{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 3084,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00486677211339579,
      "grad_norm": 4.565913677215576,
      "learning_rate": 1.0752688172043011e-06,
      "loss": 1.048,
      "mean_token_accuracy": 0.7385952278971673,
      "num_tokens": 10423682.0,
      "step": 5
    },
    {
      "epoch": 0.00973354422679158,
      "grad_norm": 1.151016354560852,
      "learning_rate": 2.4193548387096776e-06,
      "loss": 0.9962,
      "mean_token_accuracy": 0.7441306099295616,
      "num_tokens": 20833859.0,
      "step": 10
    },
    {
      "epoch": 0.01460031634018737,
      "grad_norm": 1.1799436807632446,
      "learning_rate": 3.763440860215054e-06,
      "loss": 0.9571,
      "mean_token_accuracy": 0.7468769297003746,
      "num_tokens": 31186166.0,
      "step": 15
    },
    {
      "epoch": 0.01946708845358316,
      "grad_norm": 0.5721552968025208,
      "learning_rate": 5.1075268817204305e-06,
      "loss": 0.9432,
      "mean_token_accuracy": 0.7481196910142899,
      "num_tokens": 41592551.0,
      "step": 20
    },
    {
      "epoch": 0.024333860566978952,
      "grad_norm": 1.9991700649261475,
      "learning_rate": 6.451612903225806e-06,
      "loss": 0.907,
      "mean_token_accuracy": 0.7548348233103752,
      "num_tokens": 51983936.0,
      "step": 25
    },
    {
      "epoch": 0.02920063268037474,
      "grad_norm": 0.3901795446872711,
      "learning_rate": 7.795698924731183e-06,
      "loss": 0.9084,
      "mean_token_accuracy": 0.7536361366510391,
      "num_tokens": 62413402.0,
      "step": 30
    },
    {
      "epoch": 0.03406740479377053,
      "grad_norm": 0.29165220260620117,
      "learning_rate": 9.13978494623656e-06,
      "loss": 0.8855,
      "mean_token_accuracy": 0.7587963625788688,
      "num_tokens": 72822580.0,
      "step": 35
    },
    {
      "epoch": 0.03893417690716632,
      "grad_norm": 0.27156564593315125,
      "learning_rate": 1.0483870967741936e-05,
      "loss": 0.8736,
      "mean_token_accuracy": 0.7609235867857933,
      "num_tokens": 83206353.0,
      "step": 40
    },
    {
      "epoch": 0.04380094902056211,
      "grad_norm": 0.23284034430980682,
      "learning_rate": 1.1827956989247313e-05,
      "loss": 0.8583,
      "mean_token_accuracy": 0.7639097854495048,
      "num_tokens": 93603538.0,
      "step": 45
    },
    {
      "epoch": 0.048667721133957904,
      "grad_norm": 0.21157187223434448,
      "learning_rate": 1.3172043010752688e-05,
      "loss": 0.8444,
      "mean_token_accuracy": 0.7671671479940414,
      "num_tokens": 104016442.0,
      "step": 50
    },
    {
      "epoch": 0.053534493247353696,
      "grad_norm": 0.20655414462089539,
      "learning_rate": 1.4516129032258066e-05,
      "loss": 0.8505,
      "mean_token_accuracy": 0.7648366779088974,
      "num_tokens": 114403531.0,
      "step": 55
    },
    {
      "epoch": 0.05840126536074948,
      "grad_norm": 0.22793865203857422,
      "learning_rate": 1.586021505376344e-05,
      "loss": 0.8357,
      "mean_token_accuracy": 0.7678958594799041,
      "num_tokens": 124822027.0,
      "step": 60
    },
    {
      "epoch": 0.06326803747414528,
      "grad_norm": 0.22569240629673004,
      "learning_rate": 1.7204301075268818e-05,
      "loss": 0.835,
      "mean_token_accuracy": 0.7676379770040512,
      "num_tokens": 135174509.0,
      "step": 65
    },
    {
      "epoch": 0.06813480958754106,
      "grad_norm": 0.21695566177368164,
      "learning_rate": 1.8548387096774193e-05,
      "loss": 0.828,
      "mean_token_accuracy": 0.7689614772796631,
      "num_tokens": 145595035.0,
      "step": 70
    },
    {
      "epoch": 0.07300158170093686,
      "grad_norm": 0.2320525050163269,
      "learning_rate": 1.989247311827957e-05,
      "loss": 0.8147,
      "mean_token_accuracy": 0.7722090169787407,
      "num_tokens": 155988706.0,
      "step": 75
    },
    {
      "epoch": 0.07786835381433264,
      "grad_norm": 0.2457829862833023,
      "learning_rate": 2.1236559139784946e-05,
      "loss": 0.8227,
      "mean_token_accuracy": 0.7699405491352082,
      "num_tokens": 166390326.0,
      "step": 80
    },
    {
      "epoch": 0.08273512592772843,
      "grad_norm": 0.6996296644210815,
      "learning_rate": 2.258064516129032e-05,
      "loss": 0.8105,
      "mean_token_accuracy": 0.7727881297469139,
      "num_tokens": 176814765.0,
      "step": 85
    },
    {
      "epoch": 0.08760189804112423,
      "grad_norm": 1.133107304573059,
      "learning_rate": 2.39247311827957e-05,
      "loss": 0.8416,
      "mean_token_accuracy": 0.766677676141262,
      "num_tokens": 187238817.0,
      "step": 90
    },
    {
      "epoch": 0.09246867015452001,
      "grad_norm": 1.144582986831665,
      "learning_rate": 2.5268817204301075e-05,
      "loss": 0.829,
      "mean_token_accuracy": 0.7678542837500573,
      "num_tokens": 197639228.0,
      "step": 95
    },
    {
      "epoch": 0.09733544226791581,
      "grad_norm": 0.8101761937141418,
      "learning_rate": 2.661290322580645e-05,
      "loss": 0.8327,
      "mean_token_accuracy": 0.7665805131196975,
      "num_tokens": 208027471.0,
      "step": 100
    },
    {
      "epoch": 0.10220221438131159,
      "grad_norm": 0.5259750485420227,
      "learning_rate": 2.7956989247311828e-05,
      "loss": 0.8037,
      "mean_token_accuracy": 0.77409897595644,
      "num_tokens": 218410390.0,
      "step": 105
    },
    {
      "epoch": 0.10706898649470739,
      "grad_norm": 0.32981541752815247,
      "learning_rate": 2.9301075268817207e-05,
      "loss": 0.8081,
      "mean_token_accuracy": 0.7720122143626214,
      "num_tokens": 228796552.0,
      "step": 110
    },
    {
      "epoch": 0.11193575860810318,
      "grad_norm": 0.25134915113449097,
      "learning_rate": 3.0645161290322585e-05,
      "loss": 0.8084,
      "mean_token_accuracy": 0.7717980936169624,
      "num_tokens": 239217887.0,
      "step": 115
    },
    {
      "epoch": 0.11680253072149896,
      "grad_norm": 0.28814420104026794,
      "learning_rate": 3.198924731182796e-05,
      "loss": 0.8012,
      "mean_token_accuracy": 0.7738058164715766,
      "num_tokens": 249626629.0,
      "step": 120
    },
    {
      "epoch": 0.12166930283489476,
      "grad_norm": 0.3738144636154175,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.7886,
      "mean_token_accuracy": 0.7770566537976265,
      "num_tokens": 260025918.0,
      "step": 125
    },
    {
      "epoch": 0.12653607494829056,
      "grad_norm": 0.3261353671550751,
      "learning_rate": 3.467741935483872e-05,
      "loss": 0.8015,
      "mean_token_accuracy": 0.7732368603348732,
      "num_tokens": 270425999.0,
      "step": 130
    },
    {
      "epoch": 0.13140284706168634,
      "grad_norm": 0.32858020067214966,
      "learning_rate": 3.602150537634409e-05,
      "loss": 0.801,
      "mean_token_accuracy": 0.7732197135686875,
      "num_tokens": 280820700.0,
      "step": 135
    },
    {
      "epoch": 0.13626961917508212,
      "grad_norm": 0.9234886169433594,
      "learning_rate": 3.736559139784947e-05,
      "loss": 0.8428,
      "mean_token_accuracy": 0.7741402730345726,
      "num_tokens": 291229725.0,
      "step": 140
    },
    {
      "epoch": 0.1411363912884779,
      "grad_norm": 0.9446612000465393,
      "learning_rate": 3.870967741935484e-05,
      "loss": 0.7863,
      "mean_token_accuracy": 0.7769969150424003,
      "num_tokens": 301637168.0,
      "step": 145
    },
    {
      "epoch": 0.14600316340187372,
      "grad_norm": 0.4534112811088562,
      "learning_rate": 4.005376344086022e-05,
      "loss": 0.7999,
      "mean_token_accuracy": 0.7738562002778053,
      "num_tokens": 312023441.0,
      "step": 150
    },
    {
      "epoch": 0.1508699355152695,
      "grad_norm": 0.5745447874069214,
      "learning_rate": 4.13978494623656e-05,
      "loss": 0.795,
      "mean_token_accuracy": 0.7760953888297081,
      "num_tokens": 322404386.0,
      "step": 155
    },
    {
      "epoch": 0.1557367076286653,
      "grad_norm": 14.073369026184082,
      "learning_rate": 4.2741935483870973e-05,
      "loss": 0.8243,
      "mean_token_accuracy": 0.7716104313731194,
      "num_tokens": 332833713.0,
      "step": 160
    },
    {
      "epoch": 0.16060347974206107,
      "grad_norm": 0.6174756288528442,
      "learning_rate": 4.408602150537635e-05,
      "loss": 0.7923,
      "mean_token_accuracy": 0.7748784720897675,
      "num_tokens": 343202956.0,
      "step": 165
    },
    {
      "epoch": 0.16547025185545686,
      "grad_norm": 8.148024559020996,
      "learning_rate": 4.543010752688172e-05,
      "loss": 0.8136,
      "mean_token_accuracy": 0.7737636163830757,
      "num_tokens": 353608974.0,
      "step": 170
    },
    {
      "epoch": 0.17033702396885267,
      "grad_norm": 1.5002018213272095,
      "learning_rate": 4.67741935483871e-05,
      "loss": 0.8019,
      "mean_token_accuracy": 0.7733437150716782,
      "num_tokens": 364038179.0,
      "step": 175
    },
    {
      "epoch": 0.17520379608224845,
      "grad_norm": 0.7722571492195129,
      "learning_rate": 4.811827956989248e-05,
      "loss": 0.7944,
      "mean_token_accuracy": 0.7742587998509407,
      "num_tokens": 374446297.0,
      "step": 180
    },
    {
      "epoch": 0.18007056819564424,
      "grad_norm": 0.641341507434845,
      "learning_rate": 4.9462365591397855e-05,
      "loss": 0.7904,
      "mean_token_accuracy": 0.7753829434514046,
      "num_tokens": 384873681.0,
      "step": 185
    },
    {
      "epoch": 0.18493734030904002,
      "grad_norm": 28.4090633392334,
      "learning_rate": 4.999986779279533e-05,
      "loss": 0.8188,
      "mean_token_accuracy": 0.7722472876310349,
      "num_tokens": 395267552.0,
      "step": 190
    },
    {
      "epoch": 0.1898041124224358,
      "grad_norm": 1.1488662958145142,
      "learning_rate": 4.9999059864941726e-05,
      "loss": 0.8467,
      "mean_token_accuracy": 0.77043496966362,
      "num_tokens": 405667550.0,
      "step": 195
    },
    {
      "epoch": 0.19467088453583162,
      "grad_norm": 0.5395088195800781,
      "learning_rate": 4.9997517481389156e-05,
      "loss": 0.8009,
      "mean_token_accuracy": 0.7724558785557747,
      "num_tokens": 416025514.0,
      "step": 200
    },
    {
      "epoch": 0.1995376566492274,
      "grad_norm": 0.417236328125,
      "learning_rate": 4.999524068745182e-05,
      "loss": 0.7876,
      "mean_token_accuracy": 0.7769473195075989,
      "num_tokens": 426421376.0,
      "step": 205
    },
    {
      "epoch": 0.20440442876262319,
      "grad_norm": 0.7154222726821899,
      "learning_rate": 4.999222955002041e-05,
      "loss": 0.7743,
      "mean_token_accuracy": 0.7791593804955482,
      "num_tokens": 436833923.0,
      "step": 210
    },
    {
      "epoch": 0.20927120087601897,
      "grad_norm": 0.3581502437591553,
      "learning_rate": 4.9988484157560136e-05,
      "loss": 0.8044,
      "mean_token_accuracy": 0.7741793766617775,
      "num_tokens": 447241278.0,
      "step": 215
    },
    {
      "epoch": 0.21413797298941478,
      "grad_norm": 0.31207746267318726,
      "learning_rate": 4.998400462010812e-05,
      "loss": 0.7779,
      "mean_token_accuracy": 0.7786757230758667,
      "num_tokens": 457635748.0,
      "step": 220
    },
    {
      "epoch": 0.21900474510281057,
      "grad_norm": 3.3805294036865234,
      "learning_rate": 4.9978791069270184e-05,
      "loss": 0.7813,
      "mean_token_accuracy": 0.777130514383316,
      "num_tokens": 468049627.0,
      "step": 225
    },
    {
      "epoch": 0.22387151721620635,
      "grad_norm": 0.7367106080055237,
      "learning_rate": 4.997284365821699e-05,
      "loss": 0.7921,
      "mean_token_accuracy": 0.7746963158249855,
      "num_tokens": 478447755.0,
      "step": 230
    },
    {
      "epoch": 0.22873828932960213,
      "grad_norm": 0.619115948677063,
      "learning_rate": 4.9966162561679486e-05,
      "loss": 0.7916,
      "mean_token_accuracy": 0.7771365717053413,
      "num_tokens": 488826088.0,
      "step": 235
    },
    {
      "epoch": 0.23360506144299792,
      "grad_norm": 0.7343901991844177,
      "learning_rate": 4.995874797594384e-05,
      "loss": 0.7896,
      "mean_token_accuracy": 0.7753253310918808,
      "num_tokens": 499249307.0,
      "step": 240
    },
    {
      "epoch": 0.23847183355639373,
      "grad_norm": 0.5608010292053223,
      "learning_rate": 4.9950600118845644e-05,
      "loss": 0.7772,
      "mean_token_accuracy": 0.7777869388461113,
      "num_tokens": 509653923.0,
      "step": 245
    },
    {
      "epoch": 0.24333860566978952,
      "grad_norm": 0.48460304737091064,
      "learning_rate": 4.994171922976348e-05,
      "loss": 0.7772,
      "mean_token_accuracy": 0.7779491126537323,
      "num_tokens": 520029552.0,
      "step": 250
    },
    {
      "epoch": 0.2482053777831853,
      "grad_norm": 0.2827472388744354,
      "learning_rate": 4.993210556961197e-05,
      "loss": 0.7629,
      "mean_token_accuracy": 0.7814263075590133,
      "num_tokens": 530375103.0,
      "step": 255
    },
    {
      "epoch": 0.2530721498965811,
      "grad_norm": 0.40675151348114014,
      "learning_rate": 4.9921759420834e-05,
      "loss": 0.7656,
      "mean_token_accuracy": 0.7804664924740792,
      "num_tokens": 540790537.0,
      "step": 260
    },
    {
      "epoch": 0.25793892200997687,
      "grad_norm": 0.31718799471855164,
      "learning_rate": 4.9910681087392554e-05,
      "loss": 0.7762,
      "mean_token_accuracy": 0.7776970952749253,
      "num_tokens": 551218244.0,
      "step": 265
    },
    {
      "epoch": 0.2628056941233727,
      "grad_norm": 0.22272515296936035,
      "learning_rate": 4.9898870894761664e-05,
      "loss": 0.7552,
      "mean_token_accuracy": 0.7830388635396958,
      "num_tokens": 561630527.0,
      "step": 270
    },
    {
      "epoch": 0.26767246623676844,
      "grad_norm": 0.23299150168895721,
      "learning_rate": 4.98863291899169e-05,
      "loss": 0.7732,
      "mean_token_accuracy": 0.7783086940646171,
      "num_tokens": 572051930.0,
      "step": 275
    },
    {
      "epoch": 0.27253923835016425,
      "grad_norm": 0.22763724625110626,
      "learning_rate": 4.987305634132519e-05,
      "loss": 0.7703,
      "mean_token_accuracy": 0.7788772046566009,
      "num_tokens": 582446576.0,
      "step": 280
    },
    {
      "epoch": 0.27740601046356006,
      "grad_norm": 0.22855904698371887,
      "learning_rate": 4.9859052738933966e-05,
      "loss": 0.763,
      "mean_token_accuracy": 0.7805704891681671,
      "num_tokens": 592857472.0,
      "step": 285
    },
    {
      "epoch": 0.2822727825769558,
      "grad_norm": 0.34285151958465576,
      "learning_rate": 4.984431879415973e-05,
      "loss": 0.7676,
      "mean_token_accuracy": 0.7797240734100341,
      "num_tokens": 603270638.0,
      "step": 290
    },
    {
      "epoch": 0.28713955469035163,
      "grad_norm": 10.759557723999023,
      "learning_rate": 4.9828854939875945e-05,
      "loss": 0.8007,
      "mean_token_accuracy": 0.7734850823879242,
      "num_tokens": 613640191.0,
      "step": 295
    },
    {
      "epoch": 0.29200632680374744,
      "grad_norm": 0.9145006537437439,
      "learning_rate": 4.981266163040033e-05,
      "loss": 0.7897,
      "mean_token_accuracy": 0.7749002411961555,
      "num_tokens": 624029969.0,
      "step": 300
    },
    {
      "epoch": 0.2968730989171432,
      "grad_norm": 1.5995612144470215,
      "learning_rate": 4.979573934148152e-05,
      "loss": 0.8052,
      "mean_token_accuracy": 0.7723623231053353,
      "num_tokens": 634449094.0,
      "step": 305
    },
    {
      "epoch": 0.301739871030539,
      "grad_norm": 0.7806094884872437,
      "learning_rate": 4.977808857028506e-05,
      "loss": 0.833,
      "mean_token_accuracy": 0.7730444490909576,
      "num_tokens": 644860189.0,
      "step": 310
    },
    {
      "epoch": 0.30660664314393477,
      "grad_norm": 0.7272294163703918,
      "learning_rate": 4.975970983537884e-05,
      "loss": 0.7875,
      "mean_token_accuracy": 0.778775192797184,
      "num_tokens": 655274237.0,
      "step": 315
    },
    {
      "epoch": 0.3114734152573306,
      "grad_norm": 0.8196333646774292,
      "learning_rate": 4.974060367671783e-05,
      "loss": 0.8,
      "mean_token_accuracy": 0.7742534950375557,
      "num_tokens": 665682997.0,
      "step": 320
    },
    {
      "epoch": 0.3163401873707264,
      "grad_norm": 1.1655899286270142,
      "learning_rate": 4.972077065562821e-05,
      "loss": 0.7803,
      "mean_token_accuracy": 0.7760325714945793,
      "num_tokens": 676065716.0,
      "step": 325
    },
    {
      "epoch": 0.32120695948412215,
      "grad_norm": 1.2836239337921143,
      "learning_rate": 4.970021135479093e-05,
      "loss": 0.7671,
      "mean_token_accuracy": 0.7795489281415939,
      "num_tokens": 686454897.0,
      "step": 330
    },
    {
      "epoch": 0.32607373159751796,
      "grad_norm": 1.1885932683944702,
      "learning_rate": 4.967892637822451e-05,
      "loss": 0.8185,
      "mean_token_accuracy": 0.7795033991336823,
      "num_tokens": 696854897.0,
      "step": 335
    },
    {
      "epoch": 0.3309405037109137,
      "grad_norm": 0.6116691827774048,
      "learning_rate": 4.9656916351267375e-05,
      "loss": 0.7696,
      "mean_token_accuracy": 0.7788765758275986,
      "num_tokens": 707230985.0,
      "step": 340
    },
    {
      "epoch": 0.3358072758243095,
      "grad_norm": 0.48296016454696655,
      "learning_rate": 4.963418192055942e-05,
      "loss": 0.7979,
      "mean_token_accuracy": 0.7761026263237,
      "num_tokens": 717664000.0,
      "step": 345
    },
    {
      "epoch": 0.34067404793770534,
      "grad_norm": 0.7440231442451477,
      "learning_rate": 4.961072375402305e-05,
      "loss": 0.7957,
      "mean_token_accuracy": 0.7736454382538795,
      "num_tokens": 728060607.0,
      "step": 350
    },
    {
      "epoch": 0.3455408200511011,
      "grad_norm": 0.4372887909412384,
      "learning_rate": 4.958654254084355e-05,
      "loss": 0.7729,
      "mean_token_accuracy": 0.7781047075986862,
      "num_tokens": 738480171.0,
      "step": 355
    },
    {
      "epoch": 0.3504075921644969,
      "grad_norm": 0.3816658556461334,
      "learning_rate": 4.956163899144884e-05,
      "loss": 0.7644,
      "mean_token_accuracy": 0.7801351130008698,
      "num_tokens": 748882420.0,
      "step": 360
    },
    {
      "epoch": 0.35527436427789266,
      "grad_norm": 0.28859609365463257,
      "learning_rate": 4.953601383748857e-05,
      "loss": 0.7566,
      "mean_token_accuracy": 0.782233715057373,
      "num_tokens": 759289500.0,
      "step": 365
    },
    {
      "epoch": 0.3601411363912885,
      "grad_norm": 1.1247742176055908,
      "learning_rate": 4.950966783181267e-05,
      "loss": 0.8039,
      "mean_token_accuracy": 0.7757949098944664,
      "num_tokens": 769717709.0,
      "step": 370
    },
    {
      "epoch": 0.3650079085046843,
      "grad_norm": 0.3605007231235504,
      "learning_rate": 4.9482601748449204e-05,
      "loss": 0.75,
      "mean_token_accuracy": 0.7840093672275543,
      "num_tokens": 780126203.0,
      "step": 375
    },
    {
      "epoch": 0.36987468061808004,
      "grad_norm": 0.23982560634613037,
      "learning_rate": 4.9454816382581645e-05,
      "loss": 0.7565,
      "mean_token_accuracy": 0.782028640806675,
      "num_tokens": 790525180.0,
      "step": 380
    },
    {
      "epoch": 0.37474145273147585,
      "grad_norm": 0.33285361528396606,
      "learning_rate": 4.942631255052551e-05,
      "loss": 0.772,
      "mean_token_accuracy": 0.7806572422385216,
      "num_tokens": 800894774.0,
      "step": 385
    },
    {
      "epoch": 0.3796082248448716,
      "grad_norm": 2.9638891220092773,
      "learning_rate": 4.9397091089704364e-05,
      "loss": 0.7753,
      "mean_token_accuracy": 0.7812866613268852,
      "num_tokens": 811273888.0,
      "step": 390
    },
    {
      "epoch": 0.3844749969582674,
      "grad_norm": 0.2870604991912842,
      "learning_rate": 4.936715285862523e-05,
      "loss": 0.765,
      "mean_token_accuracy": 0.7797171220183372,
      "num_tokens": 821660995.0,
      "step": 395
    },
    {
      "epoch": 0.38934176907166324,
      "grad_norm": 0.2231660783290863,
      "learning_rate": 4.9336498736853347e-05,
      "loss": 0.7607,
      "mean_token_accuracy": 0.7802403301000596,
      "num_tokens": 832075350.0,
      "step": 400
    },
    {
      "epoch": 0.394208541185059,
      "grad_norm": 0.3178626596927643,
      "learning_rate": 4.930512962498638e-05,
      "loss": 0.7695,
      "mean_token_accuracy": 0.7803532049059868,
      "num_tokens": 842504687.0,
      "step": 405
    },
    {
      "epoch": 0.3990753132984548,
      "grad_norm": 0.42232292890548706,
      "learning_rate": 4.927304644462789e-05,
      "loss": 0.7608,
      "mean_token_accuracy": 0.7807919234037399,
      "num_tokens": 852900369.0,
      "step": 410
    },
    {
      "epoch": 0.4039420854118506,
      "grad_norm": 0.4365263879299164,
      "learning_rate": 4.924025013836031e-05,
      "loss": 0.7649,
      "mean_token_accuracy": 0.7792169392108917,
      "num_tokens": 863321133.0,
      "step": 415
    },
    {
      "epoch": 0.40880885752524637,
      "grad_norm": 0.5260758996009827,
      "learning_rate": 4.920674166971725e-05,
      "loss": 0.7899,
      "mean_token_accuracy": 0.778005576133728,
      "num_tokens": 873746682.0,
      "step": 420
    },
    {
      "epoch": 0.4136756296386422,
      "grad_norm": 0.8142474293708801,
      "learning_rate": 4.9172522023155154e-05,
      "loss": 0.7547,
      "mean_token_accuracy": 0.7825237363576889,
      "num_tokens": 884167352.0,
      "step": 425
    },
    {
      "epoch": 0.41854240175203794,
      "grad_norm": 18.76129722595215,
      "learning_rate": 4.91375922040244e-05,
      "loss": 0.7422,
      "mean_token_accuracy": 0.7856142908334732,
      "num_tokens": 894528889.0,
      "step": 430
    },
    {
      "epoch": 0.42340917386543375,
      "grad_norm": 0.37281712889671326,
      "learning_rate": 4.9101953238539775e-05,
      "loss": 0.7601,
      "mean_token_accuracy": 0.7817112803459167,
      "num_tokens": 904953157.0,
      "step": 435
    },
    {
      "epoch": 0.42827594597882956,
      "grad_norm": 0.8983141779899597,
      "learning_rate": 4.90656061737503e-05,
      "loss": 0.9106,
      "mean_token_accuracy": 0.7751228958368301,
      "num_tokens": 915362194.0,
      "step": 440
    },
    {
      "epoch": 0.4331427180922253,
      "grad_norm": 0.7746883034706116,
      "learning_rate": 4.90285520775085e-05,
      "loss": 0.7598,
      "mean_token_accuracy": 0.7802918389439583,
      "num_tokens": 925781535.0,
      "step": 445
    },
    {
      "epoch": 0.43800949020562113,
      "grad_norm": 31.000160217285156,
      "learning_rate": 4.899079203843898e-05,
      "loss": 0.7903,
      "mean_token_accuracy": 0.7824798539280892,
      "num_tokens": 936182229.0,
      "step": 450
    },
    {
      "epoch": 0.4428762623190169,
      "grad_norm": 0.6914479732513428,
      "learning_rate": 4.895232716590651e-05,
      "loss": 0.7636,
      "mean_token_accuracy": 0.7789355307817459,
      "num_tokens": 946579678.0,
      "step": 455
    },
    {
      "epoch": 0.4477430344324127,
      "grad_norm": 0.7187318205833435,
      "learning_rate": 4.8913158589983374e-05,
      "loss": 0.7708,
      "mean_token_accuracy": 0.7783745303750038,
      "num_tokens": 956950464.0,
      "step": 460
    },
    {
      "epoch": 0.4526098065458085,
      "grad_norm": 0.4809802770614624,
      "learning_rate": 4.887328746141619e-05,
      "loss": 0.755,
      "mean_token_accuracy": 0.7818428546190261,
      "num_tokens": 967336916.0,
      "step": 465
    },
    {
      "epoch": 0.45747657865920427,
      "grad_norm": 20.98186492919922,
      "learning_rate": 4.8832714951592127e-05,
      "loss": 0.7934,
      "mean_token_accuracy": 0.780860711634159,
      "num_tokens": 977712134.0,
      "step": 470
    },
    {
      "epoch": 0.4623433507726001,
      "grad_norm": 0.7626291513442993,
      "learning_rate": 4.879144225250445e-05,
      "loss": 0.7542,
      "mean_token_accuracy": 0.7819822371006012,
      "num_tokens": 988101388.0,
      "step": 475
    },
    {
      "epoch": 0.46721012288599584,
      "grad_norm": 0.6073904037475586,
      "learning_rate": 4.8749470576717516e-05,
      "loss": 0.7461,
      "mean_token_accuracy": 0.7837938621640206,
      "num_tokens": 998496201.0,
      "step": 480
    },
    {
      "epoch": 0.47207689499939165,
      "grad_norm": 0.792762041091919,
      "learning_rate": 4.8706801157331154e-05,
      "loss": 0.7679,
      "mean_token_accuracy": 0.7787380769848824,
      "num_tokens": 1008904560.0,
      "step": 485
    },
    {
      "epoch": 0.47694366711278746,
      "grad_norm": 0.6579028964042664,
      "learning_rate": 4.866343524794444e-05,
      "loss": 0.7567,
      "mean_token_accuracy": 0.7814577534794808,
      "num_tokens": 1019265291.0,
      "step": 490
    },
    {
      "epoch": 0.4818104392261832,
      "grad_norm": 0.5834213495254517,
      "learning_rate": 4.8619374122618854e-05,
      "loss": 0.7545,
      "mean_token_accuracy": 0.7816786438226699,
      "num_tokens": 1029677022.0,
      "step": 495
    },
    {
      "epoch": 0.48667721133957903,
      "grad_norm": 0.6049254536628723,
      "learning_rate": 4.857461907584086e-05,
      "loss": 0.7727,
      "mean_token_accuracy": 0.7814193114638328,
      "num_tokens": 1040099304.0,
      "step": 500
    },
    {
      "epoch": 0.49154398345297484,
      "grad_norm": 0.556572437286377,
      "learning_rate": 4.8529171422483876e-05,
      "loss": 0.7389,
      "mean_token_accuracy": 0.7864722475409508,
      "num_tokens": 1050525003.0,
      "step": 505
    },
    {
      "epoch": 0.4964107555663706,
      "grad_norm": 0.5724960565567017,
      "learning_rate": 4.848303249776963e-05,
      "loss": 0.7471,
      "mean_token_accuracy": 0.7848008334636688,
      "num_tokens": 1060933707.0,
      "step": 510
    },
    {
      "epoch": 0.5012775276797664,
      "grad_norm": 0.39471882581710815,
      "learning_rate": 4.843620365722894e-05,
      "loss": 0.7603,
      "mean_token_accuracy": 0.782171542942524,
      "num_tokens": 1071341247.0,
      "step": 515
    },
    {
      "epoch": 0.5061442997931622,
      "grad_norm": 0.2869061529636383,
      "learning_rate": 4.838868627666191e-05,
      "loss": 0.757,
      "mean_token_accuracy": 0.7805218636989594,
      "num_tokens": 1081735101.0,
      "step": 520
    },
    {
      "epoch": 0.5110110719065579,
      "grad_norm": 0.2727707028388977,
      "learning_rate": 4.834048175209746e-05,
      "loss": 0.7478,
      "mean_token_accuracy": 0.783265221118927,
      "num_tokens": 1092133745.0,
      "step": 525
    },
    {
      "epoch": 0.5158778440199537,
      "grad_norm": 0.27081528306007385,
      "learning_rate": 4.8291591499752365e-05,
      "loss": 0.7501,
      "mean_token_accuracy": 0.7825851485133171,
      "num_tokens": 1102558362.0,
      "step": 530
    },
    {
      "epoch": 0.5207446161333495,
      "grad_norm": 0.2335483729839325,
      "learning_rate": 4.824201695598963e-05,
      "loss": 0.7412,
      "mean_token_accuracy": 0.7851707175374031,
      "num_tokens": 1112974836.0,
      "step": 535
    },
    {
      "epoch": 0.5256113882467454,
      "grad_norm": 8.846264839172363,
      "learning_rate": 4.819175957727625e-05,
      "loss": 0.7503,
      "mean_token_accuracy": 0.7819222822785378,
      "num_tokens": 1123396739.0,
      "step": 540
    },
    {
      "epoch": 0.5304781603601412,
      "grad_norm": 0.24144458770751953,
      "learning_rate": 4.81408208401405e-05,
      "loss": 0.7531,
      "mean_token_accuracy": 0.781779071688652,
      "num_tokens": 1133783507.0,
      "step": 545
    },
    {
      "epoch": 0.5353449324735369,
      "grad_norm": 0.2682214081287384,
      "learning_rate": 4.808920224112849e-05,
      "loss": 0.7362,
      "mean_token_accuracy": 0.7863743796944618,
      "num_tokens": 1144190109.0,
      "step": 550
    },
    {
      "epoch": 0.5402117045869327,
      "grad_norm": 0.23824985325336456,
      "learning_rate": 4.803690529676019e-05,
      "loss": 0.7319,
      "mean_token_accuracy": 0.787076610326767,
      "num_tokens": 1154605749.0,
      "step": 555
    },
    {
      "epoch": 0.5450784767003285,
      "grad_norm": 0.23254258930683136,
      "learning_rate": 4.7983931543484967e-05,
      "loss": 0.7489,
      "mean_token_accuracy": 0.7822016775608063,
      "num_tokens": 1165020195.0,
      "step": 560
    },
    {
      "epoch": 0.5499452488137243,
      "grad_norm": 0.3700663447380066,
      "learning_rate": 4.793028253763633e-05,
      "loss": 0.759,
      "mean_token_accuracy": 0.7842154935002327,
      "num_tokens": 1175427131.0,
      "step": 565
    },
    {
      "epoch": 0.5548120209271201,
      "grad_norm": 0.25556057691574097,
      "learning_rate": 4.787595985538628e-05,
      "loss": 0.7389,
      "mean_token_accuracy": 0.7848864063620568,
      "num_tokens": 1185813831.0,
      "step": 570
    },
    {
      "epoch": 0.5596787930405159,
      "grad_norm": 0.23745465278625488,
      "learning_rate": 4.7820965092698967e-05,
      "loss": 0.7488,
      "mean_token_accuracy": 0.7832117944955825,
      "num_tokens": 1196227781.0,
      "step": 575
    },
    {
      "epoch": 0.5645455651539116,
      "grad_norm": 0.38585183024406433,
      "learning_rate": 4.776529986528383e-05,
      "loss": 0.7471,
      "mean_token_accuracy": 0.7824306935071945,
      "num_tokens": 1206618364.0,
      "step": 580
    },
    {
      "epoch": 0.5694123372673074,
      "grad_norm": 0.2854125499725342,
      "learning_rate": 4.770896580854811e-05,
      "loss": 0.7434,
      "mean_token_accuracy": 0.783917561173439,
      "num_tokens": 1217012583.0,
      "step": 585
    },
    {
      "epoch": 0.5742791093807033,
      "grad_norm": 0.27572405338287354,
      "learning_rate": 4.765196457754881e-05,
      "loss": 0.7407,
      "mean_token_accuracy": 0.7843161955475807,
      "num_tokens": 1227426592.0,
      "step": 590
    },
    {
      "epoch": 0.5791458814940991,
      "grad_norm": 0.24903604388237,
      "learning_rate": 4.759429784694407e-05,
      "loss": 0.7358,
      "mean_token_accuracy": 0.7858725219964982,
      "num_tokens": 1237850858.0,
      "step": 595
    },
    {
      "epoch": 0.5840126536074949,
      "grad_norm": 0.3491290807723999,
      "learning_rate": 4.7535967310943955e-05,
      "loss": 0.7358,
      "mean_token_accuracy": 0.7857545629143715,
      "num_tokens": 1248252872.0,
      "step": 600
    },
    {
      "epoch": 0.5888794257208906,
      "grad_norm": 0.3528711795806885,
      "learning_rate": 4.7476974683260674e-05,
      "loss": 0.7427,
      "mean_token_accuracy": 0.7841674029827118,
      "num_tokens": 1258669551.0,
      "step": 605
    },
    {
      "epoch": 0.5937461978342864,
      "grad_norm": 0.26817673444747925,
      "learning_rate": 4.741732169705829e-05,
      "loss": 0.743,
      "mean_token_accuracy": 0.7836904942989349,
      "num_tokens": 1269084496.0,
      "step": 610
    },
    {
      "epoch": 0.5986129699476822,
      "grad_norm": 0.2938285768032074,
      "learning_rate": 4.7357010104901725e-05,
      "loss": 0.7566,
      "mean_token_accuracy": 0.7818590044975281,
      "num_tokens": 1279486027.0,
      "step": 615
    },
    {
      "epoch": 0.603479742061078,
      "grad_norm": 0.2213556319475174,
      "learning_rate": 4.729604167870533e-05,
      "loss": 0.7332,
      "mean_token_accuracy": 0.7862257987260819,
      "num_tokens": 1289907411.0,
      "step": 620
    },
    {
      "epoch": 0.6083465141744738,
      "grad_norm": 0.2097499966621399,
      "learning_rate": 4.723441820968076e-05,
      "loss": 0.7319,
      "mean_token_accuracy": 0.7869701266288758,
      "num_tokens": 1300307502.0,
      "step": 625
    },
    {
      "epoch": 0.6132132862878695,
      "grad_norm": 0.21281397342681885,
      "learning_rate": 4.717214150828445e-05,
      "loss": 0.7444,
      "mean_token_accuracy": 0.7830566048622132,
      "num_tokens": 1310709783.0,
      "step": 630
    },
    {
      "epoch": 0.6180800584012653,
      "grad_norm": 0.2568768858909607,
      "learning_rate": 4.710921340416431e-05,
      "loss": 0.7374,
      "mean_token_accuracy": 0.7856349423527718,
      "num_tokens": 1321040165.0,
      "step": 635
    },
    {
      "epoch": 0.6229468305146612,
      "grad_norm": 0.21003931760787964,
      "learning_rate": 4.704563574610606e-05,
      "loss": 0.7362,
      "mean_token_accuracy": 0.7855549201369285,
      "num_tokens": 1331417027.0,
      "step": 640
    },
    {
      "epoch": 0.627813602628057,
      "grad_norm": 0.46499934792518616,
      "learning_rate": 4.698141040197889e-05,
      "loss": 0.7348,
      "mean_token_accuracy": 0.7858194828033447,
      "num_tokens": 1341841944.0,
      "step": 645
    },
    {
      "epoch": 0.6326803747414528,
      "grad_norm": 0.2620554268360138,
      "learning_rate": 4.691653925868054e-05,
      "loss": 0.7367,
      "mean_token_accuracy": 0.7856289654970169,
      "num_tokens": 1352255383.0,
      "step": 650
    },
    {
      "epoch": 0.6375471468548485,
      "grad_norm": 0.2831186354160309,
      "learning_rate": 4.685102422208191e-05,
      "loss": 0.7395,
      "mean_token_accuracy": 0.7842460080981255,
      "num_tokens": 1362616796.0,
      "step": 655
    },
    {
      "epoch": 0.6424139189682443,
      "grad_norm": 5.849782943725586,
      "learning_rate": 4.678486721697107e-05,
      "loss": 0.7456,
      "mean_token_accuracy": 0.783732108771801,
      "num_tokens": 1373008154.0,
      "step": 660
    },
    {
      "epoch": 0.6472806910816401,
      "grad_norm": 0.3637860119342804,
      "learning_rate": 4.6718070186996706e-05,
      "loss": 0.7383,
      "mean_token_accuracy": 0.7853639379143715,
      "num_tokens": 1383403971.0,
      "step": 665
    },
    {
      "epoch": 0.6521474631950359,
      "grad_norm": 0.5078529119491577,
      "learning_rate": 4.665063509461097e-05,
      "loss": 0.7425,
      "mean_token_accuracy": 0.7838774964213371,
      "num_tokens": 1393808183.0,
      "step": 670
    },
    {
      "epoch": 0.6570142353084317,
      "grad_norm": 0.37028664350509644,
      "learning_rate": 4.658256392101189e-05,
      "loss": 0.7297,
      "mean_token_accuracy": 0.7871156334877014,
      "num_tokens": 1404198867.0,
      "step": 675
    },
    {
      "epoch": 0.6618810074218274,
      "grad_norm": 0.29443809390068054,
      "learning_rate": 4.651385866608512e-05,
      "loss": 0.7679,
      "mean_token_accuracy": 0.783281721174717,
      "num_tokens": 1414624972.0,
      "step": 680
    },
    {
      "epoch": 0.6667477795352232,
      "grad_norm": 0.4256887137889862,
      "learning_rate": 4.6444521348345216e-05,
      "loss": 0.7648,
      "mean_token_accuracy": 0.7818196520209313,
      "num_tokens": 1425050094.0,
      "step": 685
    },
    {
      "epoch": 0.671614551648619,
      "grad_norm": 3.1152939796447754,
      "learning_rate": 4.637455400487632e-05,
      "loss": 0.7467,
      "mean_token_accuracy": 0.7846164256334305,
      "num_tokens": 1435427712.0,
      "step": 690
    },
    {
      "epoch": 0.6764813237620149,
      "grad_norm": 3.908926010131836,
      "learning_rate": 4.630395869127226e-05,
      "loss": 0.7453,
      "mean_token_accuracy": 0.7835512667894363,
      "num_tokens": 1445814000.0,
      "step": 695
    },
    {
      "epoch": 0.6813480958754107,
      "grad_norm": 0.2961072325706482,
      "learning_rate": 4.6232737481576296e-05,
      "loss": 0.7485,
      "mean_token_accuracy": 0.7823463201522827,
      "num_tokens": 1456214792.0,
      "step": 700
    },
    {
      "epoch": 0.6862148679888064,
      "grad_norm": 0.4107789695262909,
      "learning_rate": 4.616089246822003e-05,
      "loss": 0.7446,
      "mean_token_accuracy": 0.7831661537289619,
      "num_tokens": 1466621720.0,
      "step": 705
    },
    {
      "epoch": 0.6910816401022022,
      "grad_norm": 0.3386191427707672,
      "learning_rate": 4.6088425761962005e-05,
      "loss": 0.7313,
      "mean_token_accuracy": 0.7867486298084259,
      "num_tokens": 1476992179.0,
      "step": 710
    },
    {
      "epoch": 0.695948412215598,
      "grad_norm": 0.30019065737724304,
      "learning_rate": 4.6015339491825716e-05,
      "loss": 0.7447,
      "mean_token_accuracy": 0.783013978600502,
      "num_tokens": 1487381513.0,
      "step": 715
    },
    {
      "epoch": 0.7008151843289938,
      "grad_norm": 0.29462873935699463,
      "learning_rate": 4.594163580503701e-05,
      "loss": 0.7351,
      "mean_token_accuracy": 0.786101596057415,
      "num_tokens": 1497795964.0,
      "step": 720
    },
    {
      "epoch": 0.7056819564423896,
      "grad_norm": 0.2691968083381653,
      "learning_rate": 4.586731686696102e-05,
      "loss": 0.7385,
      "mean_token_accuracy": 0.7848589658737183,
      "num_tokens": 1508184746.0,
      "step": 725
    },
    {
      "epoch": 0.7105487285557853,
      "grad_norm": 0.24561014771461487,
      "learning_rate": 4.579238486103857e-05,
      "loss": 0.7397,
      "mean_token_accuracy": 0.7841762140393257,
      "num_tokens": 1518615277.0,
      "step": 730
    },
    {
      "epoch": 0.7154155006691811,
      "grad_norm": 0.2730070650577545,
      "learning_rate": 4.5716841988722004e-05,
      "loss": 0.7369,
      "mean_token_accuracy": 0.7851657181978225,
      "num_tokens": 1529032230.0,
      "step": 735
    },
    {
      "epoch": 0.720282272782577,
      "grad_norm": 0.38489723205566406,
      "learning_rate": 4.564069046941049e-05,
      "loss": 0.7271,
      "mean_token_accuracy": 0.7883120805025101,
      "num_tokens": 1539442995.0,
      "step": 740
    },
    {
      "epoch": 0.7251490448959728,
      "grad_norm": 0.4404032230377197,
      "learning_rate": 4.556393254038486e-05,
      "loss": 0.7274,
      "mean_token_accuracy": 0.7880117669701576,
      "num_tokens": 1549866706.0,
      "step": 745
    },
    {
      "epoch": 0.7300158170093686,
      "grad_norm": 0.435287743806839,
      "learning_rate": 4.5486570456741867e-05,
      "loss": 0.735,
      "mean_token_accuracy": 0.7857326343655586,
      "num_tokens": 1560256335.0,
      "step": 750
    },
    {
      "epoch": 0.7348825891227643,
      "grad_norm": 0.3549564480781555,
      "learning_rate": 4.540860649132789e-05,
      "loss": 0.7418,
      "mean_token_accuracy": 0.7839172899723053,
      "num_tokens": 1570644171.0,
      "step": 755
    },
    {
      "epoch": 0.7397493612361601,
      "grad_norm": 0.6766982078552246,
      "learning_rate": 4.533004293467224e-05,
      "loss": 0.7751,
      "mean_token_accuracy": 0.7806326761841774,
      "num_tokens": 1581063421.0,
      "step": 760
    },
    {
      "epoch": 0.7446161333495559,
      "grad_norm": 0.4576626420021057,
      "learning_rate": 4.525088209491977e-05,
      "loss": 0.7399,
      "mean_token_accuracy": 0.784378145635128,
      "num_tokens": 1591444362.0,
      "step": 765
    },
    {
      "epoch": 0.7494829054629517,
      "grad_norm": 0.30135461688041687,
      "learning_rate": 4.517112629776314e-05,
      "loss": 0.7379,
      "mean_token_accuracy": 0.784884712100029,
      "num_tokens": 1601842276.0,
      "step": 770
    },
    {
      "epoch": 0.7543496775763475,
      "grad_norm": 0.37022897601127625,
      "learning_rate": 4.509077788637446e-05,
      "loss": 0.7655,
      "mean_token_accuracy": 0.7841911405324936,
      "num_tokens": 1612263297.0,
      "step": 775
    },
    {
      "epoch": 0.7592164496897432,
      "grad_norm": 0.2393566370010376,
      "learning_rate": 4.500983922133644e-05,
      "loss": 0.7332,
      "mean_token_accuracy": 0.7859902277588844,
      "num_tokens": 1622652062.0,
      "step": 780
    },
    {
      "epoch": 0.764083221803139,
      "grad_norm": 0.22365836799144745,
      "learning_rate": 4.4928312680573064e-05,
      "loss": 0.7337,
      "mean_token_accuracy": 0.7856424614787102,
      "num_tokens": 1633050150.0,
      "step": 785
    },
    {
      "epoch": 0.7689499939165348,
      "grad_norm": 0.4183080196380615,
      "learning_rate": 4.48462006592797e-05,
      "loss": 0.7344,
      "mean_token_accuracy": 0.7858954668045044,
      "num_tokens": 1643440964.0,
      "step": 790
    },
    {
      "epoch": 0.7738167660299307,
      "grad_norm": 0.26855677366256714,
      "learning_rate": 4.4763505569852745e-05,
      "loss": 0.7323,
      "mean_token_accuracy": 0.7858997017145157,
      "num_tokens": 1653856700.0,
      "step": 795
    },
    {
      "epoch": 0.7786835381433265,
      "grad_norm": 0.4860982596874237,
      "learning_rate": 4.468022984181875e-05,
      "loss": 0.7434,
      "mean_token_accuracy": 0.7834003984928131,
      "num_tokens": 1664275372.0,
      "step": 800
    },
    {
      "epoch": 0.7835503102567223,
      "grad_norm": 0.29402992129325867,
      "learning_rate": 4.459637592176304e-05,
      "loss": 0.726,
      "mean_token_accuracy": 0.788159690797329,
      "num_tokens": 1674697804.0,
      "step": 805
    },
    {
      "epoch": 0.788417082370118,
      "grad_norm": 0.8125059604644775,
      "learning_rate": 4.4511946273257846e-05,
      "loss": 0.7564,
      "mean_token_accuracy": 0.7818496063351631,
      "num_tokens": 1685089657.0,
      "step": 810
    },
    {
      "epoch": 0.7932838544835138,
      "grad_norm": 0.22003504633903503,
      "learning_rate": 4.44269433767899e-05,
      "loss": 0.7229,
      "mean_token_accuracy": 0.7888763830065727,
      "num_tokens": 1695504538.0,
      "step": 815
    },
    {
      "epoch": 0.7981506265969096,
      "grad_norm": 0.24877609312534332,
      "learning_rate": 4.4341369729687575e-05,
      "loss": 0.7342,
      "mean_token_accuracy": 0.7855436399579048,
      "num_tokens": 1705912399.0,
      "step": 820
    },
    {
      "epoch": 0.8030173987103054,
      "grad_norm": 0.22476691007614136,
      "learning_rate": 4.425522784604755e-05,
      "loss": 0.7262,
      "mean_token_accuracy": 0.7876876816153526,
      "num_tokens": 1716330715.0,
      "step": 825
    },
    {
      "epoch": 0.8078841708237012,
      "grad_norm": 0.22741760313510895,
      "learning_rate": 4.416852025666088e-05,
      "loss": 0.7214,
      "mean_token_accuracy": 0.7890818163752555,
      "num_tokens": 1726745360.0,
      "step": 830
    },
    {
      "epoch": 0.8127509429370969,
      "grad_norm": 0.1930360645055771,
      "learning_rate": 4.4081249508938684e-05,
      "loss": 0.7334,
      "mean_token_accuracy": 0.7860154449939728,
      "num_tokens": 1737160788.0,
      "step": 835
    },
    {
      "epoch": 0.8176177150504927,
      "grad_norm": 0.18769091367721558,
      "learning_rate": 4.3993418166837295e-05,
      "loss": 0.7227,
      "mean_token_accuracy": 0.7888397544622421,
      "num_tokens": 1747570720.0,
      "step": 840
    },
    {
      "epoch": 0.8224844871638886,
      "grad_norm": 0.1954774260520935,
      "learning_rate": 4.390502881078296e-05,
      "loss": 0.7197,
      "mean_token_accuracy": 0.789542593061924,
      "num_tokens": 1757946948.0,
      "step": 845
    },
    {
      "epoch": 0.8273512592772844,
      "grad_norm": 0.18590693175792694,
      "learning_rate": 4.3816084037595976e-05,
      "loss": 0.7381,
      "mean_token_accuracy": 0.784670104086399,
      "num_tokens": 1768329429.0,
      "step": 850
    },
    {
      "epoch": 0.8322180313906802,
      "grad_norm": 0.1882573664188385,
      "learning_rate": 4.372658646041443e-05,
      "loss": 0.74,
      "mean_token_accuracy": 0.7853763118386269,
      "num_tokens": 1778732187.0,
      "step": 855
    },
    {
      "epoch": 0.8370848035040759,
      "grad_norm": 0.2831585109233856,
      "learning_rate": 4.363653870861742e-05,
      "loss": 0.7239,
      "mean_token_accuracy": 0.7879428192973137,
      "num_tokens": 1789121965.0,
      "step": 860
    },
    {
      "epoch": 0.8419515756174717,
      "grad_norm": 0.2164989709854126,
      "learning_rate": 4.354594342774784e-05,
      "loss": 0.7282,
      "mean_token_accuracy": 0.7873011901974678,
      "num_tokens": 1799508091.0,
      "step": 865
    },
    {
      "epoch": 0.8468183477308675,
      "grad_norm": 0.248453289270401,
      "learning_rate": 4.3454803279434554e-05,
      "loss": 0.7384,
      "mean_token_accuracy": 0.787485171854496,
      "num_tokens": 1809896771.0,
      "step": 870
    },
    {
      "epoch": 0.8516851198442633,
      "grad_norm": 0.31121039390563965,
      "learning_rate": 4.336312094131434e-05,
      "loss": 0.7283,
      "mean_token_accuracy": 0.7869920879602432,
      "num_tokens": 1820277586.0,
      "step": 875
    },
    {
      "epoch": 0.8565518919576591,
      "grad_norm": 0.2747540771961212,
      "learning_rate": 4.3270899106953105e-05,
      "loss": 0.7204,
      "mean_token_accuracy": 0.7891971305012703,
      "num_tokens": 1830696673.0,
      "step": 880
    },
    {
      "epoch": 0.8614186640710548,
      "grad_norm": 549.490966796875,
      "learning_rate": 4.317814048576682e-05,
      "loss": 0.7545,
      "mean_token_accuracy": 0.7851567804813385,
      "num_tokens": 1841110223.0,
      "step": 885
    },
    {
      "epoch": 0.8662854361844506,
      "grad_norm": 0.37894293665885925,
      "learning_rate": 4.3084847802941873e-05,
      "loss": 0.7385,
      "mean_token_accuracy": 0.7846206963062287,
      "num_tokens": 1851517471.0,
      "step": 890
    },
    {
      "epoch": 0.8711522082978465,
      "grad_norm": 0.4155738055706024,
      "learning_rate": 4.299102379935505e-05,
      "loss": 0.7308,
      "mean_token_accuracy": 0.7859715074300766,
      "num_tokens": 1861909391.0,
      "step": 895
    },
    {
      "epoch": 0.8760189804112423,
      "grad_norm": 0.26846975088119507,
      "learning_rate": 4.2896671231492966e-05,
      "loss": 0.7338,
      "mean_token_accuracy": 0.7853391855955124,
      "num_tokens": 1872324199.0,
      "step": 900
    },
    {
      "epoch": 0.8808857525246381,
      "grad_norm": 0.27588847279548645,
      "learning_rate": 4.2801792871371116e-05,
      "loss": 0.7258,
      "mean_token_accuracy": 0.7877420783042908,
      "num_tokens": 1882719391.0,
      "step": 905
    },
    {
      "epoch": 0.8857525246380338,
      "grad_norm": 1.9718985557556152,
      "learning_rate": 4.2706391506452414e-05,
      "loss": 0.7293,
      "mean_token_accuracy": 0.7876736953854561,
      "num_tokens": 1893136016.0,
      "step": 910
    },
    {
      "epoch": 0.8906192967514296,
      "grad_norm": 0.7682464718818665,
      "learning_rate": 4.261046993956531e-05,
      "loss": 0.7608,
      "mean_token_accuracy": 0.7849423885345459,
      "num_tokens": 1903548307.0,
      "step": 915
    },
    {
      "epoch": 0.8954860688648254,
      "grad_norm": 0.4276520907878876,
      "learning_rate": 4.251403098882144e-05,
      "loss": 0.7199,
      "mean_token_accuracy": 0.7894313901662826,
      "num_tokens": 1913935702.0,
      "step": 920
    },
    {
      "epoch": 0.9003528409782212,
      "grad_norm": 0.3141389489173889,
      "learning_rate": 4.241707748753283e-05,
      "loss": 0.7307,
      "mean_token_accuracy": 0.7868300318717957,
      "num_tokens": 1924331800.0,
      "step": 925
    },
    {
      "epoch": 0.905219613091617,
      "grad_norm": 0.2156876027584076,
      "learning_rate": 4.2319612284128644e-05,
      "loss": 0.7315,
      "mean_token_accuracy": 0.785821071267128,
      "num_tokens": 1934730369.0,
      "step": 930
    },
    {
      "epoch": 0.9100863852050127,
      "grad_norm": 1.057871699333191,
      "learning_rate": 4.222163824207155e-05,
      "loss": 0.7369,
      "mean_token_accuracy": 0.7866499245166778,
      "num_tokens": 1945142285.0,
      "step": 935
    },
    {
      "epoch": 0.9149531573184085,
      "grad_norm": 3.8146471977233887,
      "learning_rate": 4.212315823977352e-05,
      "loss": 0.7433,
      "mean_token_accuracy": 0.7840294271707535,
      "num_tokens": 1955525051.0,
      "step": 940
    },
    {
      "epoch": 0.9198199294318044,
      "grad_norm": 0.2783801257610321,
      "learning_rate": 4.202417517051132e-05,
      "loss": 0.7211,
      "mean_token_accuracy": 0.788542227447033,
      "num_tokens": 1965949017.0,
      "step": 945
    },
    {
      "epoch": 0.9246867015452002,
      "grad_norm": 0.3669135570526123,
      "learning_rate": 4.192469194234148e-05,
      "loss": 0.7232,
      "mean_token_accuracy": 0.7881337732076645,
      "num_tokens": 1976357700.0,
      "step": 950
    },
    {
      "epoch": 0.929553473658596,
      "grad_norm": 0.2776182293891907,
      "learning_rate": 4.182471147801485e-05,
      "loss": 0.7335,
      "mean_token_accuracy": 0.7854045987129211,
      "num_tokens": 1986765822.0,
      "step": 955
    },
    {
      "epoch": 0.9344202457719917,
      "grad_norm": 0.35710859298706055,
      "learning_rate": 4.172423671489079e-05,
      "loss": 0.7422,
      "mean_token_accuracy": 0.7867365926504135,
      "num_tokens": 1997194381.0,
      "step": 960
    },
    {
      "epoch": 0.9392870178853875,
      "grad_norm": 0.39433953166007996,
      "learning_rate": 4.162327060485078e-05,
      "loss": 0.7193,
      "mean_token_accuracy": 0.7896064877510071,
      "num_tokens": 2007578618.0,
      "step": 965
    },
    {
      "epoch": 0.9441537899987833,
      "grad_norm": 0.30032631754875183,
      "learning_rate": 4.1521816114211785e-05,
      "loss": 0.726,
      "mean_token_accuracy": 0.7872108325362206,
      "num_tokens": 2017993673.0,
      "step": 970
    },
    {
      "epoch": 0.9490205621121791,
      "grad_norm": 0.2954862415790558,
      "learning_rate": 4.141987622363904e-05,
      "loss": 0.7329,
      "mean_token_accuracy": 0.7862786889076233,
      "num_tokens": 2028397525.0,
      "step": 975
    },
    {
      "epoch": 0.9538873342255749,
      "grad_norm": 0.2835451364517212,
      "learning_rate": 4.1317453928058517e-05,
      "loss": 0.7219,
      "mean_token_accuracy": 0.7887295231223106,
      "num_tokens": 2038803116.0,
      "step": 980
    },
    {
      "epoch": 0.9587541063389706,
      "grad_norm": 0.2586745619773865,
      "learning_rate": 4.12145522365689e-05,
      "loss": 0.7268,
      "mean_token_accuracy": 0.7872585833072663,
      "num_tokens": 2049218915.0,
      "step": 985
    },
    {
      "epoch": 0.9636208784523664,
      "grad_norm": 0.1965721845626831,
      "learning_rate": 4.111117417235325e-05,
      "loss": 0.7221,
      "mean_token_accuracy": 0.7883333384990692,
      "num_tokens": 2059604447.0,
      "step": 990
    },
    {
      "epoch": 0.9684876505657622,
      "grad_norm": 2.610860824584961,
      "learning_rate": 4.100732277259012e-05,
      "loss": 0.7439,
      "mean_token_accuracy": 0.7868512511253357,
      "num_tokens": 2070011199.0,
      "step": 995
    },
    {
      "epoch": 0.9733544226791581,
      "grad_norm": 0.20130938291549683,
      "learning_rate": 4.090300108836432e-05,
      "loss": 0.7237,
      "mean_token_accuracy": 0.7882704392075539,
      "num_tokens": 2080384377.0,
      "step": 1000
    },
    {
      "epoch": 0.9782211947925539,
      "grad_norm": 0.30649039149284363,
      "learning_rate": 4.0798212184577354e-05,
      "loss": 0.719,
      "mean_token_accuracy": 0.7888605058193207,
      "num_tokens": 2090803844.0,
      "step": 1005
    },
    {
      "epoch": 0.9830879669059497,
      "grad_norm": 5.640383243560791,
      "learning_rate": 4.069295913985728e-05,
      "loss": 0.7139,
      "mean_token_accuracy": 0.7909555912017823,
      "num_tokens": 2101206492.0,
      "step": 1010
    },
    {
      "epoch": 0.9879547390193454,
      "grad_norm": 0.31264573335647583,
      "learning_rate": 4.058724504646834e-05,
      "loss": 0.7298,
      "mean_token_accuracy": 0.7865370109677314,
      "num_tokens": 2111593063.0,
      "step": 1015
    },
    {
      "epoch": 0.9928215111327412,
      "grad_norm": 0.23999595642089844,
      "learning_rate": 4.048107301022005e-05,
      "loss": 0.728,
      "mean_token_accuracy": 0.7864951848983764,
      "num_tokens": 2121991314.0,
      "step": 1020
    },
    {
      "epoch": 0.997688283246137,
      "grad_norm": 2.0547049045562744,
      "learning_rate": 4.0374446150376e-05,
      "loss": 0.758,
      "mean_token_accuracy": 0.7861149102449417,
      "num_tokens": 2132385867.0,
      "step": 1025
    },
    {
      "epoch": 1.0019467088453584,
      "grad_norm": 0.25391384959220886,
      "learning_rate": 4.0267367599562185e-05,
      "loss": 0.7027,
      "mean_token_accuracy": 0.7919896006584167,
      "num_tokens": 2141481139.0,
      "step": 1030
    },
    {
      "epoch": 1.0068134809587541,
      "grad_norm": 0.2714090943336487,
      "learning_rate": 4.015984050367499e-05,
      "loss": 0.6863,
      "mean_token_accuracy": 0.7962557137012481,
      "num_tokens": 2151878249.0,
      "step": 1035
    },
    {
      "epoch": 1.0116802530721498,
      "grad_norm": 0.34391292929649353,
      "learning_rate": 4.005186802178873e-05,
      "loss": 0.7462,
      "mean_token_accuracy": 0.7915629178285599,
      "num_tokens": 2162298016.0,
      "step": 1040
    },
    {
      "epoch": 1.0165470251855457,
      "grad_norm": 0.28975221514701843,
      "learning_rate": 3.9943453326062907e-05,
      "loss": 0.7253,
      "mean_token_accuracy": 0.7907896280288697,
      "num_tokens": 2172721397.0,
      "step": 1045
    },
    {
      "epoch": 1.0214137972989414,
      "grad_norm": 0.2151109129190445,
      "learning_rate": 3.9834599601648936e-05,
      "loss": 0.6898,
      "mean_token_accuracy": 0.7955597519874573,
      "num_tokens": 2183114959.0,
      "step": 1050
    },
    {
      "epoch": 1.0262805694123374,
      "grad_norm": 0.3266723155975342,
      "learning_rate": 3.9725310046596595e-05,
      "loss": 0.7321,
      "mean_token_accuracy": 0.7886668682098389,
      "num_tokens": 2193532733.0,
      "step": 1055
    },
    {
      "epoch": 1.031147341525733,
      "grad_norm": 0.30996429920196533,
      "learning_rate": 3.961558787176012e-05,
      "loss": 0.6935,
      "mean_token_accuracy": 0.7946403086185455,
      "num_tokens": 2203935841.0,
      "step": 1060
    },
    {
      "epoch": 1.0360141136391288,
      "grad_norm": 0.47867515683174133,
      "learning_rate": 3.950543630070378e-05,
      "loss": 0.7139,
      "mean_token_accuracy": 0.7928180724382401,
      "num_tokens": 2214341336.0,
      "step": 1065
    },
    {
      "epoch": 1.0408808857525247,
      "grad_norm": 0.2457122802734375,
      "learning_rate": 3.939485856960723e-05,
      "loss": 0.6939,
      "mean_token_accuracy": 0.7945135667920112,
      "num_tokens": 2224762780.0,
      "step": 1070
    },
    {
      "epoch": 1.0457476578659204,
      "grad_norm": 0.2119971662759781,
      "learning_rate": 3.928385792717045e-05,
      "loss": 0.7025,
      "mean_token_accuracy": 0.7920261517167091,
      "num_tokens": 2235161060.0,
      "step": 1075
    },
    {
      "epoch": 1.0506144299793163,
      "grad_norm": 0.6303860545158386,
      "learning_rate": 3.917243763451824e-05,
      "loss": 0.7053,
      "mean_token_accuracy": 0.7940833002328873,
      "num_tokens": 2245533511.0,
      "step": 1080
    },
    {
      "epoch": 1.055481202092712,
      "grad_norm": 0.191057026386261,
      "learning_rate": 3.906060096510446e-05,
      "loss": 0.6882,
      "mean_token_accuracy": 0.7956651702523232,
      "num_tokens": 2255882058.0,
      "step": 1085
    },
    {
      "epoch": 1.0603479742061077,
      "grad_norm": 0.1875150054693222,
      "learning_rate": 3.894835120461584e-05,
      "loss": 0.6989,
      "mean_token_accuracy": 0.7925189465284348,
      "num_tokens": 2266279054.0,
      "step": 1090
    },
    {
      "epoch": 1.0652147463195036,
      "grad_norm": 0.19916220009326935,
      "learning_rate": 3.883569165087544e-05,
      "loss": 0.7079,
      "mean_token_accuracy": 0.7903703838586807,
      "num_tokens": 2276685571.0,
      "step": 1095
    },
    {
      "epoch": 1.0700815184328993,
      "grad_norm": 0.21875645220279694,
      "learning_rate": 3.872262561374579e-05,
      "loss": 0.6922,
      "mean_token_accuracy": 0.7947641432285308,
      "num_tokens": 2287088571.0,
      "step": 1100
    },
    {
      "epoch": 1.0749482905462953,
      "grad_norm": 0.1894306242465973,
      "learning_rate": 3.860915641503161e-05,
      "loss": 0.7108,
      "mean_token_accuracy": 0.793051327764988,
      "num_tokens": 2297510717.0,
      "step": 1105
    },
    {
      "epoch": 1.079815062659691,
      "grad_norm": 0.20602893829345703,
      "learning_rate": 3.849528738838226e-05,
      "loss": 0.6924,
      "mean_token_accuracy": 0.794889298081398,
      "num_tokens": 2307926361.0,
      "step": 1110
    },
    {
      "epoch": 1.0846818347730867,
      "grad_norm": 0.22329486906528473,
      "learning_rate": 3.8381021879193766e-05,
      "loss": 0.7236,
      "mean_token_accuracy": 0.7909580364823341,
      "num_tokens": 2318297393.0,
      "step": 1115
    },
    {
      "epoch": 1.0895486068864826,
      "grad_norm": 0.3012749254703522,
      "learning_rate": 3.826636324451054e-05,
      "loss": 0.7003,
      "mean_token_accuracy": 0.7951581731438637,
      "num_tokens": 2328697529.0,
      "step": 1120
    },
    {
      "epoch": 1.0944153789998783,
      "grad_norm": 0.23902414739131927,
      "learning_rate": 3.815131485292678e-05,
      "loss": 0.6931,
      "mean_token_accuracy": 0.7942418903112411,
      "num_tokens": 2339091854.0,
      "step": 1125
    },
    {
      "epoch": 1.0992821511132742,
      "grad_norm": 0.24394287168979645,
      "learning_rate": 3.803588008448745e-05,
      "loss": 0.6925,
      "mean_token_accuracy": 0.794149973988533,
      "num_tokens": 2349514347.0,
      "step": 1130
    },
    {
      "epoch": 1.10414892322667,
      "grad_norm": 0.20840227603912354,
      "learning_rate": 3.7920062330589044e-05,
      "loss": 0.6943,
      "mean_token_accuracy": 0.794423270225525,
      "num_tokens": 2359917223.0,
      "step": 1135
    },
    {
      "epoch": 1.1090156953400656,
      "grad_norm": 0.20512400567531586,
      "learning_rate": 3.780386499387989e-05,
      "loss": 0.689,
      "mean_token_accuracy": 0.7967657148838043,
      "num_tokens": 2370338404.0,
      "step": 1140
    },
    {
      "epoch": 1.1138824674534615,
      "grad_norm": 0.18783468008041382,
      "learning_rate": 3.7687291488160195e-05,
      "loss": 0.6859,
      "mean_token_accuracy": 0.7958439752459526,
      "num_tokens": 2380757193.0,
      "step": 1145
    },
    {
      "epoch": 1.1187492395668572,
      "grad_norm": 0.17954984307289124,
      "learning_rate": 3.75703452382818e-05,
      "loss": 0.6986,
      "mean_token_accuracy": 0.7926660105586052,
      "num_tokens": 2391170839.0,
      "step": 1150
    },
    {
      "epoch": 1.1236160116802532,
      "grad_norm": 0.17523035407066345,
      "learning_rate": 3.745302968004749e-05,
      "loss": 0.6848,
      "mean_token_accuracy": 0.7971330270171165,
      "num_tokens": 2401559393.0,
      "step": 1155
    },
    {
      "epoch": 1.1284827837936489,
      "grad_norm": 0.1739313304424286,
      "learning_rate": 3.733534826011008e-05,
      "loss": 0.6957,
      "mean_token_accuracy": 0.7937130808830262,
      "num_tokens": 2411945855.0,
      "step": 1160
    },
    {
      "epoch": 1.1333495559070446,
      "grad_norm": 0.18176232278347015,
      "learning_rate": 3.721730443587118e-05,
      "loss": 0.6866,
      "mean_token_accuracy": 0.796080568432808,
      "num_tokens": 2422348168.0,
      "step": 1165
    },
    {
      "epoch": 1.1382163280204405,
      "grad_norm": 0.1996902972459793,
      "learning_rate": 3.709890167537958e-05,
      "loss": 0.6962,
      "mean_token_accuracy": 0.7931894510984421,
      "num_tokens": 2432728467.0,
      "step": 1170
    },
    {
      "epoch": 1.1430831001338362,
      "grad_norm": 0.19395747780799866,
      "learning_rate": 3.69801434572294e-05,
      "loss": 0.6984,
      "mean_token_accuracy": 0.7936637654900551,
      "num_tokens": 2443120518.0,
      "step": 1175
    },
    {
      "epoch": 1.147949872247232,
      "grad_norm": 0.1835799664258957,
      "learning_rate": 3.6861033270457825e-05,
      "loss": 0.6895,
      "mean_token_accuracy": 0.7953867301344871,
      "num_tokens": 2453494143.0,
      "step": 1180
    },
    {
      "epoch": 1.1528166443606278,
      "grad_norm": 0.17332947254180908,
      "learning_rate": 3.6741574614442683e-05,
      "loss": 0.6985,
      "mean_token_accuracy": 0.7929242372512817,
      "num_tokens": 2463873905.0,
      "step": 1185
    },
    {
      "epoch": 1.1576834164740235,
      "grad_norm": 0.219761461019516,
      "learning_rate": 3.6621770998799574e-05,
      "loss": 0.6977,
      "mean_token_accuracy": 0.7929645851254463,
      "num_tokens": 2474276371.0,
      "step": 1190
    },
    {
      "epoch": 1.1625501885874194,
      "grad_norm": 0.22956112027168274,
      "learning_rate": 3.6501625943278805e-05,
      "loss": 0.6947,
      "mean_token_accuracy": 0.7940642952919006,
      "num_tokens": 2484684944.0,
      "step": 1195
    },
    {
      "epoch": 1.1674169607008151,
      "grad_norm": 0.23507224023342133,
      "learning_rate": 3.638114297766193e-05,
      "loss": 0.6883,
      "mean_token_accuracy": 0.7957271650433541,
      "num_tokens": 2495088310.0,
      "step": 1200
    },
    {
      "epoch": 1.172283732814211,
      "grad_norm": 0.5020534992218018,
      "learning_rate": 3.62603256416581e-05,
      "loss": 0.6941,
      "mean_token_accuracy": 0.7942167058587074,
      "num_tokens": 2505483459.0,
      "step": 1205
    },
    {
      "epoch": 1.1771505049276068,
      "grad_norm": 0.17675137519836426,
      "learning_rate": 3.613917748480004e-05,
      "loss": 0.6981,
      "mean_token_accuracy": 0.7927851811051368,
      "num_tokens": 2515861285.0,
      "step": 1210
    },
    {
      "epoch": 1.1820172770410027,
      "grad_norm": 0.18070442974567413,
      "learning_rate": 3.601770206633974e-05,
      "loss": 0.6848,
      "mean_token_accuracy": 0.7967978402972221,
      "num_tokens": 2526269738.0,
      "step": 1215
    },
    {
      "epoch": 1.1868840491543984,
      "grad_norm": 0.19230933487415314,
      "learning_rate": 3.5895902955143965e-05,
      "loss": 0.6992,
      "mean_token_accuracy": 0.7925475016236305,
      "num_tokens": 2536676940.0,
      "step": 1220
    },
    {
      "epoch": 1.191750821267794,
      "grad_norm": 0.2971474528312683,
      "learning_rate": 3.5773783729589335e-05,
      "loss": 0.6858,
      "mean_token_accuracy": 0.796607282757759,
      "num_tokens": 2547083178.0,
      "step": 1225
    },
    {
      "epoch": 1.19661759338119,
      "grad_norm": 0.23534908890724182,
      "learning_rate": 3.5651347977457214e-05,
      "loss": 0.6945,
      "mean_token_accuracy": 0.7936753481626511,
      "num_tokens": 2557490182.0,
      "step": 1230
    },
    {
      "epoch": 1.2014843654945857,
      "grad_norm": 0.19064751267433167,
      "learning_rate": 3.5528599295828305e-05,
      "loss": 0.702,
      "mean_token_accuracy": 0.7917889147996903,
      "num_tokens": 2567907664.0,
      "step": 1235
    },
    {
      "epoch": 1.2063511376079816,
      "grad_norm": 0.19908669590950012,
      "learning_rate": 3.540554129097697e-05,
      "loss": 0.6909,
      "mean_token_accuracy": 0.7944000884890556,
      "num_tokens": 2578290092.0,
      "step": 1240
    },
    {
      "epoch": 1.2112179097213773,
      "grad_norm": 0.1871795952320099,
      "learning_rate": 3.5282177578265296e-05,
      "loss": 0.6917,
      "mean_token_accuracy": 0.7947396442294121,
      "num_tokens": 2588706368.0,
      "step": 1245
    },
    {
      "epoch": 1.216084681834773,
      "grad_norm": 0.20220763981342316,
      "learning_rate": 3.515851178203684e-05,
      "loss": 0.6975,
      "mean_token_accuracy": 0.7933344259858132,
      "num_tokens": 2599090772.0,
      "step": 1250
    },
    {
      "epoch": 1.220951453948169,
      "grad_norm": 0.19249574840068817,
      "learning_rate": 3.503454753551019e-05,
      "loss": 0.6873,
      "mean_token_accuracy": 0.7958372920751572,
      "num_tokens": 2609508990.0,
      "step": 1255
    },
    {
      "epoch": 1.2258182260615647,
      "grad_norm": 0.18894228339195251,
      "learning_rate": 3.4910288480672194e-05,
      "loss": 0.6949,
      "mean_token_accuracy": 0.7937337100505829,
      "num_tokens": 2619903360.0,
      "step": 1260
    },
    {
      "epoch": 1.2306849981749606,
      "grad_norm": 0.17634110152721405,
      "learning_rate": 3.478573826817099e-05,
      "loss": 0.7062,
      "mean_token_accuracy": 0.7904315680265427,
      "num_tokens": 2630256463.0,
      "step": 1265
    },
    {
      "epoch": 1.2355517702883563,
      "grad_norm": 0.18781399726867676,
      "learning_rate": 3.466090055720869e-05,
      "loss": 0.6888,
      "mean_token_accuracy": 0.7954594030976295,
      "num_tokens": 2640663552.0,
      "step": 1270
    },
    {
      "epoch": 1.240418542401752,
      "grad_norm": 0.17590133845806122,
      "learning_rate": 3.4535779015433975e-05,
      "loss": 0.682,
      "mean_token_accuracy": 0.7974461942911149,
      "num_tokens": 2651081751.0,
      "step": 1275
    },
    {
      "epoch": 1.245285314515148,
      "grad_norm": 0.1756858080625534,
      "learning_rate": 3.441037731883424e-05,
      "loss": 0.6915,
      "mean_token_accuracy": 0.7943955942988395,
      "num_tokens": 2661486959.0,
      "step": 1280
    },
    {
      "epoch": 1.2501520866285436,
      "grad_norm": 0.28670555353164673,
      "learning_rate": 3.428469915162767e-05,
      "loss": 0.6937,
      "mean_token_accuracy": 0.7940300226211547,
      "num_tokens": 2671917735.0,
      "step": 1285
    },
    {
      "epoch": 1.2550188587419395,
      "grad_norm": 0.18882758915424347,
      "learning_rate": 3.415874820615496e-05,
      "loss": 0.6901,
      "mean_token_accuracy": 0.7952745288610459,
      "num_tokens": 2682315320.0,
      "step": 1290
    },
    {
      "epoch": 1.2598856308553352,
      "grad_norm": 0.17917940020561218,
      "learning_rate": 3.403252818277086e-05,
      "loss": 0.6981,
      "mean_token_accuracy": 0.7927278980612755,
      "num_tokens": 2692734084.0,
      "step": 1295
    },
    {
      "epoch": 1.264752402968731,
      "grad_norm": 0.2403729259967804,
      "learning_rate": 3.390604278973543e-05,
      "loss": 0.6839,
      "mean_token_accuracy": 0.7966855794191361,
      "num_tokens": 2703133101.0,
      "step": 1300
    },
    {
      "epoch": 1.2696191750821268,
      "grad_norm": 0.1808985322713852,
      "learning_rate": 3.377929574310512e-05,
      "loss": 0.6912,
      "mean_token_accuracy": 0.7946472465991974,
      "num_tokens": 2713517341.0,
      "step": 1305
    },
    {
      "epoch": 1.2744859471955226,
      "grad_norm": 0.19213491678237915,
      "learning_rate": 3.36522907666236e-05,
      "loss": 0.6898,
      "mean_token_accuracy": 0.7950350716710091,
      "num_tokens": 2723909863.0,
      "step": 1310
    },
    {
      "epoch": 1.2793527193089185,
      "grad_norm": 0.18258869647979736,
      "learning_rate": 3.352503159161234e-05,
      "loss": 0.6857,
      "mean_token_accuracy": 0.7962309658527374,
      "num_tokens": 2734325865.0,
      "step": 1315
    },
    {
      "epoch": 1.2842194914223142,
      "grad_norm": 0.1933620274066925,
      "learning_rate": 3.3397521956860985e-05,
      "loss": 0.6916,
      "mean_token_accuracy": 0.7948616608977318,
      "num_tokens": 2744738282.0,
      "step": 1320
    },
    {
      "epoch": 1.2890862635357099,
      "grad_norm": 0.21153147518634796,
      "learning_rate": 3.326976560851753e-05,
      "loss": 0.6948,
      "mean_token_accuracy": 0.7936515882611275,
      "num_tokens": 2755166223.0,
      "step": 1325
    },
    {
      "epoch": 1.2939530356491058,
      "grad_norm": 0.1847887486219406,
      "learning_rate": 3.314176629997825e-05,
      "loss": 0.6937,
      "mean_token_accuracy": 0.7938945949077606,
      "num_tokens": 2765565367.0,
      "step": 1330
    },
    {
      "epoch": 1.2988198077625015,
      "grad_norm": 0.18322524428367615,
      "learning_rate": 3.301352779177743e-05,
      "loss": 0.6923,
      "mean_token_accuracy": 0.7943141534924507,
      "num_tokens": 2775995972.0,
      "step": 1335
    },
    {
      "epoch": 1.3036865798758974,
      "grad_norm": 0.1810372918844223,
      "learning_rate": 3.288505385147687e-05,
      "loss": 0.6927,
      "mean_token_accuracy": 0.7944171726703644,
      "num_tokens": 2786351201.0,
      "step": 1340
    },
    {
      "epoch": 1.3085533519892931,
      "grad_norm": 2.6634132862091064,
      "learning_rate": 3.2756348253555215e-05,
      "loss": 0.7017,
      "mean_token_accuracy": 0.7941828548908234,
      "num_tokens": 2796761588.0,
      "step": 1345
    },
    {
      "epoch": 1.3134201241026888,
      "grad_norm": 0.1917794942855835,
      "learning_rate": 3.262741477929706e-05,
      "loss": 0.6901,
      "mean_token_accuracy": 0.7950336009263992,
      "num_tokens": 2807192367.0,
      "step": 1350
    },
    {
      "epoch": 1.3182868962160847,
      "grad_norm": 0.19772347807884216,
      "learning_rate": 3.249825721668185e-05,
      "loss": 0.6923,
      "mean_token_accuracy": 0.7945320799946785,
      "num_tokens": 2817589358.0,
      "step": 1355
    },
    {
      "epoch": 1.3231536683294804,
      "grad_norm": 0.1927238255739212,
      "learning_rate": 3.2368879360272606e-05,
      "loss": 0.6965,
      "mean_token_accuracy": 0.7931426867842675,
      "num_tokens": 2828004221.0,
      "step": 1360
    },
    {
      "epoch": 1.3280204404428764,
      "grad_norm": 0.19713135063648224,
      "learning_rate": 3.223928501110441e-05,
      "loss": 0.6992,
      "mean_token_accuracy": 0.7935102447867394,
      "num_tokens": 2838430893.0,
      "step": 1365
    },
    {
      "epoch": 1.332887212556272,
      "grad_norm": 0.26476624608039856,
      "learning_rate": 3.21094779765728e-05,
      "loss": 0.7014,
      "mean_token_accuracy": 0.7921116515994072,
      "num_tokens": 2848824501.0,
      "step": 1370
    },
    {
      "epoch": 1.3377539846696678,
      "grad_norm": 0.27639538049697876,
      "learning_rate": 3.197946207032181e-05,
      "loss": 0.7029,
      "mean_token_accuracy": 0.7926134586334228,
      "num_tokens": 2859223378.0,
      "step": 1375
    },
    {
      "epoch": 1.3426207567830637,
      "grad_norm": 0.24344052374362946,
      "learning_rate": 3.1849241112132036e-05,
      "loss": 0.6852,
      "mean_token_accuracy": 0.7962699472904206,
      "num_tokens": 2869600997.0,
      "step": 1380
    },
    {
      "epoch": 1.3474875288964594,
      "grad_norm": 0.1964772641658783,
      "learning_rate": 3.1718818927808335e-05,
      "loss": 0.6893,
      "mean_token_accuracy": 0.7950659811496734,
      "num_tokens": 2879990058.0,
      "step": 1385
    },
    {
      "epoch": 1.3523543010098553,
      "grad_norm": 0.20229022204875946,
      "learning_rate": 3.158819934906746e-05,
      "loss": 0.6968,
      "mean_token_accuracy": 0.7932897135615349,
      "num_tokens": 2890404232.0,
      "step": 1390
    },
    {
      "epoch": 1.357221073123251,
      "grad_norm": 0.18493428826332092,
      "learning_rate": 3.145738621342549e-05,
      "loss": 0.6909,
      "mean_token_accuracy": 0.7950246050953865,
      "num_tokens": 2900798146.0,
      "step": 1395
    },
    {
      "epoch": 1.3620878452366467,
      "grad_norm": 0.17084717750549316,
      "learning_rate": 3.132638336408507e-05,
      "loss": 0.6905,
      "mean_token_accuracy": 0.7950514882802964,
      "num_tokens": 2911205741.0,
      "step": 1400
    },
    {
      "epoch": 1.3669546173500426,
      "grad_norm": 0.1768016368150711,
      "learning_rate": 3.11951946498225e-05,
      "loss": 0.6841,
      "mean_token_accuracy": 0.7967058211565018,
      "num_tokens": 2921611758.0,
      "step": 1405
    },
    {
      "epoch": 1.3718213894634383,
      "grad_norm": 0.20268017053604126,
      "learning_rate": 3.1063823924874674e-05,
      "loss": 0.692,
      "mean_token_accuracy": 0.7951907083392143,
      "num_tokens": 2932025078.0,
      "step": 1410
    },
    {
      "epoch": 1.3766881615768343,
      "grad_norm": 0.1876542866230011,
      "learning_rate": 3.093227504882584e-05,
      "loss": 0.6946,
      "mean_token_accuracy": 0.7936708420515061,
      "num_tokens": 2942435899.0,
      "step": 1415
    },
    {
      "epoch": 1.38155493369023,
      "grad_norm": 0.18529033660888672,
      "learning_rate": 3.08005518864942e-05,
      "loss": 0.6955,
      "mean_token_accuracy": 0.7934351354837418,
      "num_tokens": 2952840385.0,
      "step": 1420
    },
    {
      "epoch": 1.3864217058036257,
      "grad_norm": 0.22768966853618622,
      "learning_rate": 3.0668658307818385e-05,
      "loss": 0.7604,
      "mean_token_accuracy": 0.7927023887634277,
      "num_tokens": 2963243610.0,
      "step": 1425
    },
    {
      "epoch": 1.3912884779170216,
      "grad_norm": 0.20312738418579102,
      "learning_rate": 3.053659818774372e-05,
      "loss": 0.6867,
      "mean_token_accuracy": 0.7956630036234855,
      "num_tokens": 2973656324.0,
      "step": 1430
    },
    {
      "epoch": 1.3961552500304173,
      "grad_norm": 0.20471729338169098,
      "learning_rate": 3.0404375406108427e-05,
      "loss": 0.6886,
      "mean_token_accuracy": 0.7955989301204681,
      "num_tokens": 2984078875.0,
      "step": 1435
    },
    {
      "epoch": 1.4010220221438132,
      "grad_norm": 0.18820820748806,
      "learning_rate": 3.027199384752962e-05,
      "loss": 0.689,
      "mean_token_accuracy": 0.7951110273599624,
      "num_tokens": 2994501824.0,
      "step": 1440
    },
    {
      "epoch": 1.405888794257209,
      "grad_norm": 0.17648597061634064,
      "learning_rate": 3.013945740128914e-05,
      "loss": 0.6825,
      "mean_token_accuracy": 0.7968734323978424,
      "num_tokens": 3004910306.0,
      "step": 1445
    },
    {
      "epoch": 1.4107555663706046,
      "grad_norm": 0.16382163763046265,
      "learning_rate": 3.0006769961219362e-05,
      "loss": 0.6899,
      "mean_token_accuracy": 0.7948377743363381,
      "num_tokens": 3015328215.0,
      "step": 1450
    },
    {
      "epoch": 1.4156223384840005,
      "grad_norm": 0.21011081337928772,
      "learning_rate": 2.9873935425588734e-05,
      "loss": 0.6861,
      "mean_token_accuracy": 0.7960008919239044,
      "num_tokens": 3025706448.0,
      "step": 1455
    },
    {
      "epoch": 1.4204891105973962,
      "grad_norm": 0.17258188128471375,
      "learning_rate": 2.97409576969873e-05,
      "loss": 0.6921,
      "mean_token_accuracy": 0.7939485758543015,
      "num_tokens": 3036110400.0,
      "step": 1460
    },
    {
      "epoch": 1.4253558827107922,
      "grad_norm": 0.1675489842891693,
      "learning_rate": 2.9607840682211985e-05,
      "loss": 0.6911,
      "mean_token_accuracy": 0.7944290101528168,
      "num_tokens": 3046524356.0,
      "step": 1465
    },
    {
      "epoch": 1.4302226548241879,
      "grad_norm": 0.214289590716362,
      "learning_rate": 2.9474588292151866e-05,
      "loss": 0.6902,
      "mean_token_accuracy": 0.794608598947525,
      "num_tokens": 3056955838.0,
      "step": 1470
    },
    {
      "epoch": 1.4350894269375836,
      "grad_norm": 0.21568721532821655,
      "learning_rate": 2.9341204441673266e-05,
      "loss": 0.6916,
      "mean_token_accuracy": 0.7947678297758103,
      "num_tokens": 3067339957.0,
      "step": 1475
    },
    {
      "epoch": 1.4399561990509795,
      "grad_norm": 2.9168436527252197,
      "learning_rate": 2.9207693049504697e-05,
      "loss": 0.7007,
      "mean_token_accuracy": 0.7914907485246658,
      "num_tokens": 3077738363.0,
      "step": 1480
    },
    {
      "epoch": 1.4448229711643752,
      "grad_norm": 0.1935916543006897,
      "learning_rate": 2.9074058038121794e-05,
      "loss": 0.6878,
      "mean_token_accuracy": 0.7950029358267784,
      "num_tokens": 3088152247.0,
      "step": 1485
    },
    {
      "epoch": 1.4496897432777711,
      "grad_norm": 0.20898501574993134,
      "learning_rate": 2.8940303333632024e-05,
      "loss": 0.6965,
      "mean_token_accuracy": 0.7932757169008255,
      "num_tokens": 3098546462.0,
      "step": 1490
    },
    {
      "epoch": 1.4545565153911668,
      "grad_norm": 0.1852477788925171,
      "learning_rate": 2.880643286565936e-05,
      "loss": 0.681,
      "mean_token_accuracy": 0.7974221348762512,
      "num_tokens": 3108949052.0,
      "step": 1495
    },
    {
      "epoch": 1.4594232875045625,
      "grad_norm": 1.4246114492416382,
      "learning_rate": 2.867245056722883e-05,
      "loss": 0.6866,
      "mean_token_accuracy": 0.7949179708957672,
      "num_tokens": 3119344027.0,
      "step": 1500
    },
    {
      "epoch": 1.4642900596179584,
      "grad_norm": 0.20540167391300201,
      "learning_rate": 2.853836037465097e-05,
      "loss": 0.6896,
      "mean_token_accuracy": 0.7950207859277725,
      "num_tokens": 3129754018.0,
      "step": 1505
    },
    {
      "epoch": 1.4691568317313541,
      "grad_norm": 0.1665203869342804,
      "learning_rate": 2.840416622740617e-05,
      "loss": 0.6921,
      "mean_token_accuracy": 0.79476508051157,
      "num_tokens": 3140163388.0,
      "step": 1510
    },
    {
      "epoch": 1.47402360384475,
      "grad_norm": 0.18602830171585083,
      "learning_rate": 2.8269872068028925e-05,
      "loss": 0.6863,
      "mean_token_accuracy": 0.7957534775137901,
      "num_tokens": 3150552997.0,
      "step": 1515
    },
    {
      "epoch": 1.4788903759581458,
      "grad_norm": 1.2237293720245361,
      "learning_rate": 2.813548184199205e-05,
      "loss": 0.6971,
      "mean_token_accuracy": 0.7951232016086578,
      "num_tokens": 3160958499.0,
      "step": 1520
    },
    {
      "epoch": 1.4837571480715415,
      "grad_norm": 0.5988807678222656,
      "learning_rate": 2.800099949759072e-05,
      "loss": 0.705,
      "mean_token_accuracy": 0.7925426855683326,
      "num_tokens": 3171333914.0,
      "step": 1525
    },
    {
      "epoch": 1.4886239201849374,
      "grad_norm": 0.2804173231124878,
      "learning_rate": 2.7866428985826476e-05,
      "loss": 0.7028,
      "mean_token_accuracy": 0.7928298532962799,
      "num_tokens": 3181753216.0,
      "step": 1530
    },
    {
      "epoch": 1.493490692298333,
      "grad_norm": 2.7266533374786377,
      "learning_rate": 2.773177426029116e-05,
      "loss": 0.6899,
      "mean_token_accuracy": 0.7946895241737366,
      "num_tokens": 3192146745.0,
      "step": 1535
    },
    {
      "epoch": 1.498357464411729,
      "grad_norm": 0.2572759985923767,
      "learning_rate": 2.759703927705077e-05,
      "loss": 0.7164,
      "mean_token_accuracy": 0.7924556523561478,
      "num_tokens": 3202515846.0,
      "step": 1540
    },
    {
      "epoch": 1.5032242365251247,
      "grad_norm": 0.22885890305042267,
      "learning_rate": 2.7462227994529217e-05,
      "loss": 0.7264,
      "mean_token_accuracy": 0.7924733236432076,
      "num_tokens": 3212935901.0,
      "step": 1545
    },
    {
      "epoch": 1.5080910086385204,
      "grad_norm": 1.5094523429870605,
      "learning_rate": 2.7327344373392037e-05,
      "loss": 0.7267,
      "mean_token_accuracy": 0.7923997566103935,
      "num_tokens": 3223337776.0,
      "step": 1550
    },
    {
      "epoch": 1.5129577807519163,
      "grad_norm": 0.22536472976207733,
      "learning_rate": 2.7192392376430016e-05,
      "loss": 0.6884,
      "mean_token_accuracy": 0.7956211313605308,
      "num_tokens": 3233752060.0,
      "step": 1555
    },
    {
      "epoch": 1.517824552865312,
      "grad_norm": 1.918136477470398,
      "learning_rate": 2.7057375968442782e-05,
      "loss": 0.7155,
      "mean_token_accuracy": 0.7923317447304725,
      "num_tokens": 3244142903.0,
      "step": 1560
    },
    {
      "epoch": 1.522691324978708,
      "grad_norm": 0.19763804972171783,
      "learning_rate": 2.6922299116122317e-05,
      "loss": 0.6847,
      "mean_token_accuracy": 0.7965306863188744,
      "num_tokens": 3254565699.0,
      "step": 1565
    },
    {
      "epoch": 1.5275580970921037,
      "grad_norm": 0.22156846523284912,
      "learning_rate": 2.6787165787936418e-05,
      "loss": 0.6934,
      "mean_token_accuracy": 0.7947420656681061,
      "num_tokens": 3265001124.0,
      "step": 1570
    },
    {
      "epoch": 1.5324248692054994,
      "grad_norm": 0.19398213922977448,
      "learning_rate": 2.66519799540121e-05,
      "loss": 0.6802,
      "mean_token_accuracy": 0.7974090471863746,
      "num_tokens": 3275384701.0,
      "step": 1575
    },
    {
      "epoch": 1.5372916413188953,
      "grad_norm": 0.19405736029148102,
      "learning_rate": 2.6516745586018965e-05,
      "loss": 0.6839,
      "mean_token_accuracy": 0.7966880887746811,
      "num_tokens": 3285765755.0,
      "step": 1580
    },
    {
      "epoch": 1.542158413432291,
      "grad_norm": 0.184486985206604,
      "learning_rate": 2.6381466657052522e-05,
      "loss": 0.696,
      "mean_token_accuracy": 0.7933828502893447,
      "num_tokens": 3296193995.0,
      "step": 1585
    },
    {
      "epoch": 1.547025185545687,
      "grad_norm": 0.19228088855743408,
      "learning_rate": 2.624614714151743e-05,
      "loss": 0.6849,
      "mean_token_accuracy": 0.7961423754692077,
      "num_tokens": 3306589651.0,
      "step": 1590
    },
    {
      "epoch": 1.5518919576590826,
      "grad_norm": 0.21229314804077148,
      "learning_rate": 2.611079101501078e-05,
      "loss": 0.6786,
      "mean_token_accuracy": 0.7981480285525322,
      "num_tokens": 3316990151.0,
      "step": 1595
    },
    {
      "epoch": 1.5567587297724783,
      "grad_norm": 0.17554455995559692,
      "learning_rate": 2.597540225420525e-05,
      "loss": 0.6897,
      "mean_token_accuracy": 0.7948550283908844,
      "num_tokens": 3327405990.0,
      "step": 1600
    },
    {
      "epoch": 1.5616255018858742,
      "grad_norm": 0.17284856736660004,
      "learning_rate": 2.5839984836732294e-05,
      "loss": 0.6904,
      "mean_token_accuracy": 0.7947103083133698,
      "num_tokens": 3337835598.0,
      "step": 1605
    },
    {
      "epoch": 1.56649227399927,
      "grad_norm": 0.1814858317375183,
      "learning_rate": 2.570454274106529e-05,
      "loss": 0.6844,
      "mean_token_accuracy": 0.7962429285049438,
      "num_tokens": 3348245642.0,
      "step": 1610
    },
    {
      "epoch": 1.5713590461126659,
      "grad_norm": 0.1687670648097992,
      "learning_rate": 2.556907994640264e-05,
      "loss": 0.6974,
      "mean_token_accuracy": 0.793082794547081,
      "num_tokens": 3358659240.0,
      "step": 1615
    },
    {
      "epoch": 1.5762258182260616,
      "grad_norm": 0.16963249444961548,
      "learning_rate": 2.5433600432550876e-05,
      "loss": 0.6772,
      "mean_token_accuracy": 0.7980673342943192,
      "num_tokens": 3369055570.0,
      "step": 1620
    },
    {
      "epoch": 1.5810925903394573,
      "grad_norm": 0.18524517118930817,
      "learning_rate": 2.5298108179807696e-05,
      "loss": 0.6883,
      "mean_token_accuracy": 0.7952644348144531,
      "num_tokens": 3379469764.0,
      "step": 1625
    },
    {
      "epoch": 1.5859593624528532,
      "grad_norm": 1.9075506925582886,
      "learning_rate": 2.5162607168845103e-05,
      "loss": 0.6905,
      "mean_token_accuracy": 0.7944433286786079,
      "num_tokens": 3389885768.0,
      "step": 1630
    },
    {
      "epoch": 1.5908261345662489,
      "grad_norm": 0.18973305821418762,
      "learning_rate": 2.5027101380592387e-05,
      "loss": 0.683,
      "mean_token_accuracy": 0.7967569932341576,
      "num_tokens": 3400263702.0,
      "step": 1635
    },
    {
      "epoch": 1.5956929066796448,
      "grad_norm": 0.21172359585762024,
      "learning_rate": 2.489159479611919e-05,
      "loss": 0.706,
      "mean_token_accuracy": 0.7930256441235543,
      "num_tokens": 3410662791.0,
      "step": 1640
    },
    {
      "epoch": 1.6005596787930405,
      "grad_norm": 0.17895331978797913,
      "learning_rate": 2.475609139651855e-05,
      "loss": 0.6874,
      "mean_token_accuracy": 0.7958353191614151,
      "num_tokens": 3421078158.0,
      "step": 1645
    },
    {
      "epoch": 1.6054264509064362,
      "grad_norm": 0.16819441318511963,
      "learning_rate": 2.4620595162789936e-05,
      "loss": 0.6897,
      "mean_token_accuracy": 0.7949976801872254,
      "num_tokens": 3431461927.0,
      "step": 1650
    },
    {
      "epoch": 1.6102932230198321,
      "grad_norm": 0.2880989909172058,
      "learning_rate": 2.4485110075722304e-05,
      "loss": 0.6915,
      "mean_token_accuracy": 0.7945791348814965,
      "num_tokens": 3441866917.0,
      "step": 1655
    },
    {
      "epoch": 1.6151599951332278,
      "grad_norm": 0.1703127920627594,
      "learning_rate": 2.434964011577709e-05,
      "loss": 0.6911,
      "mean_token_accuracy": 0.7944571956992149,
      "num_tokens": 3452222465.0,
      "step": 1660
    },
    {
      "epoch": 1.6200267672466238,
      "grad_norm": 0.1824474185705185,
      "learning_rate": 2.4214189262971352e-05,
      "loss": 0.6817,
      "mean_token_accuracy": 0.7970559284090996,
      "num_tokens": 3462584512.0,
      "step": 1665
    },
    {
      "epoch": 1.6248935393600195,
      "grad_norm": 0.16564872860908508,
      "learning_rate": 2.4078761496760755e-05,
      "loss": 0.6948,
      "mean_token_accuracy": 0.7930441677570343,
      "num_tokens": 3472988022.0,
      "step": 1670
    },
    {
      "epoch": 1.6297603114734152,
      "grad_norm": 0.16088281571865082,
      "learning_rate": 2.394336079592271e-05,
      "loss": 0.6913,
      "mean_token_accuracy": 0.7940496131777763,
      "num_tokens": 3483381230.0,
      "step": 1675
    },
    {
      "epoch": 1.634627083586811,
      "grad_norm": 0.7534094452857971,
      "learning_rate": 2.3807991138439477e-05,
      "loss": 0.6776,
      "mean_token_accuracy": 0.7985736101865768,
      "num_tokens": 3493749632.0,
      "step": 1680
    },
    {
      "epoch": 1.639493855700207,
      "grad_norm": 0.17122338712215424,
      "learning_rate": 2.3672656501381272e-05,
      "loss": 0.6941,
      "mean_token_accuracy": 0.7936079114675522,
      "num_tokens": 3504156246.0,
      "step": 1685
    },
    {
      "epoch": 1.6443606278136027,
      "grad_norm": 0.18514901399612427,
      "learning_rate": 2.353736086078941e-05,
      "loss": 0.6763,
      "mean_token_accuracy": 0.7983838528394699,
      "num_tokens": 3514563377.0,
      "step": 1690
    },
    {
      "epoch": 1.6492273999269984,
      "grad_norm": 0.19726413488388062,
      "learning_rate": 2.3402108191559563e-05,
      "loss": 0.6882,
      "mean_token_accuracy": 0.7953017488121986,
      "num_tokens": 3524955818.0,
      "step": 1695
    },
    {
      "epoch": 1.654094172040394,
      "grad_norm": 0.28310248255729675,
      "learning_rate": 2.3266902467324907e-05,
      "loss": 0.6841,
      "mean_token_accuracy": 0.7963617026805878,
      "num_tokens": 3535349955.0,
      "step": 1700
    },
    {
      "epoch": 1.65896094415379,
      "grad_norm": 0.1795838475227356,
      "learning_rate": 2.3131747660339394e-05,
      "loss": 0.6854,
      "mean_token_accuracy": 0.7960692644119263,
      "num_tokens": 3545778129.0,
      "step": 1705
    },
    {
      "epoch": 1.663827716267186,
      "grad_norm": 0.17762821912765503,
      "learning_rate": 2.2996647741361096e-05,
      "loss": 0.6844,
      "mean_token_accuracy": 0.7960609808564186,
      "num_tokens": 3556167810.0,
      "step": 1710
    },
    {
      "epoch": 1.6686944883805817,
      "grad_norm": 0.20176132023334503,
      "learning_rate": 2.2861606679535473e-05,
      "loss": 0.6956,
      "mean_token_accuracy": 0.7932230189442635,
      "num_tokens": 3566569451.0,
      "step": 1715
    },
    {
      "epoch": 1.6735612604939774,
      "grad_norm": 0.17322279512882233,
      "learning_rate": 2.2726628442278826e-05,
      "loss": 0.6816,
      "mean_token_accuracy": 0.7969667941331864,
      "num_tokens": 3576978386.0,
      "step": 1720
    },
    {
      "epoch": 1.678428032607373,
      "grad_norm": 0.17014460265636444,
      "learning_rate": 2.2591716995161716e-05,
      "loss": 0.6834,
      "mean_token_accuracy": 0.7964163482189178,
      "num_tokens": 3587373716.0,
      "step": 1725
    },
    {
      "epoch": 1.683294804720769,
      "grad_norm": 0.1617697775363922,
      "learning_rate": 2.245687630179242e-05,
      "loss": 0.6817,
      "mean_token_accuracy": 0.7968165248632431,
      "num_tokens": 3597773301.0,
      "step": 1730
    },
    {
      "epoch": 1.688161576834165,
      "grad_norm": 0.23249968886375427,
      "learning_rate": 2.2322110323700568e-05,
      "loss": 0.7014,
      "mean_token_accuracy": 0.7929875969886779,
      "num_tokens": 3608169136.0,
      "step": 1735
    },
    {
      "epoch": 1.6930283489475606,
      "grad_norm": 0.2602331340312958,
      "learning_rate": 2.218742302022066e-05,
      "loss": 0.6988,
      "mean_token_accuracy": 0.7926072865724564,
      "num_tokens": 3618545886.0,
      "step": 1740
    },
    {
      "epoch": 1.6978951210609563,
      "grad_norm": 0.20079894363880157,
      "learning_rate": 2.20528183483758e-05,
      "loss": 0.6854,
      "mean_token_accuracy": 0.7956514552235603,
      "num_tokens": 3628972515.0,
      "step": 1745
    },
    {
      "epoch": 1.702761893174352,
      "grad_norm": 0.32910633087158203,
      "learning_rate": 2.191830026276146e-05,
      "loss": 0.7065,
      "mean_token_accuracy": 0.7947751507163048,
      "num_tokens": 3639350449.0,
      "step": 1750
    },
    {
      "epoch": 1.707628665287748,
      "grad_norm": 0.2448817938566208,
      "learning_rate": 2.1783872715429228e-05,
      "loss": 0.6955,
      "mean_token_accuracy": 0.7938882395625114,
      "num_tokens": 3649746768.0,
      "step": 1755
    },
    {
      "epoch": 1.7124954374011438,
      "grad_norm": 0.20740175247192383,
      "learning_rate": 2.164953965577074e-05,
      "loss": 0.6925,
      "mean_token_accuracy": 0.794175212085247,
      "num_tokens": 3660145980.0,
      "step": 1760
    },
    {
      "epoch": 1.7173622095145396,
      "grad_norm": 0.16701149940490723,
      "learning_rate": 2.151530503040169e-05,
      "loss": 0.6848,
      "mean_token_accuracy": 0.7963548466563225,
      "num_tokens": 3670560895.0,
      "step": 1765
    },
    {
      "epoch": 1.7222289816279353,
      "grad_norm": 0.1693955808877945,
      "learning_rate": 2.1381172783045783e-05,
      "loss": 0.6971,
      "mean_token_accuracy": 0.7928205639123916,
      "num_tokens": 3680980039.0,
      "step": 1770
    },
    {
      "epoch": 1.727095753741331,
      "grad_norm": 0.18243108689785004,
      "learning_rate": 2.1247146854418947e-05,
      "loss": 0.6845,
      "mean_token_accuracy": 0.7961863145232201,
      "num_tokens": 3691379051.0,
      "step": 1775
    },
    {
      "epoch": 1.7319625258547269,
      "grad_norm": 0.19192378222942352,
      "learning_rate": 2.111323118211356e-05,
      "loss": 0.7258,
      "mean_token_accuracy": 0.79607575237751,
      "num_tokens": 3701754371.0,
      "step": 1780
    },
    {
      "epoch": 1.7368292979681228,
      "grad_norm": 0.17413221299648285,
      "learning_rate": 2.09794297004827e-05,
      "loss": 0.6905,
      "mean_token_accuracy": 0.7955564290285111,
      "num_tokens": 3712152830.0,
      "step": 1785
    },
    {
      "epoch": 1.7416960700815185,
      "grad_norm": 0.28621402382850647,
      "learning_rate": 2.084574634052465e-05,
      "loss": 0.6847,
      "mean_token_accuracy": 0.7981747806072235,
      "num_tokens": 3722569771.0,
      "step": 1790
    },
    {
      "epoch": 1.7465628421949142,
      "grad_norm": 0.1716383546590805,
      "learning_rate": 2.0712185029767316e-05,
      "loss": 0.6811,
      "mean_token_accuracy": 0.7969594612717629,
      "num_tokens": 3732950317.0,
      "step": 1795
    },
    {
      "epoch": 1.75142961430831,
      "grad_norm": 0.16310901939868927,
      "learning_rate": 2.0578749692152882e-05,
      "loss": 0.6957,
      "mean_token_accuracy": 0.7953600063920021,
      "num_tokens": 3743358385.0,
      "step": 1800
    },
    {
      "epoch": 1.7562963864217058,
      "grad_norm": 0.1935407817363739,
      "learning_rate": 2.0445444247922552e-05,
      "loss": 0.6892,
      "mean_token_accuracy": 0.7986009180545807,
      "num_tokens": 3753759465.0,
      "step": 1805
    },
    {
      "epoch": 1.7611631585351017,
      "grad_norm": 4.031457424163818,
      "learning_rate": 2.031227261350133e-05,
      "loss": 0.7184,
      "mean_token_accuracy": 0.7935503408312797,
      "num_tokens": 3764159764.0,
      "step": 1810
    },
    {
      "epoch": 1.7660299306484974,
      "grad_norm": 0.19278070330619812,
      "learning_rate": 2.0179238701382968e-05,
      "loss": 0.682,
      "mean_token_accuracy": 0.7970230028033256,
      "num_tokens": 3774552847.0,
      "step": 1815
    },
    {
      "epoch": 1.7708967027618931,
      "grad_norm": 0.18672893941402435,
      "learning_rate": 2.0046346420015067e-05,
      "loss": 0.6827,
      "mean_token_accuracy": 0.796694865822792,
      "num_tokens": 3784939236.0,
      "step": 1820
    },
    {
      "epoch": 1.7757634748752888,
      "grad_norm": 0.1664484441280365,
      "learning_rate": 1.991359967368416e-05,
      "loss": 0.6943,
      "mean_token_accuracy": 0.7935452565550805,
      "num_tokens": 3795335189.0,
      "step": 1825
    },
    {
      "epoch": 1.7806302469886848,
      "grad_norm": 0.1863197386264801,
      "learning_rate": 1.978100236240109e-05,
      "loss": 0.6867,
      "mean_token_accuracy": 0.7953874439001083,
      "num_tokens": 3805749456.0,
      "step": 1830
    },
    {
      "epoch": 1.7854970191020807,
      "grad_norm": 0.15571993589401245,
      "learning_rate": 1.96485583817864e-05,
      "loss": 0.6797,
      "mean_token_accuracy": 0.797387282550335,
      "num_tokens": 3816159474.0,
      "step": 1835
    },
    {
      "epoch": 1.7903637912154764,
      "grad_norm": 0.1599532961845398,
      "learning_rate": 1.951627162295587e-05,
      "loss": 0.6849,
      "mean_token_accuracy": 0.7957388177514076,
      "num_tokens": 3826550691.0,
      "step": 1840
    },
    {
      "epoch": 1.795230563328872,
      "grad_norm": 0.16260510683059692,
      "learning_rate": 1.9384145972406193e-05,
      "loss": 0.6856,
      "mean_token_accuracy": 0.795316918194294,
      "num_tokens": 3836943885.0,
      "step": 1845
    },
    {
      "epoch": 1.8000973354422678,
      "grad_norm": 0.17742158472537994,
      "learning_rate": 1.9252185311900846e-05,
      "loss": 0.699,
      "mean_token_accuracy": 0.7967996031045914,
      "num_tokens": 3847371506.0,
      "step": 1850
    },
    {
      "epoch": 1.8049641075556637,
      "grad_norm": 0.16368843615055084,
      "learning_rate": 1.9120393518355966e-05,
      "loss": 0.6849,
      "mean_token_accuracy": 0.7958560883998871,
      "num_tokens": 3857723106.0,
      "step": 1855
    },
    {
      "epoch": 1.8098308796690596,
      "grad_norm": 0.15892305970191956,
      "learning_rate": 1.8988774463726543e-05,
      "loss": 0.6772,
      "mean_token_accuracy": 0.7981528207659722,
      "num_tokens": 3868140001.0,
      "step": 1860
    },
    {
      "epoch": 1.8146976517824553,
      "grad_norm": 0.15916796028614044,
      "learning_rate": 1.8857332014892558e-05,
      "loss": 0.6864,
      "mean_token_accuracy": 0.795421127974987,
      "num_tokens": 3878547750.0,
      "step": 1865
    },
    {
      "epoch": 1.819564423895851,
      "grad_norm": 0.1582012176513672,
      "learning_rate": 1.8726070033545467e-05,
      "loss": 0.6762,
      "mean_token_accuracy": 0.7984607443213463,
      "num_tokens": 3888957777.0,
      "step": 1870
    },
    {
      "epoch": 1.8244311960092467,
      "grad_norm": 0.17070528864860535,
      "learning_rate": 1.8594992376074703e-05,
      "loss": 0.6824,
      "mean_token_accuracy": 0.7968670964241028,
      "num_tokens": 3899350092.0,
      "step": 1875
    },
    {
      "epoch": 1.8292979681226427,
      "grad_norm": 0.15726220607757568,
      "learning_rate": 1.846410289345438e-05,
      "loss": 0.6726,
      "mean_token_accuracy": 0.7993271827697754,
      "num_tokens": 3909760127.0,
      "step": 1880
    },
    {
      "epoch": 1.8341647402360386,
      "grad_norm": 0.5928107500076294,
      "learning_rate": 1.833340543113016e-05,
      "loss": 0.6765,
      "mean_token_accuracy": 0.7985545217990875,
      "num_tokens": 3920170693.0,
      "step": 1885
    },
    {
      "epoch": 1.8390315123494343,
      "grad_norm": 0.1675027757883072,
      "learning_rate": 1.82029038289063e-05,
      "loss": 0.6717,
      "mean_token_accuracy": 0.7994899287819862,
      "num_tokens": 3930605330.0,
      "step": 1890
    },
    {
      "epoch": 1.84389828446283,
      "grad_norm": 0.16610071063041687,
      "learning_rate": 1.8072601920832786e-05,
      "loss": 0.6957,
      "mean_token_accuracy": 0.7925936654210091,
      "num_tokens": 3940975153.0,
      "step": 1895
    },
    {
      "epoch": 1.8487650565762257,
      "grad_norm": 0.3610976040363312,
      "learning_rate": 1.7942503535092734e-05,
      "loss": 0.7093,
      "mean_token_accuracy": 0.7920773103833199,
      "num_tokens": 3951392262.0,
      "step": 1900
    },
    {
      "epoch": 1.8536318286896216,
      "grad_norm": 0.16417337954044342,
      "learning_rate": 1.7812612493889928e-05,
      "loss": 0.6781,
      "mean_token_accuracy": 0.7974630832672119,
      "num_tokens": 3961768781.0,
      "step": 1905
    },
    {
      "epoch": 1.8584986008030175,
      "grad_norm": 0.16069912910461426,
      "learning_rate": 1.7682932613336473e-05,
      "loss": 0.6708,
      "mean_token_accuracy": 0.7995835781097412,
      "num_tokens": 3972166854.0,
      "step": 1910
    },
    {
      "epoch": 1.8633653729164132,
      "grad_norm": 0.16137218475341797,
      "learning_rate": 1.7553467703340754e-05,
      "loss": 0.6849,
      "mean_token_accuracy": 0.7960717245936394,
      "num_tokens": 3982539969.0,
      "step": 1915
    },
    {
      "epoch": 1.868232145029809,
      "grad_norm": 0.1669033020734787,
      "learning_rate": 1.7424221567495435e-05,
      "loss": 0.677,
      "mean_token_accuracy": 0.7978730812668801,
      "num_tokens": 3992952387.0,
      "step": 1920
    },
    {
      "epoch": 1.8730989171432046,
      "grad_norm": 0.17035235464572906,
      "learning_rate": 1.729519800296575e-05,
      "loss": 0.6812,
      "mean_token_accuracy": 0.796632269024849,
      "num_tokens": 4003344208.0,
      "step": 1925
    },
    {
      "epoch": 1.8779656892566006,
      "grad_norm": 0.15629805624485016,
      "learning_rate": 1.7166400800377948e-05,
      "loss": 0.6814,
      "mean_token_accuracy": 0.7966139331459999,
      "num_tokens": 4013717458.0,
      "step": 1930
    },
    {
      "epoch": 1.8828324613699965,
      "grad_norm": 0.1578161120414734,
      "learning_rate": 1.7037833743707892e-05,
      "loss": 0.6887,
      "mean_token_accuracy": 0.7943219542503357,
      "num_tokens": 4024133580.0,
      "step": 1935
    },
    {
      "epoch": 1.8876992334833922,
      "grad_norm": 0.16146868467330933,
      "learning_rate": 1.6909500610169926e-05,
      "loss": 0.6788,
      "mean_token_accuracy": 0.7976189061999321,
      "num_tokens": 4034519223.0,
      "step": 1940
    },
    {
      "epoch": 1.892566005596788,
      "grad_norm": 1.3928577899932861,
      "learning_rate": 1.6781405170105902e-05,
      "loss": 0.6878,
      "mean_token_accuracy": 0.7952196672558784,
      "num_tokens": 4044931549.0,
      "step": 1945
    },
    {
      "epoch": 1.8974327777101836,
      "grad_norm": 0.14806461334228516,
      "learning_rate": 1.6653551186874372e-05,
      "loss": 0.6843,
      "mean_token_accuracy": 0.7961780175566673,
      "num_tokens": 4055346243.0,
      "step": 1950
    },
    {
      "epoch": 1.9022995498235795,
      "grad_norm": 0.1565493494272232,
      "learning_rate": 1.6525942416740043e-05,
      "loss": 0.6762,
      "mean_token_accuracy": 0.7980387926101684,
      "num_tokens": 4065739907.0,
      "step": 1955
    },
    {
      "epoch": 1.9071663219369754,
      "grad_norm": 0.16129525005817413,
      "learning_rate": 1.6398582608763457e-05,
      "loss": 0.6769,
      "mean_token_accuracy": 0.7987368181347847,
      "num_tokens": 4076157906.0,
      "step": 1960
    },
    {
      "epoch": 1.9120330940503711,
      "grad_norm": 0.16715414822101593,
      "learning_rate": 1.6271475504690792e-05,
      "loss": 0.6817,
      "mean_token_accuracy": 0.7968519911170006,
      "num_tokens": 4086571576.0,
      "step": 1965
    },
    {
      "epoch": 1.9168998661637668,
      "grad_norm": 0.15474775433540344,
      "learning_rate": 1.6144624838843948e-05,
      "loss": 0.6809,
      "mean_token_accuracy": 0.7968852281570434,
      "num_tokens": 4096974491.0,
      "step": 1970
    },
    {
      "epoch": 1.9217666382771625,
      "grad_norm": 0.15702135860919952,
      "learning_rate": 1.6018034338010857e-05,
      "loss": 0.682,
      "mean_token_accuracy": 0.7966296806931495,
      "num_tokens": 4107366914.0,
      "step": 1975
    },
    {
      "epoch": 1.9266334103905585,
      "grad_norm": 0.15245595574378967,
      "learning_rate": 1.589170772133597e-05,
      "loss": 0.6879,
      "mean_token_accuracy": 0.7963499456644059,
      "num_tokens": 4117785208.0,
      "step": 1980
    },
    {
      "epoch": 1.9315001825039544,
      "grad_norm": 0.15890534222126007,
      "learning_rate": 1.5765648700211e-05,
      "loss": 0.6815,
      "mean_token_accuracy": 0.7967126056551933,
      "num_tokens": 4128193822.0,
      "step": 1985
    },
    {
      "epoch": 1.93636695461735,
      "grad_norm": 0.1601361185312271,
      "learning_rate": 1.563986097816588e-05,
      "loss": 0.6836,
      "mean_token_accuracy": 0.7961337313055992,
      "num_tokens": 4138598651.0,
      "step": 1990
    },
    {
      "epoch": 1.9412337267307458,
      "grad_norm": 0.18950778245925903,
      "learning_rate": 1.5514348250759936e-05,
      "loss": 0.6801,
      "mean_token_accuracy": 0.7971198186278343,
      "num_tokens": 4149015694.0,
      "step": 1995
    },
    {
      "epoch": 1.9461004988441415,
      "grad_norm": 0.18258991837501526,
      "learning_rate": 1.5389114205473377e-05,
      "loss": 0.6881,
      "mean_token_accuracy": 0.7972483649849892,
      "num_tokens": 4159419564.0,
      "step": 2000
    },
    {
      "epoch": 1.9509672709575374,
      "grad_norm": 0.17280957102775574,
      "learning_rate": 1.5264162521598892e-05,
      "loss": 0.68,
      "mean_token_accuracy": 0.7970993369817734,
      "num_tokens": 4169809506.0,
      "step": 2005
    },
    {
      "epoch": 1.9558340430709333,
      "grad_norm": 0.17790289223194122,
      "learning_rate": 1.5139496870133569e-05,
      "loss": 0.6857,
      "mean_token_accuracy": 0.7957724526524543,
      "num_tokens": 4180230729.0,
      "step": 2010
    },
    {
      "epoch": 1.960700815184329,
      "grad_norm": 0.15443438291549683,
      "learning_rate": 1.5015120913671089e-05,
      "loss": 0.6872,
      "mean_token_accuracy": 0.7955627977848053,
      "num_tokens": 4190654998.0,
      "step": 2015
    },
    {
      "epoch": 1.9655675872977247,
      "grad_norm": 0.14602826535701752,
      "learning_rate": 1.4891038306294047e-05,
      "loss": 0.6704,
      "mean_token_accuracy": 0.7992791101336479,
      "num_tokens": 4201048904.0,
      "step": 2020
    },
    {
      "epoch": 1.9704343594111204,
      "grad_norm": 0.15038786828517914,
      "learning_rate": 1.4767252693466665e-05,
      "loss": 0.6781,
      "mean_token_accuracy": 0.7973107099533081,
      "num_tokens": 4211435131.0,
      "step": 2025
    },
    {
      "epoch": 1.9753011315245164,
      "grad_norm": 0.14765934646129608,
      "learning_rate": 1.4643767711927663e-05,
      "loss": 0.6679,
      "mean_token_accuracy": 0.800252215564251,
      "num_tokens": 4221834653.0,
      "step": 2030
    },
    {
      "epoch": 1.9801679036379123,
      "grad_norm": 0.16181115806102753,
      "learning_rate": 1.4520586989583406e-05,
      "loss": 0.6747,
      "mean_token_accuracy": 0.7984942689538002,
      "num_tokens": 4232225962.0,
      "step": 2035
    },
    {
      "epoch": 1.985034675751308,
      "grad_norm": 0.1638074815273285,
      "learning_rate": 1.439771414540131e-05,
      "loss": 0.6763,
      "mean_token_accuracy": 0.7984781607985496,
      "num_tokens": 4242599952.0,
      "step": 2040
    },
    {
      "epoch": 1.9899014478647037,
      "grad_norm": 0.15550631284713745,
      "learning_rate": 1.4275152789303575e-05,
      "loss": 0.6803,
      "mean_token_accuracy": 0.7984331622719765,
      "num_tokens": 4253009854.0,
      "step": 2045
    },
    {
      "epoch": 1.9947682199780994,
      "grad_norm": 0.16920606791973114,
      "learning_rate": 1.4152906522061048e-05,
      "loss": 0.6801,
      "mean_token_accuracy": 0.7978967219591141,
      "num_tokens": 4263423876.0,
      "step": 2050
    },
    {
      "epoch": 1.9996349920914953,
      "grad_norm": 1.4035388231277466,
      "learning_rate": 1.4030978935187511e-05,
      "loss": 0.6836,
      "mean_token_accuracy": 0.7962289318442345,
      "num_tokens": 4273848158.0,
      "step": 2055
    },
    {
      "epoch": 2.003893417690717,
      "grad_norm": 0.21786171197891235,
      "learning_rate": 1.3909373610834114e-05,
      "loss": 0.6527,
      "mean_token_accuracy": 0.8026761446680342,
      "num_tokens": 4282933283.0,
      "step": 2060
    },
    {
      "epoch": 2.0087601898041125,
      "grad_norm": 0.1736161857843399,
      "learning_rate": 1.378809412168414e-05,
      "loss": 0.6531,
      "mean_token_accuracy": 0.802866980433464,
      "num_tokens": 4293336364.0,
      "step": 2065
    },
    {
      "epoch": 2.0136269619175082,
      "grad_norm": 0.18957297503948212,
      "learning_rate": 1.3667144030848073e-05,
      "loss": 0.6629,
      "mean_token_accuracy": 0.8001277893781662,
      "num_tokens": 4303725631.0,
      "step": 2070
    },
    {
      "epoch": 2.018493734030904,
      "grad_norm": 0.17591515183448792,
      "learning_rate": 1.3546526891758871e-05,
      "loss": 0.6428,
      "mean_token_accuracy": 0.8056652173399925,
      "num_tokens": 4314119143.0,
      "step": 2075
    },
    {
      "epoch": 2.0233605061442996,
      "grad_norm": 0.16420334577560425,
      "learning_rate": 1.3426246248067594e-05,
      "loss": 0.6477,
      "mean_token_accuracy": 0.8048857882618904,
      "num_tokens": 4324530251.0,
      "step": 2080
    },
    {
      "epoch": 2.0282272782576958,
      "grad_norm": 0.15316329896450043,
      "learning_rate": 1.3306305633539323e-05,
      "loss": 0.6589,
      "mean_token_accuracy": 0.8010796695947647,
      "num_tokens": 4334943481.0,
      "step": 2085
    },
    {
      "epoch": 2.0330940503710915,
      "grad_norm": 0.1632341593503952,
      "learning_rate": 1.3186708571949274e-05,
      "loss": 0.6365,
      "mean_token_accuracy": 0.8072337210178375,
      "num_tokens": 4345349851.0,
      "step": 2090
    },
    {
      "epoch": 2.037960822484487,
      "grad_norm": 0.16255219280719757,
      "learning_rate": 1.3067458576979304e-05,
      "loss": 0.6582,
      "mean_token_accuracy": 0.8018566101789475,
      "num_tokens": 4355708690.0,
      "step": 2095
    },
    {
      "epoch": 2.042827594597883,
      "grad_norm": 0.1518089920282364,
      "learning_rate": 1.294855915211472e-05,
      "loss": 0.642,
      "mean_token_accuracy": 0.8059766829013825,
      "num_tokens": 4366109640.0,
      "step": 2100
    },
    {
      "epoch": 2.0476943667112786,
      "grad_norm": 0.15770965814590454,
      "learning_rate": 1.2830013790541279e-05,
      "loss": 0.6399,
      "mean_token_accuracy": 0.8066899254918098,
      "num_tokens": 4376511093.0,
      "step": 2105
    },
    {
      "epoch": 2.0525611388246747,
      "grad_norm": 0.16652348637580872,
      "learning_rate": 1.271182597504263e-05,
      "loss": 0.655,
      "mean_token_accuracy": 0.8023440182209015,
      "num_tokens": 4386922108.0,
      "step": 2110
    },
    {
      "epoch": 2.0574279109380704,
      "grad_norm": 0.2297440767288208,
      "learning_rate": 1.259399917789792e-05,
      "loss": 0.6484,
      "mean_token_accuracy": 0.8046289414167405,
      "num_tokens": 4397345664.0,
      "step": 2115
    },
    {
      "epoch": 2.062294683051466,
      "grad_norm": 0.15444330871105194,
      "learning_rate": 1.2476536860779823e-05,
      "loss": 0.6446,
      "mean_token_accuracy": 0.8050382226705551,
      "num_tokens": 4407741004.0,
      "step": 2120
    },
    {
      "epoch": 2.067161455164862,
      "grad_norm": 0.15926706790924072,
      "learning_rate": 1.2359442474652863e-05,
      "loss": 0.643,
      "mean_token_accuracy": 0.8062428310513496,
      "num_tokens": 4418156436.0,
      "step": 2125
    },
    {
      "epoch": 2.0720282272782575,
      "grad_norm": 0.1727658212184906,
      "learning_rate": 1.2242719459671964e-05,
      "loss": 0.6491,
      "mean_token_accuracy": 0.8044039905071259,
      "num_tokens": 4428478333.0,
      "step": 2130
    },
    {
      "epoch": 2.0768949993916537,
      "grad_norm": 0.1601225733757019,
      "learning_rate": 1.2126371245081421e-05,
      "loss": 0.6525,
      "mean_token_accuracy": 0.8034905418753624,
      "num_tokens": 4438887919.0,
      "step": 2135
    },
    {
      "epoch": 2.0817617715050494,
      "grad_norm": 0.16613970696926117,
      "learning_rate": 1.2010401249114167e-05,
      "loss": 0.6556,
      "mean_token_accuracy": 0.802705030143261,
      "num_tokens": 4449277688.0,
      "step": 2140
    },
    {
      "epoch": 2.086628543618445,
      "grad_norm": 0.16443246603012085,
      "learning_rate": 1.1894812878891287e-05,
      "loss": 0.6597,
      "mean_token_accuracy": 0.8010829821228981,
      "num_tokens": 4459685420.0,
      "step": 2145
    },
    {
      "epoch": 2.0914953157318408,
      "grad_norm": 0.1574215143918991,
      "learning_rate": 1.1779609530321983e-05,
      "loss": 0.6461,
      "mean_token_accuracy": 0.8051062658429146,
      "num_tokens": 4470083834.0,
      "step": 2150
    },
    {
      "epoch": 2.0963620878452365,
      "grad_norm": 0.15328621864318848,
      "learning_rate": 1.166479458800378e-05,
      "loss": 0.6593,
      "mean_token_accuracy": 0.801020322740078,
      "num_tokens": 4480497284.0,
      "step": 2155
    },
    {
      "epoch": 2.1012288599586326,
      "grad_norm": 0.14920932054519653,
      "learning_rate": 1.1550371425123066e-05,
      "loss": 0.6556,
      "mean_token_accuracy": 0.8022570490837098,
      "num_tokens": 4490911381.0,
      "step": 2160
    },
    {
      "epoch": 2.1060956320720283,
      "grad_norm": 0.16176725924015045,
      "learning_rate": 1.1436343403356017e-05,
      "loss": 0.6542,
      "mean_token_accuracy": 0.8028280824422837,
      "num_tokens": 4501307109.0,
      "step": 2165
    },
    {
      "epoch": 2.110962404185424,
      "grad_norm": 0.1514625996351242,
      "learning_rate": 1.1322713872769843e-05,
      "loss": 0.6467,
      "mean_token_accuracy": 0.8045107677578927,
      "num_tokens": 4511715316.0,
      "step": 2170
    },
    {
      "epoch": 2.1158291762988197,
      "grad_norm": 0.15835708379745483,
      "learning_rate": 1.120948617172432e-05,
      "loss": 0.6547,
      "mean_token_accuracy": 0.8024782687425613,
      "num_tokens": 4522096402.0,
      "step": 2175
    },
    {
      "epoch": 2.1206959484122154,
      "grad_norm": 0.15293635427951813,
      "learning_rate": 1.1096663626773778e-05,
      "loss": 0.6585,
      "mean_token_accuracy": 0.8039514124393463,
      "num_tokens": 4532505059.0,
      "step": 2180
    },
    {
      "epoch": 2.1255627205256116,
      "grad_norm": 0.1580297350883484,
      "learning_rate": 1.098424955256929e-05,
      "loss": 0.6513,
      "mean_token_accuracy": 0.8036407768726349,
      "num_tokens": 4542923337.0,
      "step": 2185
    },
    {
      "epoch": 2.1304294926390073,
      "grad_norm": 0.15784800052642822,
      "learning_rate": 1.0872247251761361e-05,
      "loss": 0.6544,
      "mean_token_accuracy": 0.8028563886880875,
      "num_tokens": 4553308100.0,
      "step": 2190
    },
    {
      "epoch": 2.135296264752403,
      "grad_norm": 0.1474941074848175,
      "learning_rate": 1.0760660014902864e-05,
      "loss": 0.6534,
      "mean_token_accuracy": 0.8028626918792725,
      "num_tokens": 4563697921.0,
      "step": 2195
    },
    {
      "epoch": 2.1401630368657987,
      "grad_norm": 0.15973593294620514,
      "learning_rate": 1.0649491120352363e-05,
      "loss": 0.6659,
      "mean_token_accuracy": 0.7996205613017082,
      "num_tokens": 4574123265.0,
      "step": 2200
    },
    {
      "epoch": 2.1450298089791944,
      "grad_norm": 0.15296930074691772,
      "learning_rate": 1.053874383417779e-05,
      "loss": 0.6535,
      "mean_token_accuracy": 0.8034606859087944,
      "num_tokens": 4584498086.0,
      "step": 2205
    },
    {
      "epoch": 2.1498965810925905,
      "grad_norm": 0.14859797060489655,
      "learning_rate": 1.0428421410060541e-05,
      "loss": 0.6536,
      "mean_token_accuracy": 0.8030571714043617,
      "num_tokens": 4594917917.0,
      "step": 2210
    },
    {
      "epoch": 2.154763353205986,
      "grad_norm": 0.15155349671840668,
      "learning_rate": 1.0318527089199815e-05,
      "loss": 0.6608,
      "mean_token_accuracy": 0.8009442865848542,
      "num_tokens": 4605325604.0,
      "step": 2215
    },
    {
      "epoch": 2.159630125319382,
      "grad_norm": 0.14784498512744904,
      "learning_rate": 1.0209064100217433e-05,
      "loss": 0.6452,
      "mean_token_accuracy": 0.8052634865045547,
      "num_tokens": 4615748432.0,
      "step": 2220
    },
    {
      "epoch": 2.1644968974327776,
      "grad_norm": 0.15303626656532288,
      "learning_rate": 1.0100035659062988e-05,
      "loss": 0.6523,
      "mean_token_accuracy": 0.8031300380825996,
      "num_tokens": 4626175659.0,
      "step": 2225
    },
    {
      "epoch": 2.1693636695461733,
      "grad_norm": 0.14879605174064636,
      "learning_rate": 9.991444968919317e-06,
      "loss": 0.6653,
      "mean_token_accuracy": 0.799750280380249,
      "num_tokens": 4636596138.0,
      "step": 2230
    },
    {
      "epoch": 2.1742304416595695,
      "grad_norm": 0.1621440201997757,
      "learning_rate": 9.883295220108454e-06,
      "loss": 0.6489,
      "mean_token_accuracy": 0.804102037847042,
      "num_tokens": 4646998708.0,
      "step": 2235
    },
    {
      "epoch": 2.179097213772965,
      "grad_norm": 0.15081925690174103,
      "learning_rate": 9.775589589997869e-06,
      "loss": 0.6503,
      "mean_token_accuracy": 0.8036722138524055,
      "num_tokens": 4657372930.0,
      "step": 2240
    },
    {
      "epoch": 2.183963985886361,
      "grad_norm": 5.372561454772949,
      "learning_rate": 9.668331242907089e-06,
      "loss": 0.6744,
      "mean_token_accuracy": 0.8020758613944053,
      "num_tokens": 4667773596.0,
      "step": 2245
    },
    {
      "epoch": 2.1888307579997566,
      "grad_norm": 0.179043710231781,
      "learning_rate": 9.56152333001481e-06,
      "loss": 0.6497,
      "mean_token_accuracy": 0.804103071987629,
      "num_tokens": 4678189224.0,
      "step": 2250
    },
    {
      "epoch": 2.1936975301131523,
      "grad_norm": 0.15939339995384216,
      "learning_rate": 9.455168989266237e-06,
      "loss": 0.665,
      "mean_token_accuracy": 0.8003452822566033,
      "num_tokens": 4688577434.0,
      "step": 2255
    },
    {
      "epoch": 2.1985643022265484,
      "grad_norm": 0.16028191149234772,
      "learning_rate": 9.349271345280927e-06,
      "loss": 0.6498,
      "mean_token_accuracy": 0.8038272604346275,
      "num_tokens": 4698981929.0,
      "step": 2260
    },
    {
      "epoch": 2.203431074339944,
      "grad_norm": 0.1571078896522522,
      "learning_rate": 9.243833509261029e-06,
      "loss": 0.6415,
      "mean_token_accuracy": 0.8061452865600586,
      "num_tokens": 4709326648.0,
      "step": 2265
    },
    {
      "epoch": 2.20829784645334,
      "grad_norm": 0.152186319231987,
      "learning_rate": 9.138858578899789e-06,
      "loss": 0.6516,
      "mean_token_accuracy": 0.8032000198960304,
      "num_tokens": 4719720199.0,
      "step": 2270
    },
    {
      "epoch": 2.2131646185667355,
      "grad_norm": 0.15774431824684143,
      "learning_rate": 9.034349638290642e-06,
      "loss": 0.6457,
      "mean_token_accuracy": 0.8048562005162239,
      "num_tokens": 4730122155.0,
      "step": 2275
    },
    {
      "epoch": 2.218031390680131,
      "grad_norm": 0.15229646861553192,
      "learning_rate": 8.930309757836517e-06,
      "loss": 0.6517,
      "mean_token_accuracy": 0.8039355188608169,
      "num_tokens": 4740461429.0,
      "step": 2280
    },
    {
      "epoch": 2.2228981627935274,
      "grad_norm": 0.1564294546842575,
      "learning_rate": 8.826741994159696e-06,
      "loss": 0.6517,
      "mean_token_accuracy": 0.8037168741226196,
      "num_tokens": 4750836580.0,
      "step": 2285
    },
    {
      "epoch": 2.227764934906923,
      "grad_norm": 0.1520540416240692,
      "learning_rate": 8.723649390011961e-06,
      "loss": 0.6352,
      "mean_token_accuracy": 0.8080217778682709,
      "num_tokens": 4761258689.0,
      "step": 2290
    },
    {
      "epoch": 2.2326317070203188,
      "grad_norm": 0.15401072800159454,
      "learning_rate": 8.62103497418525e-06,
      "loss": 0.6521,
      "mean_token_accuracy": 0.8032583862543106,
      "num_tokens": 4771642940.0,
      "step": 2295
    },
    {
      "epoch": 2.2374984791337145,
      "grad_norm": 0.15026535093784332,
      "learning_rate": 8.518901761422618e-06,
      "loss": 0.6496,
      "mean_token_accuracy": 0.8039778783917427,
      "num_tokens": 4782026422.0,
      "step": 2300
    },
    {
      "epoch": 2.24236525124711,
      "grad_norm": 0.15299025177955627,
      "learning_rate": 8.41725275232973e-06,
      "loss": 0.6427,
      "mean_token_accuracy": 0.805833987891674,
      "num_tokens": 4792427090.0,
      "step": 2305
    },
    {
      "epoch": 2.2472320233605063,
      "grad_norm": 0.15790744125843048,
      "learning_rate": 8.316090933286638e-06,
      "loss": 0.6608,
      "mean_token_accuracy": 0.8008003666996956,
      "num_tokens": 4802821178.0,
      "step": 2310
    },
    {
      "epoch": 2.252098795473902,
      "grad_norm": 0.1489599347114563,
      "learning_rate": 8.215419276360084e-06,
      "loss": 0.6413,
      "mean_token_accuracy": 0.8072931855916977,
      "num_tokens": 4813250065.0,
      "step": 2315
    },
    {
      "epoch": 2.2569655675872977,
      "grad_norm": 1.0523755550384521,
      "learning_rate": 8.115240739216181e-06,
      "loss": 0.6438,
      "mean_token_accuracy": 0.8055999353528023,
      "num_tokens": 4823671176.0,
      "step": 2320
    },
    {
      "epoch": 2.2618323397006934,
      "grad_norm": 0.1581067144870758,
      "learning_rate": 8.015558265033526e-06,
      "loss": 0.6482,
      "mean_token_accuracy": 0.8043972551822662,
      "num_tokens": 4834102213.0,
      "step": 2325
    },
    {
      "epoch": 2.266699111814089,
      "grad_norm": 0.9845110774040222,
      "learning_rate": 7.916374782416678e-06,
      "loss": 0.6492,
      "mean_token_accuracy": 0.8050491586327553,
      "num_tokens": 4844520185.0,
      "step": 2330
    },
    {
      "epoch": 2.2715658839274853,
      "grad_norm": 0.15200099349021912,
      "learning_rate": 7.817693205310194e-06,
      "loss": 0.6517,
      "mean_token_accuracy": 0.8034302070736885,
      "num_tokens": 4854899124.0,
      "step": 2335
    },
    {
      "epoch": 2.276432656040881,
      "grad_norm": 0.1534387767314911,
      "learning_rate": 7.719516432912962e-06,
      "loss": 0.6488,
      "mean_token_accuracy": 0.8041450127959251,
      "num_tokens": 4865284283.0,
      "step": 2340
    },
    {
      "epoch": 2.2812994281542767,
      "grad_norm": 0.1607615202665329,
      "learning_rate": 7.621847349593045e-06,
      "loss": 0.6551,
      "mean_token_accuracy": 0.8025413408875466,
      "num_tokens": 4875692137.0,
      "step": 2345
    },
    {
      "epoch": 2.2861662002676724,
      "grad_norm": 0.15113408863544464,
      "learning_rate": 7.524688824802953e-06,
      "loss": 0.6489,
      "mean_token_accuracy": 0.8042253971099853,
      "num_tokens": 4886093164.0,
      "step": 2350
    },
    {
      "epoch": 2.291032972381068,
      "grad_norm": 0.21686075627803802,
      "learning_rate": 7.428043712995305e-06,
      "loss": 0.6588,
      "mean_token_accuracy": 0.801556596159935,
      "num_tokens": 4896521953.0,
      "step": 2355
    },
    {
      "epoch": 2.295899744494464,
      "grad_norm": 0.14993663132190704,
      "learning_rate": 7.331914853539021e-06,
      "loss": 0.6398,
      "mean_token_accuracy": 0.8062152758240699,
      "num_tokens": 4906937622.0,
      "step": 2360
    },
    {
      "epoch": 2.30076651660786,
      "grad_norm": 0.14678749442100525,
      "learning_rate": 7.236305070635835e-06,
      "loss": 0.6514,
      "mean_token_accuracy": 0.8036011949181556,
      "num_tokens": 4917335385.0,
      "step": 2365
    },
    {
      "epoch": 2.3056332887212556,
      "grad_norm": 0.1479840725660324,
      "learning_rate": 7.1412171732373825e-06,
      "loss": 0.647,
      "mean_token_accuracy": 0.804350021481514,
      "num_tokens": 4927730661.0,
      "step": 2370
    },
    {
      "epoch": 2.3105000608346513,
      "grad_norm": 0.145925834774971,
      "learning_rate": 7.046653954962654e-06,
      "loss": 0.6528,
      "mean_token_accuracy": 0.8029148191213608,
      "num_tokens": 4938161011.0,
      "step": 2375
    },
    {
      "epoch": 2.315366832948047,
      "grad_norm": 0.14979971945285797,
      "learning_rate": 6.952618194015898e-06,
      "loss": 0.6592,
      "mean_token_accuracy": 0.8012317448854447,
      "num_tokens": 4948576011.0,
      "step": 2380
    },
    {
      "epoch": 2.320233605061443,
      "grad_norm": 0.1420050710439682,
      "learning_rate": 6.859112653105024e-06,
      "loss": 0.6667,
      "mean_token_accuracy": 0.800655497610569,
      "num_tokens": 4958969618.0,
      "step": 2385
    },
    {
      "epoch": 2.325100377174839,
      "grad_norm": 0.14455583691596985,
      "learning_rate": 6.76614007936045e-06,
      "loss": 0.6657,
      "mean_token_accuracy": 0.799447487294674,
      "num_tokens": 4969406176.0,
      "step": 2390
    },
    {
      "epoch": 2.3299671492882346,
      "grad_norm": 0.157817542552948,
      "learning_rate": 6.673703204254347e-06,
      "loss": 0.6556,
      "mean_token_accuracy": 0.802377738058567,
      "num_tokens": 4979828873.0,
      "step": 2395
    },
    {
      "epoch": 2.3348339214016303,
      "grad_norm": 1.0549571514129639,
      "learning_rate": 6.581804743520431e-06,
      "loss": 0.6486,
      "mean_token_accuracy": 0.804625403881073,
      "num_tokens": 4990248635.0,
      "step": 2400
    },
    {
      "epoch": 2.3397006935150264,
      "grad_norm": 0.14948920905590057,
      "learning_rate": 6.4904473970741616e-06,
      "loss": 0.6402,
      "mean_token_accuracy": 0.8065480872988701,
      "num_tokens": 5000645108.0,
      "step": 2405
    },
    {
      "epoch": 2.344567465628422,
      "grad_norm": 0.15044322609901428,
      "learning_rate": 6.399633848933434e-06,
      "loss": 0.6549,
      "mean_token_accuracy": 0.8022510707378387,
      "num_tokens": 5011061969.0,
      "step": 2410
    },
    {
      "epoch": 2.349434237741818,
      "grad_norm": 0.1458985060453415,
      "learning_rate": 6.309366767139682e-06,
      "loss": 0.6554,
      "mean_token_accuracy": 0.8022111177444458,
      "num_tokens": 5021482013.0,
      "step": 2415
    },
    {
      "epoch": 2.3543010098552135,
      "grad_norm": 0.1474456936120987,
      "learning_rate": 6.219648803679559e-06,
      "loss": 0.6418,
      "mean_token_accuracy": 0.8059498369693756,
      "num_tokens": 5031909206.0,
      "step": 2420
    },
    {
      "epoch": 2.359167781968609,
      "grad_norm": 0.15672780573368073,
      "learning_rate": 6.130482594406953e-06,
      "loss": 0.6609,
      "mean_token_accuracy": 0.8012527599930763,
      "num_tokens": 5042280658.0,
      "step": 2425
    },
    {
      "epoch": 2.3640345540820054,
      "grad_norm": 0.15176284313201904,
      "learning_rate": 6.041870758965612e-06,
      "loss": 0.6573,
      "mean_token_accuracy": 0.8018319338560105,
      "num_tokens": 5052692899.0,
      "step": 2430
    },
    {
      "epoch": 2.368901326195401,
      "grad_norm": 0.14805342257022858,
      "learning_rate": 5.953815900712131e-06,
      "loss": 0.6535,
      "mean_token_accuracy": 0.8027032360434532,
      "num_tokens": 5063117153.0,
      "step": 2435
    },
    {
      "epoch": 2.3737680983087968,
      "grad_norm": 0.14822912216186523,
      "learning_rate": 5.866320606639492e-06,
      "loss": 0.6505,
      "mean_token_accuracy": 0.8035964325070382,
      "num_tokens": 5073516717.0,
      "step": 2440
    },
    {
      "epoch": 2.3786348704221925,
      "grad_norm": 0.14757058024406433,
      "learning_rate": 5.779387447301063e-06,
      "loss": 0.6455,
      "mean_token_accuracy": 0.8051330745220184,
      "num_tokens": 5083916063.0,
      "step": 2445
    },
    {
      "epoch": 2.383501642535588,
      "grad_norm": 0.14956192672252655,
      "learning_rate": 5.693018976735051e-06,
      "loss": 0.6569,
      "mean_token_accuracy": 0.801778431236744,
      "num_tokens": 5094306145.0,
      "step": 2450
    },
    {
      "epoch": 2.3883684146489843,
      "grad_norm": 0.1488412469625473,
      "learning_rate": 5.607217732389503e-06,
      "loss": 0.6514,
      "mean_token_accuracy": 0.8033558040857315,
      "num_tokens": 5104674954.0,
      "step": 2455
    },
    {
      "epoch": 2.39323518676238,
      "grad_norm": 0.14660821855068207,
      "learning_rate": 5.521986235047733e-06,
      "loss": 0.6573,
      "mean_token_accuracy": 0.8016549989581108,
      "num_tokens": 5115094718.0,
      "step": 2460
    },
    {
      "epoch": 2.3981019588757757,
      "grad_norm": 0.14319941401481628,
      "learning_rate": 5.43732698875426e-06,
      "loss": 0.6516,
      "mean_token_accuracy": 0.8032140851020813,
      "num_tokens": 5125437755.0,
      "step": 2465
    },
    {
      "epoch": 2.4029687309891714,
      "grad_norm": 0.1451750099658966,
      "learning_rate": 5.353242480741247e-06,
      "loss": 0.6486,
      "mean_token_accuracy": 0.8044626891613007,
      "num_tokens": 5135840151.0,
      "step": 2470
    },
    {
      "epoch": 2.407835503102567,
      "grad_norm": 0.14004506170749664,
      "learning_rate": 5.269735181355448e-06,
      "loss": 0.6483,
      "mean_token_accuracy": 0.8048574581742287,
      "num_tokens": 5146268062.0,
      "step": 2475
    },
    {
      "epoch": 2.4127022752159633,
      "grad_norm": 0.15180236101150513,
      "learning_rate": 5.186807543985597e-06,
      "loss": 0.6463,
      "mean_token_accuracy": 0.8046328857541084,
      "num_tokens": 5156649012.0,
      "step": 2480
    },
    {
      "epoch": 2.417569047329359,
      "grad_norm": 0.15156681835651398,
      "learning_rate": 5.104462004990343e-06,
      "loss": 0.6526,
      "mean_token_accuracy": 0.8029925510287285,
      "num_tokens": 5167048729.0,
      "step": 2485
    },
    {
      "epoch": 2.4224358194427547,
      "grad_norm": 0.14392004907131195,
      "learning_rate": 5.022700983626691e-06,
      "loss": 0.6471,
      "mean_token_accuracy": 0.8045663341879845,
      "num_tokens": 5177440863.0,
      "step": 2490
    },
    {
      "epoch": 2.4273025915561504,
      "grad_norm": 0.1472492665052414,
      "learning_rate": 4.941526881978881e-06,
      "loss": 0.6494,
      "mean_token_accuracy": 0.8037728562951088,
      "num_tokens": 5187853616.0,
      "step": 2495
    },
    {
      "epoch": 2.432169363669546,
      "grad_norm": 0.14200709760189056,
      "learning_rate": 4.860942084887868e-06,
      "loss": 0.6502,
      "mean_token_accuracy": 0.8035871312022209,
      "num_tokens": 5198238175.0,
      "step": 2500
    },
    {
      "epoch": 2.437036135782942,
      "grad_norm": 0.144990012049675,
      "learning_rate": 4.7809489598812265e-06,
      "loss": 0.6464,
      "mean_token_accuracy": 0.8049494594335556,
      "num_tokens": 5208639498.0,
      "step": 2505
    },
    {
      "epoch": 2.441902907896338,
      "grad_norm": 0.1452317088842392,
      "learning_rate": 4.701549857103588e-06,
      "loss": 0.6521,
      "mean_token_accuracy": 0.8031940594315529,
      "num_tokens": 5219048645.0,
      "step": 2510
    },
    {
      "epoch": 2.4467696800097336,
      "grad_norm": 0.14874960482120514,
      "learning_rate": 4.622747109247622e-06,
      "loss": 0.6455,
      "mean_token_accuracy": 0.8050684958696366,
      "num_tokens": 5229450112.0,
      "step": 2515
    },
    {
      "epoch": 2.4516364521231293,
      "grad_norm": 0.14428818225860596,
      "learning_rate": 4.544543031485479e-06,
      "loss": 0.6505,
      "mean_token_accuracy": 0.8035919606685639,
      "num_tokens": 5239857289.0,
      "step": 2520
    },
    {
      "epoch": 2.456503224236525,
      "grad_norm": 0.14049416780471802,
      "learning_rate": 4.4669399214007785e-06,
      "loss": 0.6571,
      "mean_token_accuracy": 0.8018587455153465,
      "num_tokens": 5250254325.0,
      "step": 2525
    },
    {
      "epoch": 2.461369996349921,
      "grad_norm": 0.1468386948108673,
      "learning_rate": 4.389940058921127e-06,
      "loss": 0.6449,
      "mean_token_accuracy": 0.8047372192144394,
      "num_tokens": 5260669564.0,
      "step": 2530
    },
    {
      "epoch": 2.466236768463317,
      "grad_norm": 0.14413408935070038,
      "learning_rate": 4.313545706251096e-06,
      "loss": 0.6494,
      "mean_token_accuracy": 0.8038429006934166,
      "num_tokens": 5271055345.0,
      "step": 2535
    },
    {
      "epoch": 2.4711035405767126,
      "grad_norm": 0.14168906211853027,
      "learning_rate": 4.237759107805802e-06,
      "loss": 0.6469,
      "mean_token_accuracy": 0.804447378218174,
      "num_tokens": 5281450924.0,
      "step": 2540
    },
    {
      "epoch": 2.4759703126901083,
      "grad_norm": 0.14093631505966187,
      "learning_rate": 4.162582490144948e-06,
      "loss": 0.6531,
      "mean_token_accuracy": 0.8025687426328659,
      "num_tokens": 5291871839.0,
      "step": 2545
    },
    {
      "epoch": 2.480837084803504,
      "grad_norm": 0.13964799046516418,
      "learning_rate": 4.088018061907389e-06,
      "loss": 0.6576,
      "mean_token_accuracy": 0.8046013057231903,
      "num_tokens": 5302244487.0,
      "step": 2550
    },
    {
      "epoch": 2.4857038569169,
      "grad_norm": 2.7812206745147705,
      "learning_rate": 4.014068013746289e-06,
      "loss": 0.6576,
      "mean_token_accuracy": 0.8021691858768463,
      "num_tokens": 5312640574.0,
      "step": 2555
    },
    {
      "epoch": 2.490570629030296,
      "grad_norm": 0.14239361882209778,
      "learning_rate": 3.940734518264713e-06,
      "loss": 0.6536,
      "mean_token_accuracy": 0.803150224685669,
      "num_tokens": 5323062158.0,
      "step": 2560
    },
    {
      "epoch": 2.4954374011436915,
      "grad_norm": 0.14358608424663544,
      "learning_rate": 3.8680197299518244e-06,
      "loss": 0.654,
      "mean_token_accuracy": 0.8025683507323265,
      "num_tokens": 5333476751.0,
      "step": 2565
    },
    {
      "epoch": 2.500304173257087,
      "grad_norm": 0.1440841406583786,
      "learning_rate": 3.7959257851195974e-06,
      "loss": 0.6466,
      "mean_token_accuracy": 0.8044820740818978,
      "num_tokens": 5343875655.0,
      "step": 2570
    },
    {
      "epoch": 2.505170945370483,
      "grad_norm": 0.14689917862415314,
      "learning_rate": 3.7244548018400245e-06,
      "loss": 0.6458,
      "mean_token_accuracy": 0.8048086851835251,
      "num_tokens": 5354300701.0,
      "step": 2575
    },
    {
      "epoch": 2.510037717483879,
      "grad_norm": 0.14068074524402618,
      "learning_rate": 3.653608879882903e-06,
      "loss": 0.6536,
      "mean_token_accuracy": 0.8028267294168472,
      "num_tokens": 5364714992.0,
      "step": 2580
    },
    {
      "epoch": 2.5149044895972748,
      "grad_norm": 0.13912948966026306,
      "learning_rate": 3.583390100654152e-06,
      "loss": 0.6565,
      "mean_token_accuracy": 0.8021206289529801,
      "num_tokens": 5375131937.0,
      "step": 2585
    },
    {
      "epoch": 2.5197712617106705,
      "grad_norm": 0.14009574055671692,
      "learning_rate": 3.5138005271346643e-06,
      "loss": 0.6432,
      "mean_token_accuracy": 0.8057182282209396,
      "num_tokens": 5385557897.0,
      "step": 2590
    },
    {
      "epoch": 2.524638033824066,
      "grad_norm": 0.1442336142063141,
      "learning_rate": 3.444842203819662e-06,
      "loss": 0.6556,
      "mean_token_accuracy": 0.8037760138511658,
      "num_tokens": 5395967155.0,
      "step": 2595
    },
    {
      "epoch": 2.529504805937462,
      "grad_norm": 0.1399046927690506,
      "learning_rate": 3.3765171566586903e-06,
      "loss": 0.6475,
      "mean_token_accuracy": 0.8044023871421814,
      "num_tokens": 5406381919.0,
      "step": 2600
    },
    {
      "epoch": 2.534371578050858,
      "grad_norm": 0.14351284503936768,
      "learning_rate": 3.308827392996039e-06,
      "loss": 0.6485,
      "mean_token_accuracy": 0.8046146601438522,
      "num_tokens": 5416790850.0,
      "step": 2605
    },
    {
      "epoch": 2.5392383501642537,
      "grad_norm": 0.14063309133052826,
      "learning_rate": 3.2417749015117933e-06,
      "loss": 0.6403,
      "mean_token_accuracy": 0.8065345793962478,
      "num_tokens": 5427215731.0,
      "step": 2610
    },
    {
      "epoch": 2.5441051222776494,
      "grad_norm": 0.1423674374818802,
      "learning_rate": 3.1753616521634248e-06,
      "loss": 0.6528,
      "mean_token_accuracy": 0.8031722411513329,
      "num_tokens": 5437633265.0,
      "step": 2615
    },
    {
      "epoch": 2.548971894391045,
      "grad_norm": 0.14288178086280823,
      "learning_rate": 3.1095895961278793e-06,
      "loss": 0.6389,
      "mean_token_accuracy": 0.8069746538996696,
      "num_tokens": 5448051911.0,
      "step": 2620
    },
    {
      "epoch": 2.553838666504441,
      "grad_norm": 0.14691947400569916,
      "learning_rate": 3.044460665744284e-06,
      "loss": 0.6553,
      "mean_token_accuracy": 0.8021443754434585,
      "num_tokens": 5458435489.0,
      "step": 2625
    },
    {
      "epoch": 2.558705438617837,
      "grad_norm": 0.14588817954063416,
      "learning_rate": 2.9799767744571632e-06,
      "loss": 0.6543,
      "mean_token_accuracy": 0.8022594258189202,
      "num_tokens": 5468844675.0,
      "step": 2630
    },
    {
      "epoch": 2.5635722107312326,
      "grad_norm": 0.14357538521289825,
      "learning_rate": 2.916139816760205e-06,
      "loss": 0.6467,
      "mean_token_accuracy": 0.8045730352401733,
      "num_tokens": 5479260155.0,
      "step": 2635
    },
    {
      "epoch": 2.5684389828446283,
      "grad_norm": 0.5343859195709229,
      "learning_rate": 2.8529516681406465e-06,
      "loss": 0.6637,
      "mean_token_accuracy": 0.8032255545258522,
      "num_tokens": 5489656569.0,
      "step": 2640
    },
    {
      "epoch": 2.573305754958024,
      "grad_norm": 0.14265379309654236,
      "learning_rate": 2.790414185024126e-06,
      "loss": 0.65,
      "mean_token_accuracy": 0.8036694243550301,
      "num_tokens": 5500082764.0,
      "step": 2645
    },
    {
      "epoch": 2.5781725270714198,
      "grad_norm": 0.14276863634586334,
      "learning_rate": 2.7285292047201706e-06,
      "loss": 0.6392,
      "mean_token_accuracy": 0.8070199906826019,
      "num_tokens": 5510478604.0,
      "step": 2650
    },
    {
      "epoch": 2.583039299184816,
      "grad_norm": 0.14097563922405243,
      "learning_rate": 2.6672985453682158e-06,
      "loss": 0.6365,
      "mean_token_accuracy": 0.8079631119966507,
      "num_tokens": 5520856268.0,
      "step": 2655
    },
    {
      "epoch": 2.5879060712982116,
      "grad_norm": 0.14179939031600952,
      "learning_rate": 2.606724005884176e-06,
      "loss": 0.6555,
      "mean_token_accuracy": 0.8013408437371254,
      "num_tokens": 5531257919.0,
      "step": 2660
    },
    {
      "epoch": 2.5927728434116073,
      "grad_norm": 0.14454999566078186,
      "learning_rate": 2.5468073659076e-06,
      "loss": 0.6526,
      "mean_token_accuracy": 0.8032438889145851,
      "num_tokens": 5541672456.0,
      "step": 2665
    },
    {
      "epoch": 2.597639615525003,
      "grad_norm": 0.14427150785923004,
      "learning_rate": 2.487550385749396e-06,
      "loss": 0.6387,
      "mean_token_accuracy": 0.8072917461395264,
      "num_tokens": 5552045017.0,
      "step": 2670
    },
    {
      "epoch": 2.6025063876383987,
      "grad_norm": 0.14285337924957275,
      "learning_rate": 2.4289548063401084e-06,
      "loss": 0.6684,
      "mean_token_accuracy": 0.8027083098888397,
      "num_tokens": 5562425802.0,
      "step": 2675
    },
    {
      "epoch": 2.607373159751795,
      "grad_norm": 0.1397463083267212,
      "learning_rate": 2.371022349178764e-06,
      "loss": 0.6483,
      "mean_token_accuracy": 0.8042449176311492,
      "num_tokens": 5572823313.0,
      "step": 2680
    },
    {
      "epoch": 2.6122399318651905,
      "grad_norm": 0.14182040095329285,
      "learning_rate": 2.313754716282296e-06,
      "loss": 0.6509,
      "mean_token_accuracy": 0.8034815952181816,
      "num_tokens": 5583220155.0,
      "step": 2685
    },
    {
      "epoch": 2.6171067039785862,
      "grad_norm": 0.14046069979667664,
      "learning_rate": 2.2571535901355466e-06,
      "loss": 0.6508,
      "mean_token_accuracy": 0.8032889619469643,
      "num_tokens": 5593618711.0,
      "step": 2690
    },
    {
      "epoch": 2.621973476091982,
      "grad_norm": 0.14157386124134064,
      "learning_rate": 2.2012206336418455e-06,
      "loss": 0.6437,
      "mean_token_accuracy": 0.8052166134119034,
      "num_tokens": 5604023193.0,
      "step": 2695
    },
    {
      "epoch": 2.6268402482053776,
      "grad_norm": 0.13742253184318542,
      "learning_rate": 2.14595749007413e-06,
      "loss": 0.6433,
      "mean_token_accuracy": 0.805757936835289,
      "num_tokens": 5614434149.0,
      "step": 2700
    },
    {
      "epoch": 2.631707020318774,
      "grad_norm": 0.13803309202194214,
      "learning_rate": 2.0913657830266774e-06,
      "loss": 0.6425,
      "mean_token_accuracy": 0.8057522490620613,
      "num_tokens": 5624842003.0,
      "step": 2705
    },
    {
      "epoch": 2.6365737924321695,
      "grad_norm": 0.13676245510578156,
      "learning_rate": 2.03744711636743e-06,
      "loss": 0.6506,
      "mean_token_accuracy": 0.8036410436034203,
      "num_tokens": 5635274215.0,
      "step": 2710
    },
    {
      "epoch": 2.641440564545565,
      "grad_norm": 0.14218202233314514,
      "learning_rate": 1.9842030741908297e-06,
      "loss": 0.6496,
      "mean_token_accuracy": 0.8039317727088928,
      "num_tokens": 5645707043.0,
      "step": 2715
    },
    {
      "epoch": 2.646307336658961,
      "grad_norm": 0.13769201934337616,
      "learning_rate": 1.9316352207713105e-06,
      "loss": 0.6461,
      "mean_token_accuracy": 0.8048149377107621,
      "num_tokens": 5656104242.0,
      "step": 2720
    },
    {
      "epoch": 2.6511741087723566,
      "grad_norm": 0.13767434656620026,
      "learning_rate": 1.8797451005173384e-06,
      "loss": 0.6506,
      "mean_token_accuracy": 0.8042959049344063,
      "num_tokens": 5666510412.0,
      "step": 2725
    },
    {
      "epoch": 2.6560408808857527,
      "grad_norm": 0.13662338256835938,
      "learning_rate": 1.828534237926019e-06,
      "loss": 0.6562,
      "mean_token_accuracy": 0.8018163040280342,
      "num_tokens": 5676913518.0,
      "step": 2730
    },
    {
      "epoch": 2.6609076529991484,
      "grad_norm": 0.14092963933944702,
      "learning_rate": 1.778004137538325e-06,
      "loss": 0.6342,
      "mean_token_accuracy": 0.8087733536958694,
      "num_tokens": 5687296529.0,
      "step": 2735
    },
    {
      "epoch": 2.665774425112544,
      "grad_norm": 0.13853082060813904,
      "learning_rate": 1.7281562838948966e-06,
      "loss": 0.6403,
      "mean_token_accuracy": 0.8062141984701157,
      "num_tokens": 5697713608.0,
      "step": 2740
    },
    {
      "epoch": 2.67064119722594,
      "grad_norm": 0.14293664693832397,
      "learning_rate": 1.6789921414924054e-06,
      "loss": 0.6394,
      "mean_token_accuracy": 0.8064265713095665,
      "num_tokens": 5708107360.0,
      "step": 2745
    },
    {
      "epoch": 2.6755079693393355,
      "grad_norm": 0.14067302644252777,
      "learning_rate": 1.6305131547405621e-06,
      "loss": 0.6469,
      "mean_token_accuracy": 0.8043754413723946,
      "num_tokens": 5718500415.0,
      "step": 2750
    },
    {
      "epoch": 2.6803747414527317,
      "grad_norm": 0.1366061270236969,
      "learning_rate": 1.5827207479196443e-06,
      "loss": 0.6414,
      "mean_token_accuracy": 0.8060438290238381,
      "num_tokens": 5728916524.0,
      "step": 2755
    },
    {
      "epoch": 2.6852415135661274,
      "grad_norm": 0.13607634603977203,
      "learning_rate": 1.535616325138678e-06,
      "loss": 0.6388,
      "mean_token_accuracy": 0.8068518042564392,
      "num_tokens": 5739312774.0,
      "step": 2760
    },
    {
      "epoch": 2.690108285679523,
      "grad_norm": 0.14143404364585876,
      "learning_rate": 1.4892012702941799e-06,
      "loss": 0.6382,
      "mean_token_accuracy": 0.8065834864974022,
      "num_tokens": 5749707388.0,
      "step": 2765
    },
    {
      "epoch": 2.694975057792919,
      "grad_norm": 0.14009599387645721,
      "learning_rate": 1.44347694702949e-06,
      "loss": 0.6583,
      "mean_token_accuracy": 0.8027338907122612,
      "num_tokens": 5760123521.0,
      "step": 2770
    },
    {
      "epoch": 2.6998418299063145,
      "grad_norm": 0.1366027593612671,
      "learning_rate": 1.3984446986947093e-06,
      "loss": 0.6495,
      "mean_token_accuracy": 0.8034784376621247,
      "num_tokens": 5770532784.0,
      "step": 2775
    },
    {
      "epoch": 2.7047086020197106,
      "grad_norm": 0.13968665897846222,
      "learning_rate": 1.3541058483072605e-06,
      "loss": 0.6402,
      "mean_token_accuracy": 0.8061806738376618,
      "num_tokens": 5780914933.0,
      "step": 2780
    },
    {
      "epoch": 2.7095753741331063,
      "grad_norm": 0.13965339958667755,
      "learning_rate": 1.3104616985129796e-06,
      "loss": 0.6431,
      "mean_token_accuracy": 0.8058130428195,
      "num_tokens": 5791288795.0,
      "step": 2785
    },
    {
      "epoch": 2.714442146246502,
      "grad_norm": 0.13735094666481018,
      "learning_rate": 1.2675135315478592e-06,
      "loss": 0.6514,
      "mean_token_accuracy": 0.8053650557994843,
      "num_tokens": 5801654051.0,
      "step": 2790
    },
    {
      "epoch": 2.7193089183598977,
      "grad_norm": 0.13577793538570404,
      "learning_rate": 1.2252626092004005e-06,
      "loss": 0.6476,
      "mean_token_accuracy": 0.8044538736343384,
      "num_tokens": 5812053213.0,
      "step": 2795
    },
    {
      "epoch": 2.7241756904732934,
      "grad_norm": 0.1394418329000473,
      "learning_rate": 1.183710172774502e-06,
      "loss": 0.6542,
      "mean_token_accuracy": 0.8019895941019058,
      "num_tokens": 5822461220.0,
      "step": 2800
    },
    {
      "epoch": 2.7290424625866896,
      "grad_norm": 0.13514655828475952,
      "learning_rate": 1.1428574430530276e-06,
      "loss": 0.6529,
      "mean_token_accuracy": 0.8030089184641838,
      "num_tokens": 5832832477.0,
      "step": 2805
    },
    {
      "epoch": 2.7339092347000853,
      "grad_norm": 0.13452719151973724,
      "learning_rate": 1.102705620261918e-06,
      "loss": 0.6437,
      "mean_token_accuracy": 0.8053366273641587,
      "num_tokens": 5843259633.0,
      "step": 2810
    },
    {
      "epoch": 2.738776006813481,
      "grad_norm": 0.13801424205303192,
      "learning_rate": 1.0632558840349332e-06,
      "loss": 0.6482,
      "mean_token_accuracy": 0.8046306416392326,
      "num_tokens": 5853652033.0,
      "step": 2815
    },
    {
      "epoch": 2.7436427789268767,
      "grad_norm": 0.13894982635974884,
      "learning_rate": 1.0245093933790085e-06,
      "loss": 0.6392,
      "mean_token_accuracy": 0.8066315680742264,
      "num_tokens": 5864055200.0,
      "step": 2820
    },
    {
      "epoch": 2.7485095510402724,
      "grad_norm": 0.13615624606609344,
      "learning_rate": 9.86467286640183e-07,
      "loss": 0.6455,
      "mean_token_accuracy": 0.8048998638987541,
      "num_tokens": 5874471332.0,
      "step": 2825
    },
    {
      "epoch": 2.7533763231536685,
      "grad_norm": 0.13895970582962036,
      "learning_rate": 9.491306814701644e-07,
      "loss": 0.6518,
      "mean_token_accuracy": 0.8031819626688957,
      "num_tokens": 5884862217.0,
      "step": 2830
    },
    {
      "epoch": 2.7582430952670642,
      "grad_norm": 0.1360756903886795,
      "learning_rate": 9.125006747935122e-07,
      "loss": 0.6507,
      "mean_token_accuracy": 0.803553856909275,
      "num_tokens": 5895291181.0,
      "step": 2835
    },
    {
      "epoch": 2.76310986738046,
      "grad_norm": 0.14356902241706848,
      "learning_rate": 8.765783427753721e-07,
      "loss": 0.6589,
      "mean_token_accuracy": 0.8035019397735595,
      "num_tokens": 5905692203.0,
      "step": 2840
    },
    {
      "epoch": 2.7679766394938556,
      "grad_norm": 0.13687294721603394,
      "learning_rate": 8.413647407898989e-07,
      "loss": 0.6449,
      "mean_token_accuracy": 0.8053279563784599,
      "num_tokens": 5916105261.0,
      "step": 2845
    },
    {
      "epoch": 2.7728434116072513,
      "grad_norm": 2.104487657546997,
      "learning_rate": 8.068609033892172e-07,
      "loss": 0.6683,
      "mean_token_accuracy": 0.8018761411309242,
      "num_tokens": 5926512937.0,
      "step": 2850
    },
    {
      "epoch": 2.7777101837206475,
      "grad_norm": 0.1347740888595581,
      "learning_rate": 7.730678442730538e-07,
      "loss": 0.6543,
      "mean_token_accuracy": 0.8025069952011108,
      "num_tokens": 5936882445.0,
      "step": 2855
    },
    {
      "epoch": 2.782576955834043,
      "grad_norm": 0.1408262401819229,
      "learning_rate": 7.399865562589315e-07,
      "loss": 0.6403,
      "mean_token_accuracy": 0.8059840679168702,
      "num_tokens": 5947291010.0,
      "step": 2860
    },
    {
      "epoch": 2.787443727947439,
      "grad_norm": 0.13597391545772552,
      "learning_rate": 7.076180112530251e-07,
      "loss": 0.6514,
      "mean_token_accuracy": 0.8032924175262451,
      "num_tokens": 5957704983.0,
      "step": 2865
    },
    {
      "epoch": 2.7923105000608346,
      "grad_norm": 0.13730192184448242,
      "learning_rate": 6.759631602215877e-07,
      "loss": 0.6529,
      "mean_token_accuracy": 0.8028345763683319,
      "num_tokens": 5968091178.0,
      "step": 2870
    },
    {
      "epoch": 2.7971772721742303,
      "grad_norm": 0.14050757884979248,
      "learning_rate": 6.450229331630253e-07,
      "loss": 0.6574,
      "mean_token_accuracy": 0.8013913810253144,
      "num_tokens": 5978454542.0,
      "step": 2875
    },
    {
      "epoch": 2.8020440442876264,
      "grad_norm": 0.1342129409313202,
      "learning_rate": 6.147982390805607e-07,
      "loss": 0.6418,
      "mean_token_accuracy": 0.8058375447988511,
      "num_tokens": 5988864624.0,
      "step": 2880
    },
    {
      "epoch": 2.806910816401022,
      "grad_norm": 0.13734175264835358,
      "learning_rate": 5.852899659555377e-07,
      "loss": 0.6586,
      "mean_token_accuracy": 0.8014825373888016,
      "num_tokens": 5999228523.0,
      "step": 2885
    },
    {
      "epoch": 2.811777588514418,
      "grad_norm": 0.1406588852405548,
      "learning_rate": 5.564989807213372e-07,
      "loss": 0.6562,
      "mean_token_accuracy": 0.8022344693541527,
      "num_tokens": 6009608517.0,
      "step": 2890
    },
    {
      "epoch": 2.8166443606278135,
      "grad_norm": 0.1352994740009308,
      "learning_rate": 5.284261292378967e-07,
      "loss": 0.6392,
      "mean_token_accuracy": 0.8066071510314942,
      "num_tokens": 6019993160.0,
      "step": 2895
    },
    {
      "epoch": 2.8215111327412092,
      "grad_norm": 0.14118821918964386,
      "learning_rate": 5.010722362668558e-07,
      "loss": 0.6536,
      "mean_token_accuracy": 0.8026082828640938,
      "num_tokens": 6030378114.0,
      "step": 2900
    },
    {
      "epoch": 2.8263779048546054,
      "grad_norm": 0.1393430531024933,
      "learning_rate": 4.744381054473446e-07,
      "loss": 0.6509,
      "mean_token_accuracy": 0.8036423295736312,
      "num_tokens": 6040802056.0,
      "step": 2905
    },
    {
      "epoch": 2.831244676968001,
      "grad_norm": 0.13414908945560455,
      "learning_rate": 4.4852451927235304e-07,
      "loss": 0.654,
      "mean_token_accuracy": 0.8024564862251282,
      "num_tokens": 6051194677.0,
      "step": 2910
    },
    {
      "epoch": 2.836111449081397,
      "grad_norm": 0.1357431411743164,
      "learning_rate": 4.23332239065749e-07,
      "loss": 0.6598,
      "mean_token_accuracy": 0.8011080399155617,
      "num_tokens": 6061609663.0,
      "step": 2915
    },
    {
      "epoch": 2.8409782211947925,
      "grad_norm": 0.1353534460067749,
      "learning_rate": 3.9886200495992675e-07,
      "loss": 0.6507,
      "mean_token_accuracy": 0.8035222545266152,
      "num_tokens": 6072027152.0,
      "step": 2920
    },
    {
      "epoch": 2.845844993308188,
      "grad_norm": 0.1338973492383957,
      "learning_rate": 3.751145358740271e-07,
      "loss": 0.6471,
      "mean_token_accuracy": 0.8042368397116662,
      "num_tokens": 6082428941.0,
      "step": 2925
    },
    {
      "epoch": 2.8507117654215843,
      "grad_norm": 0.1394575983285904,
      "learning_rate": 3.520905294928489e-07,
      "loss": 0.6498,
      "mean_token_accuracy": 0.8037312358617783,
      "num_tokens": 6092796335.0,
      "step": 2930
    },
    {
      "epoch": 2.85557853753498,
      "grad_norm": 0.1385304182767868,
      "learning_rate": 3.297906622463404e-07,
      "loss": 0.6532,
      "mean_token_accuracy": 0.8027949735522271,
      "num_tokens": 6103220950.0,
      "step": 2935
    },
    {
      "epoch": 2.8604453096483757,
      "grad_norm": 0.1354062408208847,
      "learning_rate": 3.0821558928972315e-07,
      "loss": 0.6561,
      "mean_token_accuracy": 0.8019639119505882,
      "num_tokens": 6113628580.0,
      "step": 2940
    },
    {
      "epoch": 2.8653120817617714,
      "grad_norm": 0.13501453399658203,
      "learning_rate": 2.8736594448424415e-07,
      "loss": 0.6418,
      "mean_token_accuracy": 0.8057534992694855,
      "num_tokens": 6124019420.0,
      "step": 2945
    },
    {
      "epoch": 2.870178853875167,
      "grad_norm": 0.13876374065876007,
      "learning_rate": 2.6724234037855403e-07,
      "loss": 0.6533,
      "mean_token_accuracy": 0.8029359236359597,
      "num_tokens": 6134422643.0,
      "step": 2950
    },
    {
      "epoch": 2.8750456259885633,
      "grad_norm": 0.1367383748292923,
      "learning_rate": 2.4784536819071616e-07,
      "loss": 0.6401,
      "mean_token_accuracy": 0.8066906943917275,
      "num_tokens": 6144805443.0,
      "step": 2955
    },
    {
      "epoch": 2.879912398101959,
      "grad_norm": 0.13855135440826416,
      "learning_rate": 2.291755977908344e-07,
      "loss": 0.6725,
      "mean_token_accuracy": 0.8019560590386391,
      "num_tokens": 6155204947.0,
      "step": 2960
    },
    {
      "epoch": 2.8847791702153547,
      "grad_norm": 0.13742128014564514,
      "learning_rate": 2.1123357768430275e-07,
      "loss": 0.6477,
      "mean_token_accuracy": 0.8043317839503288,
      "num_tokens": 6165626876.0,
      "step": 2965
    },
    {
      "epoch": 2.8896459423287504,
      "grad_norm": 0.13694816827774048,
      "learning_rate": 1.9401983499569842e-07,
      "loss": 0.6458,
      "mean_token_accuracy": 0.8052307814359665,
      "num_tokens": 6176047864.0,
      "step": 2970
    },
    {
      "epoch": 2.894512714442146,
      "grad_norm": 0.13271553814411163,
      "learning_rate": 1.7753487545330017e-07,
      "loss": 0.6423,
      "mean_token_accuracy": 0.8058422580361366,
      "num_tokens": 6186457248.0,
      "step": 2975
    },
    {
      "epoch": 2.8993794865555422,
      "grad_norm": 0.13335135579109192,
      "learning_rate": 1.6177918337422216e-07,
      "loss": 0.6447,
      "mean_token_accuracy": 0.8048436254262924,
      "num_tokens": 6196841925.0,
      "step": 2980
    },
    {
      "epoch": 2.904246258668938,
      "grad_norm": 0.13320565223693848,
      "learning_rate": 1.4675322165017825e-07,
      "loss": 0.6381,
      "mean_token_accuracy": 0.8069759920239449,
      "num_tokens": 6207243897.0,
      "step": 2985
    },
    {
      "epoch": 2.9091130307823336,
      "grad_norm": 0.1356399655342102,
      "learning_rate": 1.324574317339039e-07,
      "loss": 0.6553,
      "mean_token_accuracy": 0.8022593021392822,
      "num_tokens": 6217647625.0,
      "step": 2990
    },
    {
      "epoch": 2.9139798028957293,
      "grad_norm": 0.1343669593334198,
      "learning_rate": 1.1889223362616663e-07,
      "loss": 0.6444,
      "mean_token_accuracy": 0.805311168730259,
      "num_tokens": 6228070179.0,
      "step": 2995
    },
    {
      "epoch": 2.918846575009125,
      "grad_norm": 0.13633297383785248,
      "learning_rate": 1.0605802586343972e-07,
      "loss": 0.6387,
      "mean_token_accuracy": 0.806846372783184,
      "num_tokens": 6238472187.0,
      "step": 3000
    },
    {
      "epoch": 2.923713347122521,
      "grad_norm": 0.13619497418403625,
      "learning_rate": 9.395518550617821e-08,
      "loss": 0.6478,
      "mean_token_accuracy": 0.8041173994541169,
      "num_tokens": 6248892325.0,
      "step": 3005
    },
    {
      "epoch": 2.928580119235917,
      "grad_norm": 0.13829292356967926,
      "learning_rate": 8.258406812775565e-08,
      "loss": 0.6553,
      "mean_token_accuracy": 0.8021312355995178,
      "num_tokens": 6259303068.0,
      "step": 3010
    },
    {
      "epoch": 2.9334468913493126,
      "grad_norm": 0.1339482069015503,
      "learning_rate": 7.194500780401958e-08,
      "loss": 0.6509,
      "mean_token_accuracy": 0.8036595299839974,
      "num_tokens": 6269699571.0,
      "step": 3015
    },
    {
      "epoch": 2.9383136634627083,
      "grad_norm": 0.13798336684703827,
      "learning_rate": 6.203831710346054e-08,
      "loss": 0.6464,
      "mean_token_accuracy": 0.8045310348272323,
      "num_tokens": 6280107083.0,
      "step": 3020
    },
    {
      "epoch": 2.943180435576104,
      "grad_norm": 0.13729774951934814,
      "learning_rate": 5.2864287078041584e-08,
      "loss": 0.6534,
      "mean_token_accuracy": 0.8032229900360107,
      "num_tokens": 6290526693.0,
      "step": 3025
    },
    {
      "epoch": 2.9480472076895,
      "grad_norm": 0.1346033215522766,
      "learning_rate": 4.442318725464134e-08,
      "loss": 0.6486,
      "mean_token_accuracy": 0.8043657347559929,
      "num_tokens": 6300944044.0,
      "step": 3030
    },
    {
      "epoch": 2.952913979802896,
      "grad_norm": 0.13460230827331543,
      "learning_rate": 3.6715265627146334e-08,
      "loss": 0.6404,
      "mean_token_accuracy": 0.806394700706005,
      "num_tokens": 6311344004.0,
      "step": 3035
    },
    {
      "epoch": 2.9577807519162915,
      "grad_norm": 0.13553784787654877,
      "learning_rate": 2.974074864914578e-08,
      "loss": 0.6519,
      "mean_token_accuracy": 0.8032595947384834,
      "num_tokens": 6321759862.0,
      "step": 3040
    },
    {
      "epoch": 2.9626475240296872,
      "grad_norm": 0.13157348334789276,
      "learning_rate": 2.349984122729798e-08,
      "loss": 0.6545,
      "mean_token_accuracy": 0.8027060940861702,
      "num_tokens": 6332124656.0,
      "step": 3045
    },
    {
      "epoch": 2.967514296143083,
      "grad_norm": 0.13371992111206055,
      "learning_rate": 1.7992726715299058e-08,
      "loss": 0.6529,
      "mean_token_accuracy": 0.8031978636980057,
      "num_tokens": 6342516474.0,
      "step": 3050
    },
    {
      "epoch": 2.972381068256479,
      "grad_norm": 0.1322321593761444,
      "learning_rate": 1.3219566908495573e-08,
      "loss": 0.6575,
      "mean_token_accuracy": 0.8019333899021148,
      "num_tokens": 6352925552.0,
      "step": 3055
    },
    {
      "epoch": 2.9772478403698748,
      "grad_norm": 0.1346210539340973,
      "learning_rate": 9.180502039141114e-09,
      "loss": 0.6448,
      "mean_token_accuracy": 0.8050085991621018,
      "num_tokens": 6363314842.0,
      "step": 3060
    },
    {
      "epoch": 2.9821146124832705,
      "grad_norm": 0.13946297764778137,
      "learning_rate": 5.8756507722634854e-09,
      "loss": 0.6487,
      "mean_token_accuracy": 0.8037395745515823,
      "num_tokens": 6373692258.0,
      "step": 3065
    },
    {
      "epoch": 2.986981384596666,
      "grad_norm": 0.1357472985982895,
      "learning_rate": 3.3051102021924853e-09,
      "loss": 0.6434,
      "mean_token_accuracy": 0.8055857300758362,
      "num_tokens": 6384086474.0,
      "step": 3070
    },
    {
      "epoch": 2.991848156710062,
      "grad_norm": 0.13831458985805511,
      "learning_rate": 1.4689558496927547e-09,
      "loss": 0.6533,
      "mean_token_accuracy": 0.802681352198124,
      "num_tokens": 6394497121.0,
      "step": 3075
    },
    {
      "epoch": 2.996714928823458,
      "grad_norm": 0.1369500756263733,
      "learning_rate": 3.6724165975166034e-10,
      "loss": 0.6448,
      "mean_token_accuracy": 0.8047536134719848,
      "num_tokens": 6404912429.0,
      "step": 3080
    }
  ],
  "logging_steps": 5,
  "max_steps": 3084,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 320,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.7049719711133598e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}