280,
      "step": 36235
    },
    {
      "epoch": 18.9937106918239,
      "grad_norm": 0.47621873021125793,
      "learning_rate": 3.8498599249501e-07,
      "loss": 0.4047,
      "num_input_tokens_seen": 23715224,
      "step": 36240
    },
    {
      "epoch": 18.996331236897273,
      "grad_norm": 0.5316367149353027,
      "learning_rate": 3.8298963513470345e-07,
      "loss": 0.3177,
      "num_input_tokens_seen": 23719224,
      "step": 36245
    },
    {
      "epoch": 18.99895178197065,
      "grad_norm": 0.47653505206108093,
      "learning_rate": 3.8099842737211734e-07,
      "loss": 0.4458,
      "num_input_tokens_seen": 23721848,
      "step": 36250
    },
    {
      "epoch": 19.0,
      "eval_loss": 0.4471572935581207,
      "eval_runtime": 9.6487,
      "eval_samples_per_second": 87.887,
      "eval_steps_per_second": 21.972,
      "num_input_tokens_seen": 23722376,
      "step": 36252
    },
    {
      "epoch": 19.001572327044027,
      "grad_norm": 0.7641491293907166,
      "learning_rate": 3.7901236962379627e-07,
      "loss": 0.4043,
      "num_input_tokens_seen": 23723848,
      "step": 36255
    },
    {
      "epoch": 19.0041928721174,
      "grad_norm": 0.4411640465259552,
      "learning_rate": 3.7703146230519126e-07,
      "loss": 0.4779,
      "num_input_tokens_seen": 23727912,
      "step": 36260
    },
    {
      "epoch": 19.006813417190777,
      "grad_norm": 0.7670329213142395,
      "learning_rate": 3.7505570583069306e-07,
      "loss": 0.4018,
      "num_input_tokens_seen": 23730696,
      "step": 36265
    },
    {
      "epoch": 19.00943396226415,
      "grad_norm": 0.34827110171318054,
      "learning_rate": 3.7308510061360435e-07,
      "loss": 0.3709,
      "num_input_tokens_seen": 23733960,
      "step": 36270
    },
    {
      "epoch": 19.012054507337528,
      "grad_norm": 0.764362633228302,
      "learning_rate": 3.7111964706615374e-07,
      "loss": 0.4815,
      "num_input_tokens_seen": 23736712,
      "step": 36275
    },
    {
      "epoch": 19.0146750524109,
      "grad_norm": 0.5981764197349548,
      "learning_rate": 3.691593455994929e-07,
      "loss": 0.5737,
      "num_input_tokens_seen": 23739624,
      "step": 36280
    },
    {
      "epoch": 19.017295597484278,
      "grad_norm": 0.6055063605308533,
      "learning_rate": 3.672041966236911e-07,
      "loss": 0.351,
      "num_input_tokens_seen": 23742696,
      "step": 36285
    },
    {
      "epoch": 19.01991614255765,
      "grad_norm": 0.5715456604957581,
      "learning_rate": 3.6525420054774884e-07,
      "loss": 0.5181,
      "num_input_tokens_seen": 23746472,
      "step": 36290
    },
    {
      "epoch": 19.02253668763103,
      "grad_norm": 0.4452049732208252,
      "learning_rate": 3.6330935777957597e-07,
      "loss": 0.3034,
      "num_input_tokens_seen": 23749640,
      "step": 36295
    },
    {
      "epoch": 19.0251572327044,
      "grad_norm": 0.2981938421726227,
      "learning_rate": 3.613696687260193e-07,
      "loss": 0.4816,
      "num_input_tokens_seen": 23752328,
      "step": 36300
    },
    {
      "epoch": 19.02777777777778,
      "grad_norm": 0.3768625259399414,
      "learning_rate": 3.59435133792832e-07,
      "loss": 0.4066,
      "num_input_tokens_seen": 23756392,
      "step": 36305
    },
    {
      "epoch": 19.030398322851152,
      "grad_norm": 0.510184109210968,
      "learning_rate": 3.575057533847043e-07,
      "loss": 0.5069,
      "num_input_tokens_seen": 23759784,
      "step": 36310
    },
    {
      "epoch": 19.03301886792453,
      "grad_norm": 0.3425469398498535,
      "learning_rate": 3.555815279052327e-07,
      "loss": 0.4432,
      "num_input_tokens_seen": 23762408,
      "step": 36315
    },
    {
      "epoch": 19.035639412997902,
      "grad_norm": 0.28444579243659973,
      "learning_rate": 3.5366245775694805e-07,
      "loss": 0.4669,
      "num_input_tokens_seen": 23765960,
      "step": 36320
    },
    {
      "epoch": 19.03825995807128,
      "grad_norm": 0.5660430788993835,
      "learning_rate": 3.517485433412987e-07,
      "loss": 0.4987,
      "num_input_tokens_seen": 23769032,
      "step": 36325
    },
    {
      "epoch": 19.040880503144653,
      "grad_norm": 0.5652046799659729,
      "learning_rate": 3.4983978505865044e-07,
      "loss": 0.3657,
      "num_input_tokens_seen": 23771528,
      "step": 36330
    },
    {
      "epoch": 19.04350104821803,
      "grad_norm": 0.5436833500862122,
      "learning_rate": 3.47936183308295e-07,
      "loss": 0.5117,
      "num_input_tokens_seen": 23775400,
      "step": 36335
    },
    {
      "epoch": 19.046121593291403,
      "grad_norm": 0.6491172909736633,
      "learning_rate": 3.460377384884472e-07,
      "loss": 0.4078,
      "num_input_tokens_seen": 23778536,
      "step": 36340
    },
    {
      "epoch": 19.04874213836478,
      "grad_norm": 0.3663213849067688,
      "learning_rate": 3.4414445099623384e-07,
      "loss": 0.4321,
      "num_input_tokens_seen": 23781832,
      "step": 36345
    },
    {
      "epoch": 19.051362683438157,
      "grad_norm": 0.40992140769958496,
      "learning_rate": 3.42256321227713e-07,
      "loss": 0.4818,
      "num_input_tokens_seen": 23789288,
      "step": 36350
    },
    {
      "epoch": 19.05398322851153,
      "grad_norm": 0.3722268342971802,
      "learning_rate": 3.403733495778605e-07,
      "loss": 0.4478,
      "num_input_tokens_seen": 23792296,
      "step": 36355
    },
    {
      "epoch": 19.056603773584907,
      "grad_norm": 0.29100972414016724,
      "learning_rate": 3.3849553644056955e-07,
      "loss": 0.4277,
      "num_input_tokens_seen": 23795464,
      "step": 36360
    },
    {
      "epoch": 19.05922431865828,
      "grad_norm": 0.5232933163642883,
      "learning_rate": 3.36622882208662e-07,
      "loss": 0.5358,
      "num_input_tokens_seen": 23798504,
      "step": 36365
    },
    {
      "epoch": 19.061844863731658,
      "grad_norm": 0.4990130364894867,
      "learning_rate": 3.347553872738718e-07,
      "loss": 0.4828,
      "num_input_tokens_seen": 23802024,
      "step": 36370
    },
    {
      "epoch": 19.06446540880503,
      "grad_norm": 0.4455731511116028,
      "learning_rate": 3.3289305202685583e-07,
      "loss": 0.3309,
      "num_input_tokens_seen": 23805512,
      "step": 36375
    },
    {
      "epoch": 19.067085953878408,
      "grad_norm": 0.4996175765991211,
      "learning_rate": 3.310358768571997e-07,
      "loss": 0.3701,
      "num_input_tokens_seen": 23808712,
      "step": 36380
    },
    {
      "epoch": 19.06970649895178,
      "grad_norm": 0.5195133090019226,
      "learning_rate": 3.2918386215339827e-07,
      "loss": 0.3681,
      "num_input_tokens_seen": 23814280,
      "step": 36385
    },
    {
      "epoch": 19.072327044025158,
      "grad_norm": 0.8077579736709595,
      "learning_rate": 3.2733700830287205e-07,
      "loss": 0.4143,
      "num_input_tokens_seen": 23817448,
      "step": 36390
    },
    {
      "epoch": 19.07494758909853,
      "grad_norm": 0.5049005746841431,
      "learning_rate": 3.254953156919649e-07,
      "loss": 0.5189,
      "num_input_tokens_seen": 23820520,
      "step": 36395
    },
    {
      "epoch": 19.07756813417191,
      "grad_norm": 0.11367081850767136,
      "learning_rate": 3.2365878470593527e-07,
      "loss": 0.3122,
      "num_input_tokens_seen": 23826664,
      "step": 36400
    },
    {
      "epoch": 19.080188679245282,
      "grad_norm": 0.3073461055755615,
      "learning_rate": 3.2182741572896747e-07,
      "loss": 0.3736,
      "num_input_tokens_seen": 23830120,
      "step": 36405
    },
    {
      "epoch": 19.08280922431866,
      "grad_norm": 0.5514822006225586,
      "learning_rate": 3.2000120914415786e-07,
      "loss": 0.5754,
      "num_input_tokens_seen": 23833480,
      "step": 36410
    },
    {
      "epoch": 19.085429769392032,
      "grad_norm": 0.5346662402153015,
      "learning_rate": 3.181801653335315e-07,
      "loss": 0.5042,
      "num_input_tokens_seen": 23836712,
      "step": 36415
    },
    {
      "epoch": 19.08805031446541,
      "grad_norm": 0.8093337416648865,
      "learning_rate": 3.163642846780307e-07,
      "loss": 0.593,
      "num_input_tokens_seen": 23840072,
      "step": 36420
    },
    {
      "epoch": 19.090670859538783,
      "grad_norm": 0.4039062261581421,
      "learning_rate": 3.1455356755751565e-07,
      "loss": 0.4784,
      "num_input_tokens_seen": 23843176,
      "step": 36425
    },
    {
      "epoch": 19.09329140461216,
      "grad_norm": 0.2561565041542053,
      "learning_rate": 3.127480143507666e-07,
      "loss": 0.3998,
      "num_input_tokens_seen": 23846344,
      "step": 36430
    },
    {
      "epoch": 19.095911949685533,
      "grad_norm": 0.46588101983070374,
      "learning_rate": 3.1094762543548704e-07,
      "loss": 0.4039,
      "num_input_tokens_seen": 23850216,
      "step": 36435
    },
    {
      "epoch": 19.09853249475891,
      "grad_norm": 1.2455222606658936,
      "learning_rate": 3.091524011882951e-07,
      "loss": 0.583,
      "num_input_tokens_seen": 23853352,
      "step": 36440
    },
    {
      "epoch": 19.101153039832287,
      "grad_norm": 0.37909701466560364,
      "learning_rate": 3.073623419847349e-07,
      "loss": 0.4584,
      "num_input_tokens_seen": 23855912,
      "step": 36445
    },
    {
      "epoch": 19.10377358490566,
      "grad_norm": 0.9304777979850769,
      "learning_rate": 3.0557744819926235e-07,
      "loss": 0.3065,
      "num_input_tokens_seen": 23858120,
      "step": 36450
    },
    {
      "epoch": 19.106394129979037,
      "grad_norm": 0.455859899520874,
      "learning_rate": 3.037977202052594e-07,
      "loss": 0.44,
      "num_input_tokens_seen": 23861736,
      "step": 36455
    },
    {
      "epoch": 19.10901467505241,
      "grad_norm": 0.46168872714042664,
      "learning_rate": 3.020231583750255e-07,
      "loss": 0.3193,
      "num_input_tokens_seen": 23865864,
      "step": 36460
    },
    {
      "epoch": 19.111635220125788,
      "grad_norm": 0.7506423592567444,
      "learning_rate": 3.002537630797747e-07,
      "loss": 0.4242,
      "num_input_tokens_seen": 23869160,
      "step": 36465
    },
    {
      "epoch": 19.11425576519916,
      "grad_norm": 0.4280827045440674,
      "learning_rate": 2.984895346896499e-07,
      "loss": 0.3886,
      "num_input_tokens_seen": 23873064,
      "step": 36470
    },
    {
      "epoch": 19.116876310272538,
      "grad_norm": 0.36114701628685,
      "learning_rate": 2.967304735737031e-07,
      "loss": 0.4226,
      "num_input_tokens_seen": 23876328,
      "step": 36475
    },
    {
      "epoch": 19.11949685534591,
      "grad_norm": 0.300685852766037,
      "learning_rate": 2.9497658009991214e-07,
      "loss": 0.4326,
      "num_input_tokens_seen": 23878984,
      "step": 36480
    },
    {
      "epoch": 19.122117400419288,
      "grad_norm": 0.396723210811615,
      "learning_rate": 2.932278546351724e-07,
      "loss": 0.2984,
      "num_input_tokens_seen": 23881800,
      "step": 36485
    },
    {
      "epoch": 19.12473794549266,
      "grad_norm": 0.5150547623634338,
      "learning_rate": 2.9148429754529137e-07,
      "loss": 0.4205,
      "num_input_tokens_seen": 23884296,
      "step": 36490
    },
    {
      "epoch": 19.12735849056604,
      "grad_norm": 0.8811467885971069,
      "learning_rate": 2.897459091950078e-07,
      "loss": 0.6868,
      "num_input_tokens_seen": 23887560,
      "step": 36495
    },
    {
      "epoch": 19.129979035639412,
      "grad_norm": 0.41868799924850464,
      "learning_rate": 2.880126899479724e-07,
      "loss": 0.3586,
      "num_input_tokens_seen": 23890408,
      "step": 36500
    },
    {
      "epoch": 19.13259958071279,
      "grad_norm": 0.6209421753883362,
      "learning_rate": 2.862846401667507e-07,
      "loss": 0.4751,
      "num_input_tokens_seen": 23893288,
      "step": 36505
    },
    {
      "epoch": 19.135220125786162,
      "grad_norm": 0.722720742225647,
      "learning_rate": 2.8456176021283697e-07,
      "loss": 0.6397,
      "num_input_tokens_seen": 23897000,
      "step": 36510
    },
    {
      "epoch": 19.13784067085954,
      "grad_norm": 0.5617395043373108,
      "learning_rate": 2.8284405044662897e-07,
      "loss": 0.4555,
      "num_input_tokens_seen": 23900392,
      "step": 36515
    },
    {
      "epoch": 19.140461215932913,
      "grad_norm": 0.3373260200023651,
      "learning_rate": 2.8113151122745874e-07,
      "loss": 0.344,
      "num_input_tokens_seen": 23904520,
      "step": 36520
    },
    {
      "epoch": 19.14308176100629,
      "grad_norm": 0.6567053198814392,
      "learning_rate": 2.7942414291357023e-07,
      "loss": 0.4936,
      "num_input_tokens_seen": 23907560,
      "step": 36525
    },
    {
      "epoch": 19.145702306079663,
      "grad_norm": 0.4380923807621002,
      "learning_rate": 2.777219458621222e-07,
      "loss": 0.4782,
      "num_input_tokens_seen": 23911784,
      "step": 36530
    },
    {
      "epoch": 19.14832285115304,
      "grad_norm": 0.3354288637638092,
      "learning_rate": 2.7602492042919923e-07,
      "loss": 0.4234,
      "num_input_tokens_seen": 23914952,
      "step": 36535
    },
    {
      "epoch": 19.150943396226417,
      "grad_norm": 0.5246031284332275,
      "learning_rate": 2.7433306696979523e-07,
      "loss": 0.48,
      "num_input_tokens_seen": 23918408,
      "step": 36540
    },
    {
      "epoch": 19.15356394129979,
      "grad_norm": 0.5666950345039368,
      "learning_rate": 2.7264638583782973e-07,
      "loss": 0.4752,
      "num_input_tokens_seen": 23921896,
      "step": 36545
    },
    {
      "epoch": 19.156184486373167,
      "grad_norm": 0.3769274950027466,
      "learning_rate": 2.7096487738613177e-07,
      "loss": 0.4329,
      "num_input_tokens_seen": 23925512,
      "step": 36550
    },
    {
      "epoch": 19.15880503144654,
      "grad_norm": 0.42399096488952637,
      "learning_rate": 2.6928854196646147e-07,
      "loss": 0.4388,
      "num_input_tokens_seen": 23928072,
      "step": 36555
    },
    {
      "epoch": 19.161425576519918,
      "grad_norm": 0.5314890146255493,
      "learning_rate": 2.6761737992948565e-07,
      "loss": 0.4229,
      "num_input_tokens_seen": 23930632,
      "step": 36560
    },
    {
      "epoch": 19.16404612159329,
      "grad_norm": 0.3806911110877991,
      "learning_rate": 2.6595139162479134e-07,
      "loss": 0.394,
      "num_input_tokens_seen": 23934408,
      "step": 36565
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.3888084590435028,
      "learning_rate": 2.6429057740088303e-07,
      "loss": 0.2901,
      "num_input_tokens_seen": 23937768,
      "step": 36570
    },
    {
      "epoch": 19.16928721174004,
      "grad_norm": 0.40768271684646606,
      "learning_rate": 2.6263493760518564e-07,
      "loss": 0.3385,
      "num_input_tokens_seen": 23941288,
      "step": 36575
    },
    {
      "epoch": 19.171907756813418,
      "grad_norm": 0.4787967801094055,
      "learning_rate": 2.6098447258404157e-07,
      "loss": 0.3607,
      "num_input_tokens_seen": 23945384,
      "step": 36580
    },
    {
      "epoch": 19.17452830188679,
      "grad_norm": 0.38877618312835693,
      "learning_rate": 2.593391826827107e-07,
      "loss": 0.3654,
      "num_input_tokens_seen": 23948680,
      "step": 36585
    },
    {
      "epoch": 19.17714884696017,
      "grad_norm": 1.0851346254348755,
      "learning_rate": 2.5769906824536225e-07,
      "loss": 0.4472,
      "num_input_tokens_seen": 23951912,
      "step": 36590
    },
    {
      "epoch": 19.179769392033542,
      "grad_norm": 0.4876876175403595,
      "learning_rate": 2.560641296150967e-07,
      "loss": 0.4536,
      "num_input_tokens_seen": 23954568,
      "step": 36595
    },
    {
      "epoch": 19.18238993710692,
      "grad_norm": 0.3732403814792633,
      "learning_rate": 2.544343671339211e-07,
      "loss": 0.6004,
      "num_input_tokens_seen": 23958376,
      "step": 36600
    },
    {
      "epoch": 19.185010482180292,
      "grad_norm": 0.5087453722953796,
      "learning_rate": 2.5280978114276264e-07,
      "loss": 0.5081,
      "num_input_tokens_seen": 23961768,
      "step": 36605
    },
    {
      "epoch": 19.18763102725367,
      "grad_norm": 0.48950982093811035,
      "learning_rate": 2.511903719814662e-07,
      "loss": 0.5018,
      "num_input_tokens_seen": 23964936,
      "step": 36610
    },
    {
      "epoch": 19.190251572327043,
      "grad_norm": 0.4997764825820923,
      "learning_rate": 2.4957613998879695e-07,
      "loss": 0.5151,
      "num_input_tokens_seen": 23967976,
      "step": 36615
    },
    {
      "epoch": 19.19287211740042,
      "grad_norm": 0.751437246799469,
      "learning_rate": 2.479670855024291e-07,
      "loss": 0.4479,
      "num_input_tokens_seen": 23970984,
      "step": 36620
    },
    {
      "epoch": 19.195492662473793,
      "grad_norm": 0.7008517384529114,
      "learning_rate": 2.463632088589629e-07,
      "loss": 0.3801,
      "num_input_tokens_seen": 23975400,
      "step": 36625
    },
    {
      "epoch": 19.19811320754717,
      "grad_norm": 0.7370015382766724,
      "learning_rate": 2.447645103939078e-07,
      "loss": 0.5183,
      "num_input_tokens_seen": 23978504,
      "step": 36630
    },
    {
      "epoch": 19.200733752620547,
      "grad_norm": 0.7851257920265198,
      "learning_rate": 2.431709904416962e-07,
      "loss": 0.3687,
      "num_input_tokens_seen": 23981992,
      "step": 36635
    },
    {
      "epoch": 19.20335429769392,
      "grad_norm": 0.4397372603416443,
      "learning_rate": 2.4158264933567264e-07,
      "loss": 0.4337,
      "num_input_tokens_seen": 23985416,
      "step": 36640
    },
    {
      "epoch": 19.205974842767297,
      "grad_norm": 0.6581450700759888,
      "learning_rate": 2.399994874080963e-07,
      "loss": 0.4726,
      "num_input_tokens_seen": 23988328,
      "step": 36645
    },
    {
      "epoch": 19.20859538784067,
      "grad_norm": 0.5084232091903687,
      "learning_rate": 2.3842150499015504e-07,
      "loss": 0.5188,
      "num_input_tokens_seen": 23991464,
      "step": 36650
    },
    {
      "epoch": 19.211215932914047,
      "grad_norm": 0.5460798740386963,
      "learning_rate": 2.368487024119348e-07,
      "loss": 0.4443,
      "num_input_tokens_seen": 23994536,
      "step": 36655
    },
    {
      "epoch": 19.21383647798742,
      "grad_norm": 0.6951980590820312,
      "learning_rate": 2.3528108000245573e-07,
      "loss": 0.3677,
      "num_input_tokens_seen": 23998472,
      "step": 36660
    },
    {
      "epoch": 19.216457023060798,
      "grad_norm": 0.4124065041542053,
      "learning_rate": 2.3371863808964445e-07,
      "loss": 0.4309,
      "num_input_tokens_seen": 24001384,
      "step": 36665
    },
    {
      "epoch": 19.21907756813417,
      "grad_norm": 0.4708881378173828,
      "learning_rate": 2.32161377000345e-07,
      "loss": 0.5052,
      "num_input_tokens_seen": 24004712,
      "step": 36670
    },
    {
      "epoch": 19.221698113207548,
      "grad_norm": 0.671906054019928,
      "learning_rate": 2.306092970603163e-07,
      "loss": 0.4156,
      "num_input_tokens_seen": 24007144,
      "step": 36675
    },
    {
      "epoch": 19.22431865828092,
      "grad_norm": 0.864493191242218,
      "learning_rate": 2.2906239859424306e-07,
      "loss": 0.4177,
      "num_input_tokens_seen": 24010056,
      "step": 36680
    },
    {
      "epoch": 19.2269392033543,
      "grad_norm": 0.4258370101451874,
      "learning_rate": 2.2752068192571084e-07,
      "loss": 0.3779,
      "num_input_tokens_seen": 24013352,
      "step": 36685
    },
    {
      "epoch": 19.229559748427672,
      "grad_norm": 0.5903211236000061,
      "learning_rate": 2.2598414737723396e-07,
      "loss": 0.4099,
      "num_input_tokens_seen": 24016488,
      "step": 36690
    },
    {
      "epoch": 19.23218029350105,
      "grad_norm": 0.5016258358955383,
      "learning_rate": 2.2445279527023578e-07,
      "loss": 0.6167,
      "num_input_tokens_seen": 24020008,
      "step": 36695
    },
    {
      "epoch": 19.234800838574422,
      "grad_norm": 0.5416702628135681,
      "learning_rate": 2.2292662592506007e-07,
      "loss": 0.474,
      "num_input_tokens_seen": 24022472,
      "step": 36700
    },
    {
      "epoch": 19.2374213836478,
      "grad_norm": 0.3762895464897156,
      "learning_rate": 2.2140563966095985e-07,
      "loss": 0.5194,
      "num_input_tokens_seen": 24026856,
      "step": 36705
    },
    {
      "epoch": 19.240041928721173,
      "grad_norm": 0.4185940623283386,
      "learning_rate": 2.1988983679611386e-07,
      "loss": 0.4516,
      "num_input_tokens_seen": 24030376,
      "step": 36710
    },
    {
      "epoch": 19.24266247379455,
      "grad_norm": 0.524783194065094,
      "learning_rate": 2.1837921764760462e-07,
      "loss": 0.4088,
      "num_input_tokens_seen": 24033384,
      "step": 36715
    },
    {
      "epoch": 19.245283018867923,
      "grad_norm": 0.5972617864608765,
      "learning_rate": 2.1687378253144043e-07,
      "loss": 0.3498,
      "num_input_tokens_seen": 24039144,
      "step": 36720
    },
    {
      "epoch": 19.2479035639413,
      "grad_norm": 0.5682650208473206,
      "learning_rate": 2.153735317625416e-07,
      "loss": 0.3779,
      "num_input_tokens_seen": 24043112,
      "step": 36725
    },
    {
      "epoch": 19.250524109014677,
      "grad_norm": 0.4835644066333771,
      "learning_rate": 2.1387846565474045e-07,
      "loss": 0.5069,
      "num_input_tokens_seen": 24046824,
      "step": 36730
    },
    {
      "epoch": 19.25314465408805,
      "grad_norm": 0.7470293641090393,
      "learning_rate": 2.123885845207896e-07,
      "loss": 0.5061,
      "num_input_tokens_seen": 24049576,
      "step": 36735
    },
    {
      "epoch": 19.255765199161427,
      "grad_norm": 0.2832844853401184,
      "learning_rate": 2.1090388867235356e-07,
      "loss": 0.5164,
      "num_input_tokens_seen": 24052456,
      "step": 36740
    },
    {
      "epoch": 19.2583857442348,
      "grad_norm": 1.0538091659545898,
      "learning_rate": 2.0942437842001728e-07,
      "loss": 0.3488,
      "num_input_tokens_seen": 24055272,
      "step": 36745
    },
    {
      "epoch": 19.261006289308177,
      "grad_norm": 0.44833827018737793,
      "learning_rate": 2.0795005407327483e-07,
      "loss": 0.3899,
      "num_input_tokens_seen": 24058728,
      "step": 36750
    },
    {
      "epoch": 19.26362683438155,
      "grad_norm": 0.40221354365348816,
      "learning_rate": 2.0648091594053508e-07,
      "loss": 0.4602,
      "num_input_tokens_seen": 24061896,
      "step": 36755
    },
    {
      "epoch": 19.266247379454928,
      "grad_norm": 0.7165574431419373,
      "learning_rate": 2.0501696432913275e-07,
      "loss": 0.5511,
      "num_input_tokens_seen": 24065000,
      "step": 36760
    },
    {
      "epoch": 19.2688679245283,
      "grad_norm": 0.49239540100097656,
      "learning_rate": 2.035581995453034e-07,
      "loss": 0.4307,
      "num_input_tokens_seen": 24068264,
      "step": 36765
    },
    {
      "epoch": 19.271488469601678,
      "grad_norm": 0.6603814959526062,
      "learning_rate": 2.0210462189420577e-07,
      "loss": 0.418,
      "num_input_tokens_seen": 24072008,
      "step": 36770
    },
    {
      "epoch": 19.27410901467505,
      "grad_norm": 0.2868794798851013,
      "learning_rate": 2.006562316799132e-07,
      "loss": 0.4425,
      "num_input_tokens_seen": 24075752,
      "step": 36775
    },
    {
      "epoch": 19.27672955974843,
      "grad_norm": 0.4777512848377228,
      "learning_rate": 1.992130292054084e-07,
      "loss": 0.3819,
      "num_input_tokens_seen": 24079528,
      "step": 36780
    },
    {
      "epoch": 19.279350104821802,
      "grad_norm": 0.3026517331600189,
      "learning_rate": 1.9777501477259708e-07,
      "loss": 0.3917,
      "num_input_tokens_seen": 24085544,
      "step": 36785
    },
    {
      "epoch": 19.28197064989518,
      "grad_norm": 0.7588669061660767,
      "learning_rate": 1.9634218868229414e-07,
      "loss": 0.4577,
      "num_input_tokens_seen": 24089480,
      "step": 36790
    },
    {
      "epoch": 19.284591194968552,
      "grad_norm": 0.4374316334724426,
      "learning_rate": 1.9491455123423207e-07,
      "loss": 0.3561,
      "num_input_tokens_seen": 24092968,
      "step": 36795
    },
    {
      "epoch": 19.28721174004193,
      "grad_norm": 0.31410840153694153,
      "learning_rate": 1.9349210272705253e-07,
      "loss": 0.3588,
      "num_input_tokens_seen": 24096680,
      "step": 36800
    },
    {
      "epoch": 19.289832285115303,
      "grad_norm": 0.5200327634811401,
      "learning_rate": 1.920748434583175e-07,
      "loss": 0.4173,
      "num_input_tokens_seen": 24100456,
      "step": 36805
    },
    {
      "epoch": 19.29245283018868,
      "grad_norm": 0.5407039523124695,
      "learning_rate": 1.906627737245009e-07,
      "loss": 0.4683,
      "num_input_tokens_seen": 24103816,
      "step": 36810
    },
    {
      "epoch": 19.295073375262053,
      "grad_norm": 0.4816030263900757,
      "learning_rate": 1.8925589382099152e-07,
      "loss": 0.4196,
      "num_input_tokens_seen": 24107752,
      "step": 36815
    },
    {
      "epoch": 19.29769392033543,
      "grad_norm": 0.6015945076942444,
      "learning_rate": 1.8785420404209553e-07,
      "loss": 0.4548,
      "num_input_tokens_seen": 24110856,
      "step": 36820
    },
    {
      "epoch": 19.300314465408803,
      "grad_norm": 0.9923291802406311,
      "learning_rate": 1.864577046810284e-07,
      "loss": 0.557,
      "num_input_tokens_seen": 24113896,
      "step": 36825
    },
    {
      "epoch": 19.30293501048218,
      "grad_norm": 0.48854175209999084,
      "learning_rate": 1.8506639602992038e-07,
      "loss": 0.4757,
      "num_input_tokens_seen": 24116872,
      "step": 36830
    },
    {
      "epoch": 19.305555555555557,
      "grad_norm": 0.2978104054927826,
      "learning_rate": 1.836802783798164e-07,
      "loss": 0.3723,
      "num_input_tokens_seen": 24120072,
      "step": 36835
    },
    {
      "epoch": 19.30817610062893,
      "grad_norm": 0.34662047028541565,
      "learning_rate": 1.822993520206817e-07,
      "loss": 0.3635,
      "num_input_tokens_seen": 24124232,
      "step": 36840
    },
    {
      "epoch": 19.310796645702307,
      "grad_norm": 0.4762248396873474,
      "learning_rate": 1.8092361724138528e-07,
      "loss": 0.3842,
      "num_input_tokens_seen": 24127720,
      "step": 36845
    },
    {
      "epoch": 19.31341719077568,
      "grad_norm": 0.5890054106712341,
      "learning_rate": 1.7955307432971914e-07,
      "loss": 0.4556,
      "num_input_tokens_seen": 24131624,
      "step": 36850
    },
    {
      "epoch": 19.316037735849058,
      "grad_norm": 0.5497915148735046,
      "learning_rate": 1.7818772357238445e-07,
      "loss": 0.3665,
      "num_input_tokens_seen": 24134568,
      "step": 36855
    },
    {
      "epoch": 19.31865828092243,
      "grad_norm": 0.9310927987098694,
      "learning_rate": 1.7682756525499166e-07,
      "loss": 0.3528,
      "num_input_tokens_seen": 24137768,
      "step": 36860
    },
    {
      "epoch": 19.321278825995808,
      "grad_norm": 1.3284834623336792,
      "learning_rate": 1.7547259966207708e-07,
      "loss": 0.4692,
      "num_input_tokens_seen": 24142184,
      "step": 36865
    },
    {
      "epoch": 19.32389937106918,
      "grad_norm": 0.4699355661869049,
      "learning_rate": 1.7412282707708062e-07,
      "loss": 0.425,
      "num_input_tokens_seen": 24145192,
      "step": 36870
    },
    {
      "epoch": 19.32651991614256,
      "grad_norm": 0.40170663595199585,
      "learning_rate": 1.7277824778235974e-07,
      "loss": 0.4383,
      "num_input_tokens_seen": 24148648,
      "step": 36875
    },
    {
      "epoch": 19.329140461215932,
      "grad_norm": 0.23371554911136627,
      "learning_rate": 1.714388620591867e-07,
      "loss": 0.3424,
      "num_input_tokens_seen": 24152616,
      "step": 36880
    },
    {
      "epoch": 19.33176100628931,
      "grad_norm": 0.5405797362327576,
      "learning_rate": 1.701046701877429e-07,
      "loss": 0.5412,
      "num_input_tokens_seen": 24155272,
      "step": 36885
    },
    {
      "epoch": 19.334381551362682,
      "grad_norm": 0.27982139587402344,
      "learning_rate": 1.6877567244713011e-07,
      "loss": 0.4483,
      "num_input_tokens_seen": 24158536,
      "step": 36890
    },
    {
      "epoch": 19.33700209643606,
      "grad_norm": 0.24612703919410706,
      "learning_rate": 1.6745186911535372e-07,
      "loss": 0.4759,
      "num_input_tokens_seen": 24163144,
      "step": 36895
    },
    {
      "epoch": 19.339622641509433,
      "grad_norm": 0.4932178556919098,
      "learning_rate": 1.6613326046934497e-07,
      "loss": 0.3978,
      "num_input_tokens_seen": 24165864,
      "step": 36900
    },
    {
      "epoch": 19.34224318658281,
      "grad_norm": 0.3812399208545685,
      "learning_rate": 1.64819846784936e-07,
      "loss": 0.4274,
      "num_input_tokens_seen": 24169032,
      "step": 36905
    },
    {
      "epoch": 19.344863731656183,
      "grad_norm": 0.7087461352348328,
      "learning_rate": 1.6351162833688204e-07,
      "loss": 0.4951,
      "num_input_tokens_seen": 24171592,
      "step": 36910
    },
    {
      "epoch": 19.34748427672956,
      "grad_norm": 0.43684855103492737,
      "learning_rate": 1.6220860539884475e-07,
      "loss": 0.4703,
      "num_input_tokens_seen": 24174824,
      "step": 36915
    },
    {
      "epoch": 19.350104821802937,
      "grad_norm": 0.5053076148033142,
      "learning_rate": 1.6091077824340327e-07,
      "loss": 0.568,
      "num_input_tokens_seen": 24177864,
      "step": 36920
    },
    {
      "epoch": 19.35272536687631,
      "grad_norm": 0.4132927656173706,
      "learning_rate": 1.5961814714204605e-07,
      "loss": 0.4183,
      "num_input_tokens_seen": 24181288,
      "step": 36925
    },
    {
      "epoch": 19.355345911949687,
      "grad_norm": 0.4989885091781616,
      "learning_rate": 1.5833071236518172e-07,
      "loss": 0.5874,
      "num_input_tokens_seen": 24184680,
      "step": 36930
    },
    {
      "epoch": 19.35796645702306,
      "grad_norm": 0.5000324249267578,
      "learning_rate": 1.5704847418212266e-07,
      "loss": 0.3281,
      "num_input_tokens_seen": 24188072,
      "step": 36935
    },
    {
      "epoch": 19.360587002096437,
      "grad_norm": 0.6379866003990173,
      "learning_rate": 1.5577143286109875e-07,
      "loss": 0.4025,
      "num_input_tokens_seen": 24190664,
      "step": 36940
    },
    {
      "epoch": 19.36320754716981,
      "grad_norm": 0.3476327657699585,
      "learning_rate": 1.5449958866925185e-07,
      "loss": 0.2532,
      "num_input_tokens_seen": 24193832,
      "step": 36945
    },
    {
      "epoch": 19.365828092243188,
      "grad_norm": 0.6093940734863281,
      "learning_rate": 1.532329418726386e-07,
      "loss": 0.4474,
      "num_input_tokens_seen": 24197256,
      "step": 36950
    },
    {
      "epoch": 19.36844863731656,
      "grad_norm": 0.6028652191162109,
      "learning_rate": 1.5197149273623036e-07,
      "loss": 0.3701,
      "num_input_tokens_seen": 24199816,
      "step": 36955
    },
    {
      "epoch": 19.371069182389938,
      "grad_norm": 0.41828200221061707,
      "learning_rate": 1.5071524152389936e-07,
      "loss": 0.5197,
      "num_input_tokens_seen": 24203272,
      "step": 36960
    },
    {
      "epoch": 19.37368972746331,
      "grad_norm": 0.18041707575321198,
      "learning_rate": 1.4946418849844934e-07,
      "loss": 0.358,
      "num_input_tokens_seen": 24206728,
      "step": 36965
    },
    {
      "epoch": 19.37631027253669,
      "grad_norm": 0.6393333673477173,
      "learning_rate": 1.4821833392157923e-07,
      "loss": 0.3606,
      "num_input_tokens_seen": 24209416,
      "step": 36970
    },
    {
      "epoch": 19.378930817610062,
      "grad_norm": 0.6128814816474915,
      "learning_rate": 1.4697767805390838e-07,
      "loss": 0.4829,
      "num_input_tokens_seen": 24212456,
      "step": 36975
    },
    {
      "epoch": 19.38155136268344,
      "grad_norm": 0.8123725652694702,
      "learning_rate": 1.4574222115497084e-07,
      "loss": 0.4714,
      "num_input_tokens_seen": 24214856,
      "step": 36980
    },
    {
      "epoch": 19.384171907756812,
      "grad_norm": 0.4334161877632141,
      "learning_rate": 1.4451196348320706e-07,
      "loss": 0.4125,
      "num_input_tokens_seen": 24218024,
      "step": 36985
    },
    {
      "epoch": 19.38679245283019,
      "grad_norm": 0.5951847434043884,
      "learning_rate": 1.4328690529597232e-07,
      "loss": 0.4245,
      "num_input_tokens_seen": 24220808,
      "step": 36990
    },
    {
      "epoch": 19.389412997903563,
      "grad_norm": 0.3020763397216797,
      "learning_rate": 1.4206704684953943e-07,
      "loss": 0.5759,
      "num_input_tokens_seen": 24223816,
      "step": 36995
    },
    {
      "epoch": 19.39203354297694,
      "grad_norm": 0.27085164189338684,
      "learning_rate": 1.4085238839908476e-07,
      "loss": 0.4207,
      "num_input_tokens_seen": 24227464,
      "step": 37000
    },
    {
      "epoch": 19.394654088050313,
      "grad_norm": 0.29632458090782166,
      "learning_rate": 1.3964293019870235e-07,
      "loss": 0.5092,
      "num_input_tokens_seen": 24230376,
      "step": 37005
    },
    {
      "epoch": 19.39727463312369,
      "grad_norm": 0.5464555621147156,
      "learning_rate": 1.3843867250139807e-07,
      "loss": 0.4,
      "num_input_tokens_seen": 24232744,
      "step": 37010
    },
    {
      "epoch": 19.399895178197063,
      "grad_norm": 0.42810121178627014,
      "learning_rate": 1.3723961555908715e-07,
      "loss": 0.428,
      "num_input_tokens_seen": 24235848,
      "step": 37015
    },
    {
      "epoch": 19.40251572327044,
      "grad_norm": 0.5566081404685974,
      "learning_rate": 1.3604575962259947e-07,
      "loss": 0.4388,
      "num_input_tokens_seen": 24239912,
      "step": 37020
    },
    {
      "epoch": 19.405136268343817,
      "grad_norm": 0.3403046131134033,
      "learning_rate": 1.3485710494167692e-07,
      "loss": 0.6207,
      "num_input_tokens_seen": 24243784,
      "step": 37025
    },
    {
      "epoch": 19.40775681341719,
      "grad_norm": 0.5192477703094482,
      "learning_rate": 1.3367365176496783e-07,
      "loss": 0.3682,
      "num_input_tokens_seen": 24246792,
      "step": 37030
    },
    {
      "epoch": 19.410377358490567,
      "grad_norm": 0.3319516181945801,
      "learning_rate": 1.324954003400436e-07,
      "loss": 0.3644,
      "num_input_tokens_seen": 24249704,
      "step": 37035
    },
    {
      "epoch": 19.41299790356394,
      "grad_norm": 0.42728596925735474,
      "learning_rate": 1.313223509133793e-07,
      "loss": 0.4004,
      "num_input_tokens_seen": 24252616,
      "step": 37040
    },
    {
      "epoch": 19.415618448637318,
      "grad_norm": 0.48222899436950684,
      "learning_rate": 1.301545037303592e-07,
      "loss": 0.3728,
      "num_input_tokens_seen": 24256136,
      "step": 37045
    },
    {
      "epoch": 19.41823899371069,
      "grad_norm": 0.42887869477272034,
      "learning_rate": 1.2899185903529066e-07,
      "loss": 0.4362,
      "num_input_tokens_seen": 24259400,
      "step": 37050
    },
    {
      "epoch": 19.420859538784068,
      "grad_norm": 0.2684969902038574,
      "learning_rate": 1.278344170713791e-07,
      "loss": 0.3691,
      "num_input_tokens_seen": 24263880,
      "step": 37055
    },
    {
      "epoch": 19.42348008385744,
      "grad_norm": 0.5037758350372314,
      "learning_rate": 1.2668217808075033e-07,
      "loss": 0.4592,
      "num_input_tokens_seen": 24266888,
      "step": 37060
    },
    {
      "epoch": 19.42610062893082,
      "grad_norm": 0.6915864944458008,
      "learning_rate": 1.255351423044393e-07,
      "loss": 0.4713,
      "num_input_tokens_seen": 24269768,
      "step": 37065
    },
    {
      "epoch": 19.428721174004192,
      "grad_norm": 0.48273351788520813,
      "learning_rate": 1.2439330998239572e-07,
      "loss": 0.3605,
      "num_input_tokens_seen": 24272808,
      "step": 37070
    },
    {
      "epoch": 19.43134171907757,
      "grad_norm": 0.4518774449825287,
      "learning_rate": 1.2325668135347578e-07,
      "loss": 0.379,
      "num_input_tokens_seen": 24275592,
      "step": 37075
    },
    {
      "epoch": 19.433962264150942,
      "grad_norm": 0.5131540894508362,
      "learning_rate": 1.221252566554476e-07,
      "loss": 0.454,
      "num_input_tokens_seen": 24278984,
      "step": 37080
    },
    {
      "epoch": 19.43658280922432,
      "grad_norm": 0.4402117431163788,
      "learning_rate": 1.2099903612499685e-07,
      "loss": 0.5593,
      "num_input_tokens_seen": 24282248,
      "step": 37085
    },
    {
      "epoch": 19.439203354297693,
      "grad_norm": 0.727025032043457,
      "learning_rate": 1.198780199977101e-07,
      "loss": 0.4011,
      "num_input_tokens_seen": 24285960,
      "step": 37090
    },
    {
      "epoch": 19.44182389937107,
      "grad_norm": 0.4135187566280365,
      "learning_rate": 1.1876220850809416e-07,
      "loss": 0.3507,
      "num_input_tokens_seen": 24288392,
      "step": 37095
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.36397308111190796,
      "learning_rate": 1.1765160188956514e-07,
      "loss": 0.4188,
      "num_input_tokens_seen": 24291880,
      "step": 37100
    },
    {
      "epoch": 19.44706498951782,
      "grad_norm": 0.32488393783569336,
      "learning_rate": 1.1654620037445107e-07,
      "loss": 0.3691,
      "num_input_tokens_seen": 24295112,
      "step": 37105
    },
    {
      "epoch": 19.449685534591197,
      "grad_norm": 0.4024854898452759,
      "learning_rate": 1.1544600419398088e-07,
      "loss": 0.4445,
      "num_input_tokens_seen": 24298152,
      "step": 37110
    },
    {
      "epoch": 19.45230607966457,
      "grad_norm": 0.4721045196056366,
      "learning_rate": 1.1435101357831213e-07,
      "loss": 0.3702,
      "num_input_tokens_seen": 24302024,
      "step": 37115
    },
    {
      "epoch": 19.454926624737947,
      "grad_norm": 0.7712602615356445,
      "learning_rate": 1.1326122875650047e-07,
      "loss": 0.5007,
      "num_input_tokens_seen": 24305480,
      "step": 37120
    },
    {
      "epoch": 19.45754716981132,
      "grad_norm": 0.4375622868537903,
      "learning_rate": 1.1217664995651633e-07,
      "loss": 0.4122,
      "num_input_tokens_seen": 24309800,
      "step": 37125
    },
    {
      "epoch": 19.460167714884697,
      "grad_norm": 0.6968990564346313,
      "learning_rate": 1.1109727740523935e-07,
      "loss": 0.4566,
      "num_input_tokens_seen": 24313096,
      "step": 37130
    },
    {
      "epoch": 19.46278825995807,
      "grad_norm": 0.5004681944847107,
      "learning_rate": 1.1002311132846944e-07,
      "loss": 0.5421,
      "num_input_tokens_seen": 24315784,
      "step": 37135
    },
    {
      "epoch": 19.465408805031448,
      "grad_norm": 0.6188836097717285,
      "learning_rate": 1.0895415195090186e-07,
      "loss": 0.4698,
      "num_input_tokens_seen": 24318440,
      "step": 37140
    },
    {
      "epoch": 19.46802935010482,
      "grad_norm": 0.3857711851596832,
      "learning_rate": 1.0789039949615221e-07,
      "loss": 0.5032,
      "num_input_tokens_seen": 24321544,
      "step": 37145
    },
    {
      "epoch": 19.470649895178198,
      "grad_norm": 0.3943652808666229,
      "learning_rate": 1.0683185418674801e-07,
      "loss": 0.431,
      "num_input_tokens_seen": 24325192,
      "step": 37150
    },
    {
      "epoch": 19.47327044025157,
      "grad_norm": 0.753406286239624,
      "learning_rate": 1.0577851624412604e-07,
      "loss": 0.4697,
      "num_input_tokens_seen": 24328072,
      "step": 37155
    },
    {
      "epoch": 19.47589098532495,
      "grad_norm": 0.6353267431259155,
      "learning_rate": 1.0473038588862672e-07,
      "loss": 0.3224,
      "num_input_tokens_seen": 24331208,
      "step": 37160
    },
    {
      "epoch": 19.478511530398322,
      "grad_norm": 0.43362489342689514,
      "learning_rate": 1.0368746333951351e-07,
      "loss": 0.3525,
      "num_input_tokens_seen": 24334600,
      "step": 37165
    },
    {
      "epoch": 19.4811320754717,
      "grad_norm": 0.5307271480560303,
      "learning_rate": 1.0264974881494805e-07,
      "loss": 0.5675,
      "num_input_tokens_seen": 24338120,
      "step": 37170
    },
    {
      "epoch": 19.483752620545072,
      "grad_norm": 0.35592398047447205,
      "learning_rate": 1.0161724253201222e-07,
      "loss": 0.4408,
      "num_input_tokens_seen": 24341672,
      "step": 37175
    },
    {
      "epoch": 19.48637316561845,
      "grad_norm": 0.5952874422073364,
      "learning_rate": 1.0058994470669436e-07,
      "loss": 0.5176,
      "num_input_tokens_seen": 24345512,
      "step": 37180
    },
    {
      "epoch": 19.488993710691823,
      "grad_norm": 0.4451923966407776,
      "learning_rate": 9.956785555388926e-08,
      "loss": 0.4592,
      "num_input_tokens_seen": 24348616,
      "step": 37185
    },
    {
      "epoch": 19.4916142557652,
      "grad_norm": 0.42531025409698486,
      "learning_rate": 9.855097528741475e-08,
      "loss": 0.3775,
      "num_input_tokens_seen": 24351496,
      "step": 37190
    },
    {
      "epoch": 19.494234800838573,
      "grad_norm": 0.555091917514801,
      "learning_rate": 9.753930411998402e-08,
      "loss": 0.435,
      "num_input_tokens_seen": 24354248,
      "step": 37195
    },
    {
      "epoch": 19.49685534591195,
      "grad_norm": 0.6133989095687866,
      "learning_rate": 9.653284226322501e-08,
      "loss": 0.5727,
      "num_input_tokens_seen": 24357320,
      "step": 37200
    },
    {
      "epoch": 19.499475890985323,
      "grad_norm": 0.37114232778549194,
      "learning_rate": 9.553158992768318e-08,
      "loss": 0.5321,
      "num_input_tokens_seen": 24360520,
      "step": 37205
    },
    {
      "epoch": 19.5020964360587,
      "grad_norm": 0.3402423858642578,
      "learning_rate": 9.453554732281044e-08,
      "loss": 0.4008,
      "num_input_tokens_seen": 24364552,
      "step": 37210
    },
    {
      "epoch": 19.504716981132077,
      "grad_norm": 0.5987699627876282,
      "learning_rate": 9.354471465696236e-08,
      "loss": 0.4013,
      "num_input_tokens_seen": 24367464,
      "step": 37215
    },
    {
      "epoch": 19.50733752620545,
      "grad_norm": 0.9093635082244873,
      "learning_rate": 9.255909213741199e-08,
      "loss": 0.4894,
      "num_input_tokens_seen": 24370536,
      "step": 37220
    },
    {
      "epoch": 19.509958071278827,
      "grad_norm": 0.28882431983947754,
      "learning_rate": 9.157867997034164e-08,
      "loss": 0.5609,
      "num_input_tokens_seen": 24374760,
      "step": 37225
    },
    {
      "epoch": 19.5125786163522,
      "grad_norm": 0.8931187987327576,
      "learning_rate": 9.060347836084004e-08,
      "loss": 0.4816,
      "num_input_tokens_seen": 24377896,
      "step": 37230
    },
    {
      "epoch": 19.515199161425578,
      "grad_norm": 0.5378032922744751,
      "learning_rate": 8.96334875129079e-08,
      "loss": 0.5027,
      "num_input_tokens_seen": 24380392,
      "step": 37235
    },
    {
      "epoch": 19.51781970649895,
      "grad_norm": 0.5587863326072693,
      "learning_rate": 8.866870762946067e-08,
      "loss": 0.4517,
      "num_input_tokens_seen": 24383272,
      "step": 37240
    },
    {
      "epoch": 19.520440251572328,
      "grad_norm": 0.5119616389274597,
      "learning_rate": 8.77091389123147e-08,
      "loss": 0.563,
      "num_input_tokens_seen": 24386600,
      "step": 37245
    },
    {
      "epoch": 19.5230607966457,
      "grad_norm": 0.5532591938972473,
      "learning_rate": 8.675478156220385e-08,
      "loss": 0.4258,
      "num_input_tokens_seen": 24389832,
      "step": 37250
    },
    {
      "epoch": 19.52568134171908,
      "grad_norm": 0.30119219422340393,
      "learning_rate": 8.580563577876566e-08,
      "loss": 0.4423,
      "num_input_tokens_seen": 24393096,
      "step": 37255
    },
    {
      "epoch": 19.528301886792452,
      "grad_norm": 0.4477858245372772,
      "learning_rate": 8.486170176055242e-08,
      "loss": 0.4128,
      "num_input_tokens_seen": 24396008,
      "step": 37260
    },
    {
      "epoch": 19.53092243186583,
      "grad_norm": 0.36463385820388794,
      "learning_rate": 8.392297970502282e-08,
      "loss": 0.5053,
      "num_input_tokens_seen": 24398632,
      "step": 37265
    },
    {
      "epoch": 19.533542976939202,
      "grad_norm": 0.49859189987182617,
      "learning_rate": 8.298946980855315e-08,
      "loss": 0.4934,
      "num_input_tokens_seen": 24401992,
      "step": 37270
    },
    {
      "epoch": 19.53616352201258,
      "grad_norm": 0.39368098974227905,
      "learning_rate": 8.206117226641497e-08,
      "loss": 0.5349,
      "num_input_tokens_seen": 24405384,
      "step": 37275
    },
    {
      "epoch": 19.538784067085953,
      "grad_norm": 0.6640636920928955,
      "learning_rate": 8.113808727280015e-08,
      "loss": 0.5025,
      "num_input_tokens_seen": 24408136,
      "step": 37280
    },
    {
      "epoch": 19.54140461215933,
      "grad_norm": 0.29125019907951355,
      "learning_rate": 8.022021502080979e-08,
      "loss": 0.5659,
      "num_input_tokens_seen": 24412040,
      "step": 37285
    },
    {
      "epoch": 19.544025157232703,
      "grad_norm": 0.4710071384906769,
      "learning_rate": 7.930755570244863e-08,
      "loss": 0.4733,
      "num_input_tokens_seen": 24415080,
      "step": 37290
    },
    {
      "epoch": 19.54664570230608,
      "grad_norm": 0.8527817130088806,
      "learning_rate": 7.840010950863619e-08,
      "loss": 0.4559,
      "num_input_tokens_seen": 24417736,
      "step": 37295
    },
    {
      "epoch": 19.549266247379457,
      "grad_norm": 0.40400779247283936,
      "learning_rate": 7.749787662920116e-08,
      "loss": 0.5097,
      "num_input_tokens_seen": 24421224,
      "step": 37300
    },
    {
      "epoch": 19.55188679245283,
      "grad_norm": 0.33129602670669556,
      "learning_rate": 7.660085725288146e-08,
      "loss": 0.4275,
      "num_input_tokens_seen": 24427304,
      "step": 37305
    },
    {
      "epoch": 19.554507337526207,
      "grad_norm": 0.6610384583473206,
      "learning_rate": 7.570905156731867e-08,
      "loss": 0.4547,
      "num_input_tokens_seen": 24431080,
      "step": 37310
    },
    {
      "epoch": 19.55712788259958,
      "grad_norm": 0.3956250548362732,
      "learning_rate": 7.482245975907188e-08,
      "loss": 0.5015,
      "num_input_tokens_seen": 24433960,
      "step": 37315
    },
    {
      "epoch": 19.559748427672957,
      "grad_norm": 0.4233977496623993,
      "learning_rate": 7.394108201360383e-08,
      "loss": 0.3632,
      "num_input_tokens_seen": 24436616,
      "step": 37320
    },
    {
      "epoch": 19.56236897274633,
      "grad_norm": 0.9820997714996338,
      "learning_rate": 7.306491851529206e-08,
      "loss": 0.4943,
      "num_input_tokens_seen": 24439240,
      "step": 37325
    },
    {
      "epoch": 19.564989517819708,
      "grad_norm": 0.4786880612373352,
      "learning_rate": 7.219396944741774e-08,
      "loss": 0.5061,
      "num_input_tokens_seen": 24441992,
      "step": 37330
    },
    {
      "epoch": 19.56761006289308,
      "grad_norm": 0.4181821346282959,
      "learning_rate": 7.132823499217123e-08,
      "loss": 0.4113,
      "num_input_tokens_seen": 24444776,
      "step": 37335
    },
    {
      "epoch": 19.570230607966458,
      "grad_norm": 0.35816240310668945,
      "learning_rate": 7.046771533065766e-08,
      "loss": 0.4403,
      "num_input_tokens_seen": 24447720,
      "step": 37340
    },
    {
      "epoch": 19.57285115303983,
      "grad_norm": 0.29447805881500244,
      "learning_rate": 6.961241064288859e-08,
      "loss": 0.4262,
      "num_input_tokens_seen": 24452520,
      "step": 37345
    },
    {
      "epoch": 19.57547169811321,
      "grad_norm": 0.695354700088501,
      "learning_rate": 6.876232110778202e-08,
      "loss": 0.3811,
      "num_input_tokens_seen": 24454888,
      "step": 37350
    },
    {
      "epoch": 19.578092243186582,
      "grad_norm": 0.47248023748397827,
      "learning_rate": 6.791744690316515e-08,
      "loss": 0.3727,
      "num_input_tokens_seen": 24458152,
      "step": 37355
    },
    {
      "epoch": 19.58071278825996,
      "grad_norm": 0.2849186360836029,
      "learning_rate": 6.707778820577715e-08,
      "loss": 0.496,
      "num_input_tokens_seen": 24460904,
      "step": 37360
    },
    {
      "epoch": 19.583333333333332,
      "grad_norm": 0.5182445049285889,
      "learning_rate": 6.624334519126919e-08,
      "loss": 0.4166,
      "num_input_tokens_seen": 24464328,
      "step": 37365
    },
    {
      "epoch": 19.58595387840671,
      "grad_norm": 0.5726698637008667,
      "learning_rate": 6.541411803419329e-08,
      "loss": 0.5408,
      "num_input_tokens_seen": 24468072,
      "step": 37370
    },
    {
      "epoch": 19.588574423480082,
      "grad_norm": 0.45199882984161377,
      "learning_rate": 6.459010690801626e-08,
      "loss": 0.5271,
      "num_input_tokens_seen": 24472424,
      "step": 37375
    },
    {
      "epoch": 19.59119496855346,
      "grad_norm": 0.41495975852012634,
      "learning_rate": 6.377131198510855e-08,
      "loss": 0.288,
      "num_input_tokens_seen": 24475656,
      "step": 37380
    },
    {
      "epoch": 19.593815513626833,
      "grad_norm": 0.6282992959022522,
      "learning_rate": 6.295773343675537e-08,
      "loss": 0.351,
      "num_input_tokens_seen": 24479368,
      "step": 37385
    },
    {
      "epoch": 19.59643605870021,
      "grad_norm": 0.4553171694278717,
      "learning_rate": 6.214937143314837e-08,
      "loss": 0.3809,
      "num_input_tokens_seen": 24483048,
      "step": 37390
    },
    {
      "epoch": 19.599056603773583,
      "grad_norm": 0.4080677628517151,
      "learning_rate": 6.134622614338559e-08,
      "loss": 0.4489,
      "num_input_tokens_seen": 24486152,
      "step": 37395
    },
    {
      "epoch": 19.60167714884696,
      "grad_norm": 0.38353031873703003,
      "learning_rate": 6.05482977354771e-08,
      "loss": 0.392,
      "num_input_tokens_seen": 24489192,
      "step": 37400
    },
    {
      "epoch": 19.604297693920337,
      "grad_norm": 0.34740275144577026,
      "learning_rate": 5.975558637634215e-08,
      "loss": 0.5232,
      "num_input_tokens_seen": 24492392,
      "step": 37405
    },
    {
      "epoch": 19.60691823899371,
      "grad_norm": 0.8556483387947083,
      "learning_rate": 5.896809223180366e-08,
      "loss": 0.427,
      "num_input_tokens_seen": 24494760,
      "step": 37410
    },
    {
      "epoch": 19.609538784067087,
      "grad_norm": 0.6230118870735168,
      "learning_rate": 5.8185815466596516e-08,
      "loss": 0.4464,
      "num_input_tokens_seen": 24497672,
      "step": 37415
    },
    {
      "epoch": 19.61215932914046,
      "grad_norm": 0.3888949751853943,
      "learning_rate": 5.7408756244370365e-08,
      "loss": 0.4477,
      "num_input_tokens_seen": 24501704,
      "step": 37420
    },
    {
      "epoch": 19.614779874213838,
      "grad_norm": 0.5409068465232849,
      "learning_rate": 5.663691472766741e-08,
      "loss": 0.3818,
      "num_input_tokens_seen": 24505576,
      "step": 37425
    },
    {
      "epoch": 19.61740041928721,
      "grad_norm": 0.4844219386577606,
      "learning_rate": 5.587029107795572e-08,
      "loss": 0.493,
      "num_input_tokens_seen": 24508616,
      "step": 37430
    },
    {
      "epoch": 19.620020964360588,
      "grad_norm": 0.3668571412563324,
      "learning_rate": 5.510888545559867e-08,
      "loss": 0.4462,
      "num_input_tokens_seen": 24511528,
      "step": 37435
    },
    {
      "epoch": 19.62264150943396,
      "grad_norm": 0.680854320526123,
      "learning_rate": 5.435269801987997e-08,
      "loss": 0.5457,
      "num_input_tokens_seen": 24514632,
      "step": 37440
    },
    {
      "epoch": 19.62526205450734,
      "grad_norm": 0.49054867029190063,
      "learning_rate": 5.360172892897863e-08,
      "loss": 0.4877,
      "num_input_tokens_seen": 24518088,
      "step": 37445
    },
    {
      "epoch": 19.627882599580712,
      "grad_norm": 0.45408254861831665,
      "learning_rate": 5.285597833999401e-08,
      "loss": 0.4443,
      "num_input_tokens_seen": 24520904,
      "step": 37450
    },
    {
      "epoch": 19.63050314465409,
      "grad_norm": 0.5352818369865417,
      "learning_rate": 5.2115446408926295e-08,
      "loss": 0.4424,
      "num_input_tokens_seen": 24524104,
      "step": 37455
    },
    {
      "epoch": 19.633123689727462,
      "grad_norm": 0.5527013540267944,
      "learning_rate": 5.138013329068769e-08,
      "loss": 0.4926,
      "num_input_tokens_seen": 24527624,
      "step": 37460
    },
    {
      "epoch": 19.63574423480084,
      "grad_norm": 0.45675036311149597,
      "learning_rate": 5.065003913909405e-08,
      "loss": 0.4463,
      "num_input_tokens_seen": 24529960,
      "step": 37465
    },
    {
      "epoch": 19.638364779874212,
      "grad_norm": 0.8942098021507263,
      "learning_rate": 4.9925164106875975e-08,
      "loss": 0.7539,
      "num_input_tokens_seen": 24532616,
      "step": 37470
    },
    {
      "epoch": 19.64098532494759,
      "grad_norm": 0.6493972539901733,
      "learning_rate": 4.9205508345670506e-08,
      "loss": 0.336,
      "num_input_tokens_seen": 24535880,
      "step": 37475
    },
    {
      "epoch": 19.643605870020963,
      "grad_norm": 0.5401037335395813,
      "learning_rate": 4.8491072006015546e-08,
      "loss": 0.4411,
      "num_input_tokens_seen": 24540296,
      "step": 37480
    },
    {
      "epoch": 19.64622641509434,
      "grad_norm": 0.24944068491458893,
      "learning_rate": 4.778185523736933e-08,
      "loss": 0.4191,
      "num_input_tokens_seen": 24543496,
      "step": 37485
    },
    {
      "epoch": 19.648846960167717,
      "grad_norm": 0.7906742691993713,
      "learning_rate": 4.7077858188085414e-08,
      "loss": 0.3741,
      "num_input_tokens_seen": 24550056,
      "step": 37490
    },
    {
      "epoch": 19.65146750524109,
      "grad_norm": 0.3867514133453369,
      "learning_rate": 4.6379081005437644e-08,
      "loss": 0.4294,
      "num_input_tokens_seen": 24553608,
      "step": 37495
    },
    {
      "epoch": 19.654088050314467,
      "grad_norm": 0.9979223608970642,
      "learning_rate": 4.568552383560076e-08,
      "loss": 0.5885,
      "num_input_tokens_seen": 24557224,
      "step": 37500
    },
    {
      "epoch": 19.65670859538784,
      "grad_norm": 0.4692364037036896,
      "learning_rate": 4.4997186823655945e-08,
      "loss": 0.6334,
      "num_input_tokens_seen": 24560616,
      "step": 37505
    },
    {
      "epoch": 19.659329140461217,
      "grad_norm": 0.4361220598220825,
      "learning_rate": 4.431407011360189e-08,
      "loss": 0.701,
      "num_input_tokens_seen": 24563656,
      "step": 37510
    },
    {
      "epoch": 19.66194968553459,
      "grad_norm": 0.4327263832092285,
      "learning_rate": 4.3636173848332627e-08,
      "loss": 0.4058,
      "num_input_tokens_seen": 24566312,
      "step": 37515
    },
    {
      "epoch": 19.664570230607968,
      "grad_norm": 0.3688875436782837,
      "learning_rate": 4.2963498169656946e-08,
      "loss": 0.5336,
      "num_input_tokens_seen": 24569640,
      "step": 37520
    },
    {
      "epoch": 19.66719077568134,
      "grad_norm": 0.49020451307296753,
      "learning_rate": 4.229604321829561e-08,
      "loss": 0.5204,
      "num_input_tokens_seen": 24572264,
      "step": 37525
    },
    {
      "epoch": 19.669811320754718,
      "grad_norm": 0.5000865459442139,
      "learning_rate": 4.16338091338675e-08,
      "loss": 0.4216,
      "num_input_tokens_seen": 24575048,
      "step": 37530
    },
    {
      "epoch": 19.67243186582809,
      "grad_norm": 0.4338620901107788,
      "learning_rate": 4.097679605490623e-08,
      "loss": 0.4457,
      "num_input_tokens_seen": 24577704,
      "step": 37535
    },
    {
      "epoch": 19.67505241090147,
      "grad_norm": 0.4925406873226166,
      "learning_rate": 4.032500411885465e-08,
      "loss": 0.4392,
      "num_input_tokens_seen": 24581352,
      "step": 37540
    },
    {
      "epoch": 19.677672955974842,
      "grad_norm": 0.5090775489807129,
      "learning_rate": 3.967843346205369e-08,
      "loss": 0.4402,
      "num_input_tokens_seen": 24584168,
      "step": 37545
    },
    {
      "epoch": 19.68029350104822,
      "grad_norm": 0.44714120030403137,
      "learning_rate": 3.903708421976182e-08,
      "loss": 0.326,
      "num_input_tokens_seen": 24588200,
      "step": 37550
    },
    {
      "epoch": 19.682914046121592,
      "grad_norm": 0.5567954182624817,
      "learning_rate": 3.840095652614395e-08,
      "loss": 0.6148,
      "num_input_tokens_seen": 24590568,
      "step": 37555
    },
    {
      "epoch": 19.68553459119497,
      "grad_norm": 0.5622734427452087,
      "learning_rate": 3.777005051426863e-08,
      "loss": 0.4287,
      "num_input_tokens_seen": 24593320,
      "step": 37560
    },
    {
      "epoch": 19.688155136268342,
      "grad_norm": 0.8229444622993469,
      "learning_rate": 3.71443663161164e-08,
      "loss": 0.5627,
      "num_input_tokens_seen": 24595848,
      "step": 37565
    },
    {
      "epoch": 19.69077568134172,
      "grad_norm": 0.3955531716346741,
      "learning_rate": 3.652390406256867e-08,
      "loss": 0.3812,
      "num_input_tokens_seen": 24598600,
      "step": 37570
    },
    {
      "epoch": 19.693396226415093,
      "grad_norm": 0.5058721899986267,
      "learning_rate": 3.5908663883421624e-08,
      "loss": 0.4468,
      "num_input_tokens_seen": 24601544,
      "step": 37575
    },
    {
      "epoch": 19.69601677148847,
      "grad_norm": 0.4421572983264923,
      "learning_rate": 3.529864590737786e-08,
      "loss": 0.4225,
      "num_input_tokens_seen": 24604552,
      "step": 37580
    },
    {
      "epoch": 19.698637316561843,
      "grad_norm": 0.4886750280857086,
      "learning_rate": 3.4693850262046415e-08,
      "loss": 0.5191,
      "num_input_tokens_seen": 24607432,
      "step": 37585
    },
    {
      "epoch": 19.70125786163522,
      "grad_norm": 0.7078866362571716,
      "learning_rate": 3.409427707394275e-08,
      "loss": 0.4729,
      "num_input_tokens_seen": 24611464,
      "step": 37590
    },
    {
      "epoch": 19.703878406708597,
      "grad_norm": 0.7772899270057678,
      "learning_rate": 3.349992646848599e-08,
      "loss": 0.5035,
      "num_input_tokens_seen": 24614344,
      "step": 37595
    },
    {
      "epoch": 19.70649895178197,
      "grad_norm": 0.5056313276290894,
      "learning_rate": 3.291079857001556e-08,
      "loss": 0.4931,
      "num_input_tokens_seen": 24617864,
      "step": 37600
    },
    {
      "epoch": 19.709119496855347,
      "grad_norm": 0.5775796175003052,
      "learning_rate": 3.2326893501763436e-08,
      "loss": 0.3941,
      "num_input_tokens_seen": 24620904,
      "step": 37605
    },
    {
      "epoch": 19.71174004192872,
      "grad_norm": 0.6647204160690308,
      "learning_rate": 3.174821138588191e-08,
      "loss": 0.3351,
      "num_input_tokens_seen": 24623560,
      "step": 37610
    },
    {
      "epoch": 19.714360587002098,
      "grad_norm": 0.4743652045726776,
      "learning_rate": 3.11747523434186e-08,
      "loss": 0.3867,
      "num_input_tokens_seen": 24627016,
      "step": 37615
    },
    {
      "epoch": 19.71698113207547,
      "grad_norm": 0.27481406927108765,
      "learning_rate": 3.060651649434143e-08,
      "loss": 0.3794,
      "num_input_tokens_seen": 24630600,
      "step": 37620
    },
    {
      "epoch": 19.719601677148848,
      "grad_norm": 0.49680131673812866,
      "learning_rate": 3.004350395751088e-08,
      "loss": 0.5276,
      "num_input_tokens_seen": 24634248,
      "step": 37625
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.47047504782676697,
      "learning_rate": 2.94857148507105e-08,
      "loss": 0.3134,
      "num_input_tokens_seen": 24637352,
      "step": 37630
    },
    {
      "epoch": 19.7248427672956,
      "grad_norm": 0.5366686582565308,
      "learning_rate": 2.893314929061641e-08,
      "loss": 0.4371,
      "num_input_tokens_seen": 24639912,
      "step": 37635
    },
    {
      "epoch": 19.72746331236897,
      "grad_norm": 0.4516195058822632,
      "learning_rate": 2.8385807392825037e-08,
      "loss": 0.3728,
      "num_input_tokens_seen": 24642568,
      "step": 37640
    },
    {
      "epoch": 19.73008385744235,
      "grad_norm": 0.5243638157844543,
      "learning_rate": 2.7843689271830897e-08,
      "loss": 0.3378,
      "num_input_tokens_seen": 24645416,
      "step": 37645
    },
    {
      "epoch": 19.732704402515722,
      "grad_norm": 0.5873776078224182,
      "learning_rate": 2.7306795041040502e-08,
      "loss": 0.422,
      "num_input_tokens_seen": 24648360,
      "step": 37650
    },
    {
      "epoch": 19.7353249475891,
      "grad_norm": 1.1059396266937256,
      "learning_rate": 2.677512481276401e-08,
      "loss": 0.5336,
      "num_input_tokens_seen": 24650728,
      "step": 37655
    },
    {
      "epoch": 19.737945492662472,
      "grad_norm": 1.2737003564834595,
      "learning_rate": 2.6248678698220785e-08,
      "loss": 0.5166,
      "num_input_tokens_seen": 24653928,
      "step": 37660
    },
    {
      "epoch": 19.74056603773585,
      "grad_norm": 0.34912756085395813,
      "learning_rate": 2.5727456807542183e-08,
      "loss": 0.4102,
      "num_input_tokens_seen": 24657608,
      "step": 37665
    },
    {
      "epoch": 19.743186582809223,
      "grad_norm": 0.9864776134490967,
      "learning_rate": 2.5211459249754875e-08,
      "loss": 0.3345,
      "num_input_tokens_seen": 24660744,
      "step": 37670
    },
    {
      "epoch": 19.7458071278826,
      "grad_norm": 0.5858588814735413,
      "learning_rate": 2.4700686132803076e-08,
      "loss": 0.4155,
      "num_input_tokens_seen": 24664136,
      "step": 37675
    },
    {
      "epoch": 19.748427672955973,
      "grad_norm": 0.40748098492622375,
      "learning_rate": 2.4195137563534643e-08,
      "loss": 0.4554,
      "num_input_tokens_seen": 24666920,
      "step": 37680
    },
    {
      "epoch": 19.75104821802935,
      "grad_norm": 0.712846040725708,
      "learning_rate": 2.369481364770387e-08,
      "loss": 0.467,
      "num_input_tokens_seen": 24669800,
      "step": 37685
    },
    {
      "epoch": 19.753668763102727,
      "grad_norm": 0.33522161841392517,
      "learning_rate": 2.3199714489974266e-08,
      "loss": 0.4551,
      "num_input_tokens_seen": 24672488,
      "step": 37690
    },
    {
      "epoch": 19.7562893081761,
      "grad_norm": 0.5464876890182495,
      "learning_rate": 2.2709840193912978e-08,
      "loss": 0.4847,
      "num_input_tokens_seen": 24676168,
      "step": 37695
    },
    {
      "epoch": 19.758909853249477,
      "grad_norm": 0.8185508847236633,
      "learning_rate": 2.2225190861999146e-08,
      "loss": 0.5227,
      "num_input_tokens_seen": 24679176,
      "step": 37700
    },
    {
      "epoch": 19.76153039832285,
      "grad_norm": 0.6969055533409119,
      "learning_rate": 2.1745766595612783e-08,
      "loss": 0.3551,
      "num_input_tokens_seen": 24681864,
      "step": 37705
    },
    {
      "epoch": 19.764150943396228,
      "grad_norm": 0.45263805985450745,
      "learning_rate": 2.1271567495045886e-08,
      "loss": 0.4831,
      "num_input_tokens_seen": 24684712,
      "step": 37710
    },
    {
      "epoch": 19.7667714884696,
      "grad_norm": 0.44236522912979126,
      "learning_rate": 2.0802593659491333e-08,
      "loss": 0.5207,
      "num_input_tokens_seen": 24687336,
      "step": 37715
    },
    {
      "epoch": 19.769392033542978,
      "grad_norm": 0.5392473340034485,
      "learning_rate": 2.0338845187059528e-08,
      "loss": 0.322,
      "num_input_tokens_seen": 24690664,
      "step": 37720
    },
    {
      "epoch": 19.77201257861635,
      "grad_norm": 0.41401904821395874,
      "learning_rate": 1.988032217475899e-08,
      "loss": 0.4606,
      "num_input_tokens_seen": 24694216,
      "step": 37725
    },
    {
      "epoch": 19.77463312368973,
      "grad_norm": 1.129934549331665,
      "learning_rate": 1.942702471850466e-08,
      "loss": 0.4856,
      "num_input_tokens_seen": 24696936,
      "step": 37730
    },
    {
      "epoch": 19.7772536687631,
      "grad_norm": 0.5158014893531799,
      "learning_rate": 1.897895291312346e-08,
      "loss": 0.4045,
      "num_input_tokens_seen": 24699400,
      "step": 37735
    },
    {
      "epoch": 19.77987421383648,
      "grad_norm": 0.5553308725357056,
      "learning_rate": 1.8536106852348746e-08,
      "loss": 0.4791,
      "num_input_tokens_seen": 24702376,
      "step": 37740
    },
    {
      "epoch": 19.782494758909852,
      "grad_norm": 0.3544132709503174,
      "learning_rate": 1.8098486628814748e-08,
      "loss": 0.3487,
      "num_input_tokens_seen": 24706568,
      "step": 37745
    },
    {
      "epoch": 19.78511530398323,
      "grad_norm": 0.3610681891441345,
      "learning_rate": 1.7666092334067684e-08,
      "loss": 0.4764,
      "num_input_tokens_seen": 24709384,
      "step": 37750
    },
    {
      "epoch": 19.787735849056602,
      "grad_norm": 0.39503970742225647,
      "learning_rate": 1.7238924058562977e-08,
      "loss": 0.3901,
      "num_input_tokens_seen": 24712552,
      "step": 37755
    },
    {
      "epoch": 19.79035639412998,
      "grad_norm": 0.4996175467967987,
      "learning_rate": 1.6816981891656924e-08,
      "loss": 0.3615,
      "num_input_tokens_seen": 24715400,
      "step": 37760
    },
    {
      "epoch": 19.792976939203353,
      "grad_norm": 0.6463298201560974,
      "learning_rate": 1.6400265921615032e-08,
      "loss": 0.4697,
      "num_input_tokens_seen": 24718376,
      "step": 37765
    },
    {
      "epoch": 19.79559748427673,
      "grad_norm": 0.5576300024986267,
      "learning_rate": 1.598877623560924e-08,
      "loss": 0.5776,
      "num_input_tokens_seen": 24721416,
      "step": 37770
    },
    {
      "epoch": 19.798218029350103,
      "grad_norm": 0.49093523621559143,
      "learning_rate": 1.5582512919717908e-08,
      "loss": 0.4905,
      "num_input_tokens_seen": 24724232,
      "step": 37775
    },
    {
      "epoch": 19.80083857442348,
      "grad_norm": 0.462528258562088,
      "learning_rate": 1.5181476058928613e-08,
      "loss": 0.5983,
      "num_input_tokens_seen": 24731656,
      "step": 37780
    },
    {
      "epoch": 19.803459119496857,
      "grad_norm": 0.639585018157959,
      "learning_rate": 1.4785665737129805e-08,
      "loss": 0.5191,
      "num_input_tokens_seen": 24734344,
      "step": 37785
    },
    {
      "epoch": 19.80607966457023,
      "grad_norm": 0.5650954842567444,
      "learning_rate": 1.4395082037127471e-08,
      "loss": 0.4177,
      "num_input_tokens_seen": 24737416,
      "step": 37790
    },
    {
      "epoch": 19.808700209643607,
      "grad_norm": 0.3098074197769165,
      "learning_rate": 1.4009725040620148e-08,
      "loss": 0.3912,
      "num_input_tokens_seen": 24740648,
      "step": 37795
    },
    {
      "epoch": 19.81132075471698,
      "grad_norm": 0.2641153931617737,
      "learning_rate": 1.3629594828223902e-08,
      "loss": 0.3647,
      "num_input_tokens_seen": 24744712,
      "step": 37800
    },
    {
      "epoch": 19.813941299790358,
      "grad_norm": 0.5199506878852844,
      "learning_rate": 1.3254691479458458e-08,
      "loss": 0.3835,
      "num_input_tokens_seen": 24747592,
      "step": 37805
    },
    {
      "epoch": 19.81656184486373,
      "grad_norm": 0.42697566747665405,
      "learning_rate": 1.2885015072747197e-08,
      "loss": 0.4036,
      "num_input_tokens_seen": 24750920,
      "step": 37810
    },
    {
      "epoch": 19.819182389937108,
      "grad_norm": 0.30768316984176636,
      "learning_rate": 1.2520565685419927e-08,
      "loss": 0.3622,
      "num_input_tokens_seen": 24756200,
      "step": 37815
    },
    {
      "epoch": 19.82180293501048,
      "grad_norm": 0.5707959532737732,
      "learning_rate": 1.2161343393721214e-08,
      "loss": 0.4073,
      "num_input_tokens_seen": 24759144,
      "step": 37820
    },
    {
      "epoch": 19.82442348008386,
      "grad_norm": 0.7727971076965332,
      "learning_rate": 1.1807348272790952e-08,
      "loss": 0.421,
      "num_input_tokens_seen": 24761992,
      "step": 37825
    },
    {
      "epoch": 19.82704402515723,
      "grad_norm": 0.838967502117157,
      "learning_rate": 1.1458580396683794e-08,
      "loss": 0.4357,
      "num_input_tokens_seen": 24764744,
      "step": 37830
    },
    {
      "epoch": 19.82966457023061,
      "grad_norm": 0.4871392846107483,
      "learning_rate": 1.1115039838358044e-08,
      "loss": 0.5559,
      "num_input_tokens_seen": 24767336,
      "step": 37835
    },
    {
      "epoch": 19.832285115303982,
      "grad_norm": 0.41756319999694824,
      "learning_rate": 1.0776726669678439e-08,
      "loss": 0.4541,
      "num_input_tokens_seen": 24770344,
      "step": 37840
    },
    {
      "epoch": 19.83490566037736,
      "grad_norm": 0.7861230969429016,
      "learning_rate": 1.044364096141337e-08,
      "loss": 0.4644,
      "num_input_tokens_seen": 24773704,
      "step": 37845
    },
    {
      "epoch": 19.837526205450732,
      "grad_norm": 0.4437408745288849,
      "learning_rate": 1.0115782783243211e-08,
      "loss": 0.4391,
      "num_input_tokens_seen": 24776840,
      "step": 37850
    },
    {
      "epoch": 19.84014675052411,
      "grad_norm": 0.6942259669303894,
      "learning_rate": 9.793152203751988e-09,
      "loss": 0.4413,
      "num_input_tokens_seen": 24780904,
      "step": 37855
    },
    {
      "epoch": 19.842767295597483,
      "grad_norm": 0.38884100317955017,
      "learning_rate": 9.475749290430158e-09,
      "loss": 0.384,
      "num_input_tokens_seen": 24784104,
      "step": 37860
    },
    {
      "epoch": 19.84538784067086,
      "grad_norm": 0.3167460262775421,
      "learning_rate": 9.16357410967461e-09,
      "loss": 0.4103,
      "num_input_tokens_seen": 24788712,
      "step": 37865
    },
    {
      "epoch": 19.848008385744233,
      "grad_norm": 0.4950598180294037,
      "learning_rate": 8.856626726788664e-09,
      "loss": 0.4023,
      "num_input_tokens_seen": 24791784,
      "step": 37870
    },
    {
      "epoch": 19.85062893081761,
      "grad_norm": 0.3537774682044983,
      "learning_rate": 8.554907205982066e-09,
      "loss": 0.5174,
      "num_input_tokens_seen": 24795016,
      "step": 37875
    },
    {
      "epoch": 19.853249475890987,
      "grad_norm": 0.5130225419998169,
      "learning_rate": 8.258415610373772e-09,
      "loss": 0.4913,
      "num_input_tokens_seen": 24798600,
      "step": 37880
    },
    {
      "epoch": 19.85587002096436,
      "grad_norm": 0.6898267269134521,
      "learning_rate": 7.967152001980838e-09,
      "loss": 0.5301,
      "num_input_tokens_seen": 24801288,
      "step": 37885
    },
    {
      "epoch": 19.858490566037737,
      "grad_norm": 0.30650272965431213,
      "learning_rate": 7.681116441737856e-09,
      "loss": 0.3859,
      "num_input_tokens_seen": 24803976,
      "step": 37890
    },
    {
      "epoch": 19.86111111111111,
      "grad_norm": 0.6131160855293274,
      "learning_rate": 7.4003089894747464e-09,
      "loss": 0.5237,
      "num_input_tokens_seen": 24806632,
      "step": 37895
    },
    {
      "epoch": 19.863731656184488,
      "grad_norm": 0.5778619647026062,
      "learning_rate": 7.12472970393896e-09,
      "loss": 0.3721,
      "num_input_tokens_seen": 24809160,
      "step": 37900
    },
    {
      "epoch": 19.86635220125786,
      "grad_norm": 0.4302766025066376,
      "learning_rate": 6.854378642776049e-09,
      "loss": 0.5168,
      "num_input_tokens_seen": 24812584,
      "step": 37905
    },
    {
      "epoch": 19.868972746331238,
      "grad_norm": 0.5051471590995789,
      "learning_rate": 6.589255862538002e-09,
      "loss": 0.5348,
      "num_input_tokens_seen": 24816104,
      "step": 37910
    },
    {
      "epoch": 19.87159329140461,
      "grad_norm": 0.5976493954658508,
      "learning_rate": 6.32936141869156e-09,
      "loss": 0.4417,
      "num_input_tokens_seen": 24819112,
      "step": 37915
    },
    {
      "epoch": 19.87421383647799,
      "grad_norm": 0.5276509523391724,
      "learning_rate": 6.074695365596017e-09,
      "loss": 0.3217,
      "num_input_tokens_seen": 24822504,
      "step": 37920
    },
    {
      "epoch": 19.87683438155136,
      "grad_norm": 0.4809226095676422,
      "learning_rate": 5.8252577565309775e-09,
      "loss": 0.3992,
      "num_input_tokens_seen": 24825224,
      "step": 37925
    },
    {
      "epoch": 19.87945492662474,
      "grad_norm": 0.9491074681282043,
      "learning_rate": 5.581048643671371e-09,
      "loss": 0.5147,
      "num_input_tokens_seen": 24828104,
      "step": 37930
    },
    {
      "epoch": 19.882075471698112,
      "grad_norm": 0.3115301728248596,
      "learning_rate": 5.342068078106888e-09,
      "loss": 0.4027,
      "num_input_tokens_seen": 24831592,
      "step": 37935
    },
    {
      "epoch": 19.88469601677149,
      "grad_norm": 0.4621322453022003,
      "learning_rate": 5.1083161098253175e-09,
      "loss": 0.4657,
      "num_input_tokens_seen": 24834824,
      "step": 37940
    },
    {
      "epoch": 19.887316561844862,
      "grad_norm": 0.4292811155319214,
      "learning_rate": 4.87979278772921e-09,
      "loss": 0.4608,
      "num_input_tokens_seen": 24838184,
      "step": 37945
    },
    {
      "epoch": 19.88993710691824,
      "grad_norm": 0.3165460526943207,
      "learning_rate": 4.656498159619216e-09,
      "loss": 0.4081,
      "num_input_tokens_seen": 24841576,
      "step": 37950
    },
    {
      "epoch": 19.892557651991613,
      "grad_norm": 1.1505403518676758,
      "learning_rate": 4.438432272210746e-09,
      "loss": 0.536,
      "num_input_tokens_seen": 24844232,
      "step": 37955
    },
    {
      "epoch": 19.89517819706499,
      "grad_norm": 0.4923076629638672,
      "learning_rate": 4.225595171114538e-09,
      "loss": 0.3955,
      "num_input_tokens_seen": 24847144,
      "step": 37960
    },
    {
      "epoch": 19.897798742138363,
      "grad_norm": 0.7219642996788025,
      "learning_rate": 4.017986900858861e-09,
      "loss": 0.4294,
      "num_input_tokens_seen": 24850568,
      "step": 37965
    },
    {
      "epoch": 19.90041928721174,
      "grad_norm": 0.5585688352584839,
      "learning_rate": 3.815607504870089e-09,
      "loss": 0.3631,
      "num_input_tokens_seen": 24852808,
      "step": 37970
    },
    {
      "epoch": 19.903039832285117,
      "grad_norm": 0.43503716588020325,
      "learning_rate": 3.618457025486577e-09,
      "loss": 0.5043,
      "num_input_tokens_seen": 24856744,
      "step": 37975
    },
    {
      "epoch": 19.90566037735849,
      "grad_norm": 0.4505484700202942,
      "learning_rate": 3.426535503947559e-09,
      "loss": 0.4774,
      "num_input_tokens_seen": 24860296,
      "step": 37980
    },
    {
      "epoch": 19.908280922431867,
      "grad_norm": 0.4168458878993988,
      "learning_rate": 3.2398429803986997e-09,
      "loss": 0.3905,
      "num_input_tokens_seen": 24863560,
      "step": 37985
    },
    {
      "epoch": 19.91090146750524,
      "grad_norm": 0.4456072747707367,
      "learning_rate": 3.058379493900421e-09,
      "loss": 0.3853,
      "num_input_tokens_seen": 24866344,
      "step": 37990
    },
    {
      "epoch": 19.913522012578618,
      "grad_norm": 0.3715256452560425,
      "learning_rate": 2.882145082405696e-09,
      "loss": 0.4757,
      "num_input_tokens_seen": 24869800,
      "step": 37995
    },
    {
      "epoch": 19.91614255765199,
      "grad_norm": 0.530022144317627,
      "learning_rate": 2.711139782785033e-09,
      "loss": 0.3723,
      "num_input_tokens_seen": 24872264,
      "step": 38000
    },
    {
      "epoch": 19.918763102725368,
      "grad_norm": 0.49171754717826843,
      "learning_rate": 2.5453636308098163e-09,
      "loss": 0.4603,
      "num_input_tokens_seen": 24875816,
      "step": 38005
    },
    {
      "epoch": 19.92138364779874,
      "grad_norm": 0.4121414124965668,
      "learning_rate": 2.3848166611606383e-09,
      "loss": 0.4326,
      "num_input_tokens_seen": 24878856,
      "step": 38010
    },
    {
      "epoch": 19.92400419287212,
      "grad_norm": 0.5756369233131409,
      "learning_rate": 2.2294989074161943e-09,
      "loss": 0.4298,
      "num_input_tokens_seen": 24881864,
      "step": 38015
    },
    {
      "epoch": 19.92662473794549,
      "grad_norm": 0.636431097984314,
      "learning_rate": 2.0794104020727125e-09,
      "loss": 0.4486,
      "num_input_tokens_seen": 24884488,
      "step": 38020
    },
    {
      "epoch": 19.92924528301887,
      "grad_norm": 0.440079003572464,
      "learning_rate": 1.9345511765272993e-09,
      "loss": 0.3378,
      "num_input_tokens_seen": 24887944,
      "step": 38025
    },
    {
      "epoch": 19.931865828092242,
      "grad_norm": 0.37305954098701477,
      "learning_rate": 1.794921261077942e-09,
      "loss": 0.4162,
      "num_input_tokens_seen": 24891400,
      "step": 38030
    },
    {
      "epoch": 19.93448637316562,
      "grad_norm": 0.685915470123291,
      "learning_rate": 1.6605206849373833e-09,
      "loss": 0.3552,
      "num_input_tokens_seen": 24893672,
      "step": 38035
    },
    {
      "epoch": 19.937106918238992,
      "grad_norm": 0.520357072353363,
      "learning_rate": 1.531349476219246e-09,
      "loss": 0.4454,
      "num_input_tokens_seen": 24897544,
      "step": 38040
    },
    {
      "epoch": 19.93972746331237,
      "grad_norm": 0.2918347418308258,
      "learning_rate": 1.4074076619463584e-09,
      "loss": 0.6031,
      "num_input_tokens_seen": 24900904,
      "step": 38045
    },
    {
      "epoch": 19.942348008385743,
      "grad_norm": 0.22383345663547516,
      "learning_rate": 1.2886952680452036e-09,
      "loss": 0.3959,
      "num_input_tokens_seen": 24905064,
      "step": 38050
    },
    {
      "epoch": 19.94496855345912,
      "grad_norm": 0.4527980089187622,
      "learning_rate": 1.1752123193459197e-09,
      "loss": 0.4479,
      "num_input_tokens_seen": 24908392,
      "step": 38055
    },
    {
      "epoch": 19.947589098532493,
      "grad_norm": 0.3902800679206848,
      "learning_rate": 1.0669588395934017e-09,
      "loss": 0.4144,
      "num_input_tokens_seen": 24911336,
      "step": 38060
    },
    {
      "epoch": 19.95020964360587,
      "grad_norm": 0.4598628580570221,
      "learning_rate": 9.639348514278722e-10,
      "loss": 0.3771,
      "num_input_tokens_seen": 24915880,
      "step": 38065
    },
    {
      "epoch": 19.952830188679247,
      "grad_norm": 0.32229575514793396,
      "learning_rate": 8.661403764043119e-10,
      "loss": 0.5652,
      "num_input_tokens_seen": 24919464,
      "step": 38070
    },
    {
      "epoch": 19.95545073375262,
      "grad_norm": 0.5587350130081177,
      "learning_rate": 7.735754349785795e-10,
      "loss": 0.2789,
      "num_input_tokens_seen": 24922312,
      "step": 38075
    },
    {
      "epoch": 19.958071278825997,
      "grad_norm": 0.3344362676143646,
      "learning_rate": 6.862400465157403e-10,
      "loss": 0.4645,
      "num_input_tokens_seen": 24925928,
      "step": 38080
    },
    {
      "epoch": 19.96069182389937,
      "grad_norm": 0.6734849810600281,
      "learning_rate": 6.041342292817387e-10,
      "loss": 0.3784,
      "num_input_tokens_seen": 24929224,
      "step": 38085
    },
    {
      "epoch": 19.963312368972748,
      "grad_norm": 0.47502586245536804,
      "learning_rate": 5.272580004572758e-10,
      "loss": 0.381,
      "num_input_tokens_seen": 24931752,
      "step": 38090
    },
    {
      "epoch": 19.96593291404612,
      "grad_norm": 0.8061944246292114,
      "learning_rate": 4.556113761211567e-10,
      "loss": 0.6642,
      "num_input_tokens_seen": 24934504,
      "step": 38095
    },
    {
      "epoch": 19.968553459119498,
      "grad_norm": 0.5492798089981079,
      "learning_rate": 3.8919437125861657e-10,
      "loss": 0.5346,
      "num_input_tokens_seen": 24937512,
      "step": 38100
    },
    {
      "epoch": 19.97117400419287,
      "grad_norm": 0.37479841709136963,
      "learning_rate": 3.280069997696478e-10,
      "loss": 0.3288,
      "num_input_tokens_seen": 24940520,
      "step": 38105
    },
    {
      "epoch": 19.97379454926625,
      "grad_norm": 0.4705444276332855,
      "learning_rate": 2.720492744467951e-10,
      "loss": 0.3693,
      "num_input_tokens_seen": 24942888,
      "step": 38110
    },
    {
      "epoch": 19.97641509433962,
      "grad_norm": 0.7125320434570312,
      "learning_rate": 2.2132120700291136e-10,
      "loss": 0.4818,
      "num_input_tokens_seen": 24946408,
      "step": 38115
    },
    {
      "epoch": 19.979035639413,
      "grad_norm": 0.7221015095710754,
      "learning_rate": 1.7582280804340213e-10,
      "loss": 0.653,
      "num_input_tokens_seen": 24948968,
      "step": 38120
    },
    {
      "epoch": 19.981656184486372,
      "grad_norm": 0.4286576807498932,
      "learning_rate": 1.3555408708842977e-10,
      "loss": 0.3699,
      "num_input_tokens_seen": 24951752,
      "step": 38125
    },
    {
      "epoch": 19.98427672955975,
      "grad_norm": 0.6260583400726318,
      "learning_rate": 1.0051505256458704e-10,
      "loss": 0.4542,
      "num_input_tokens_seen": 24954696,
      "step": 38130
    },
    {
      "epoch": 19.986897274633122,
      "grad_norm": 0.5100289583206177,
      "learning_rate": 7.070571179657037e-11,
      "loss": 0.4797,
      "num_input_tokens_seen": 24959048,
      "step": 38135
    },
    {
      "epoch": 19.9895178197065,
      "grad_norm": 0.30811142921447754,
      "learning_rate": 4.612607102383315e-11,
      "loss": 0.3115,
      "num_input_tokens_seen": 24961640,
      "step": 38140
    },
    {
      "epoch": 19.992138364779873,
      "grad_norm": 0.7547236680984497,
      "learning_rate": 2.677613538670798e-11,
      "loss": 0.3854,
      "num_input_tokens_seen": 24963976,
      "step": 38145
    },
    {
      "epoch": 19.99475890985325,
      "grad_norm": 0.4806714653968811,
      "learning_rate": 1.2655908931957782e-11,
      "loss": 0.4028,
      "num_input_tokens_seen": 24966760,
      "step": 38150
    },
    {
      "epoch": 19.997379454926623,
      "grad_norm": 0.35909751057624817,
      "learning_rate": 3.765394615551366e-12,
      "loss": 0.3566,
      "num_input_tokens_seen": 24970216,
      "step": 38155
    },
    {
      "epoch": 20.0,
      "grad_norm": 7.554598331451416,
      "learning_rate": 1.0459429711229262e-13,
      "loss": 0.7972,
      "num_input_tokens_seen": 24973864,
      "step": 38160
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.4470931589603424,
      "eval_runtime": 9.6228,
      "eval_samples_per_second": 88.124,
      "eval_steps_per_second": 22.031,
      "num_input_tokens_seen": 24973864,
      "step": 38160
    },
    {
      "epoch": 20.0,
      "num_input_tokens_seen": 24973864,
      "step": 38160,
      "total_flos": 1.1246017027819438e+18,
      "train_loss": 0.48482324835264456,
      "train_runtime": 6757.3789,
      "train_samples_per_second": 22.58,
      "train_steps_per_second": 5.647
    }
  ],
  "logging_steps": 5,
  "max_steps": 38160,
  "num_input_tokens_seen": 24973864,
  "num_train_epochs": 20,
  "save_steps": 1908,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1246017027819438e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}