{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 2814,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.19921875,
      "epoch": 0.0010666666666666667,
      "grad_norm": 66.93943786621094,
      "learning_rate": 0.0,
      "loss": 7.131989479064941,
      "mean_token_accuracy": 0.3328644782304764,
      "num_tokens": 16384.0,
      "step": 1
    },
    {
      "entropy": 1.4782986111111112,
      "epoch": 0.010666666666666666,
      "grad_norm": 18.575763702392578,
      "learning_rate": 3.6e-05,
      "loss": 6.127874586317274,
      "mean_token_accuracy": 0.3650997512870365,
      "num_tokens": 163840.0,
      "step": 10
    },
    {
      "entropy": 2.76875,
      "epoch": 0.021333333333333333,
      "grad_norm": 2.84274959564209,
      "learning_rate": 7.6e-05,
      "loss": 3.52994384765625,
      "mean_token_accuracy": 0.46443248987197877,
      "num_tokens": 327680.0,
      "step": 20
    },
    {
      "entropy": 2.596875,
      "epoch": 0.032,
      "grad_norm": 1.3951237201690674,
      "learning_rate": 0.000116,
      "loss": 2.576917266845703,
      "mean_token_accuracy": 0.5652274936437607,
      "num_tokens": 491520.0,
      "step": 30
    },
    {
      "entropy": 1.93046875,
      "epoch": 0.042666666666666665,
      "grad_norm": 1.058626651763916,
      "learning_rate": 0.00015600000000000002,
      "loss": 1.8930469512939454,
      "mean_token_accuracy": 0.6515900194644928,
      "num_tokens": 655360.0,
      "step": 40
    },
    {
      "entropy": 1.360546875,
      "epoch": 0.05333333333333334,
      "grad_norm": 1.3331706523895264,
      "learning_rate": 0.000196,
      "loss": 1.3695893287658691,
      "mean_token_accuracy": 0.7230797469615936,
      "num_tokens": 819200.0,
      "step": 50
    },
    {
      "entropy": 1.085546875,
      "epoch": 0.064,
      "grad_norm": 1.189793348312378,
      "learning_rate": 0.00019999476791363665,
      "loss": 1.095842456817627,
      "mean_token_accuracy": 0.754965752363205,
      "num_tokens": 983040.0,
      "step": 60
    },
    {
      "entropy": 1.0015625,
      "epoch": 0.07466666666666667,
      "grad_norm": 1.3060555458068848,
      "learning_rate": 0.0001999766823920699,
      "loss": 1.000664234161377,
      "mean_token_accuracy": 0.7602739751338958,
      "num_tokens": 1146880.0,
      "step": 70
    },
    {
      "entropy": 0.85703125,
      "epoch": 0.08533333333333333,
      "grad_norm": 1.0107979774475098,
      "learning_rate": 0.0001999456811774999,
      "loss": 0.8682949066162109,
      "mean_token_accuracy": 0.7829990148544311,
      "num_tokens": 1310720.0,
      "step": 80
    },
    {
      "entropy": 0.7951171875,
      "epoch": 0.096,
      "grad_norm": 1.029617428779602,
      "learning_rate": 0.00019990176827488216,
      "loss": 0.8053689002990723,
      "mean_token_accuracy": 0.7931323438882828,
      "num_tokens": 1474560.0,
      "step": 90
    },
    {
      "entropy": 0.738671875,
      "epoch": 0.10666666666666667,
      "grad_norm": 1.226258397102356,
      "learning_rate": 0.00019984494935719523,
      "loss": 0.7419948577880859,
      "mean_token_accuracy": 0.8018162906169891,
      "num_tokens": 1638400.0,
      "step": 100
    },
    {
      "entropy": 0.66796875,
      "epoch": 0.11733333333333333,
      "grad_norm": 1.1927998065948486,
      "learning_rate": 0.00019977523176470765,
      "loss": 0.6762372493743897,
      "mean_token_accuracy": 0.8141634076833725,
      "num_tokens": 1802240.0,
      "step": 110
    },
    {
      "entropy": 0.6681640625,
      "epoch": 0.128,
      "grad_norm": 1.2808526754379272,
      "learning_rate": 0.00019969262450402972,
      "loss": 0.6766439914703369,
      "mean_token_accuracy": 0.8112585633993149,
      "num_tokens": 1966080.0,
      "step": 120
    },
    {
      "entropy": 0.6388671875,
      "epoch": 0.13866666666666666,
      "grad_norm": 1.0940663814544678,
      "learning_rate": 0.0001995971382469501,
      "loss": 0.6461334705352784,
      "mean_token_accuracy": 0.8173679143190384,
      "num_tokens": 2129920.0,
      "step": 130
    },
    {
      "entropy": 0.618359375,
      "epoch": 0.14933333333333335,
      "grad_norm": 1.1703182458877563,
      "learning_rate": 0.00019948878532905694,
      "loss": 0.6226601600646973,
      "mean_token_accuracy": 0.8222725123167038,
      "num_tokens": 2293760.0,
      "step": 140
    },
    {
      "entropy": 0.5904296875,
      "epoch": 0.16,
      "grad_norm": 1.2715009450912476,
      "learning_rate": 0.00019936757974814442,
      "loss": 0.592757272720337,
      "mean_token_accuracy": 0.827409490942955,
      "num_tokens": 2457600.0,
      "step": 150
    },
    {
      "entropy": 0.5919921875,
      "epoch": 0.17066666666666666,
      "grad_norm": 1.1974737644195557,
      "learning_rate": 0.0001992335371624044,
      "loss": 0.6006930351257325,
      "mean_token_accuracy": 0.8252568453550339,
      "num_tokens": 2621440.0,
      "step": 160
    },
    {
      "entropy": 0.585546875,
      "epoch": 0.18133333333333335,
      "grad_norm": 1.219773530960083,
      "learning_rate": 0.00019908667488840356,
      "loss": 0.5985349655151367,
      "mean_token_accuracy": 0.8253180116415024,
      "num_tokens": 2785280.0,
      "step": 170
    },
    {
      "entropy": 0.566015625,
      "epoch": 0.192,
      "grad_norm": 1.3660892248153687,
      "learning_rate": 0.0001989270118988463,
      "loss": 0.5701795101165772,
      "mean_token_accuracy": 0.8307485342025757,
      "num_tokens": 2949120.0,
      "step": 180
    },
    {
      "entropy": 0.5650390625,
      "epoch": 0.20266666666666666,
      "grad_norm": 1.4739092588424683,
      "learning_rate": 0.00019875456882012383,
      "loss": 0.5757900238037109,
      "mean_token_accuracy": 0.8268407583236694,
      "num_tokens": 3112960.0,
      "step": 190
    },
    {
      "entropy": 0.55224609375,
      "epoch": 0.21333333333333335,
      "grad_norm": 1.305107831954956,
      "learning_rate": 0.00019856936792964934,
      "loss": 0.5630440711975098,
      "mean_token_accuracy": 0.8288221627473831,
      "num_tokens": 3276800.0,
      "step": 200
    },
    {
      "entropy": 0.53828125,
      "epoch": 0.224,
      "grad_norm": 0.9271957874298096,
      "learning_rate": 0.0001983714331529801,
      "loss": 0.5515984535217285,
      "mean_token_accuracy": 0.8324302762746811,
      "num_tokens": 3440640.0,
      "step": 210
    },
    {
      "entropy": 0.553515625,
      "epoch": 0.23466666666666666,
      "grad_norm": 1.7457942962646484,
      "learning_rate": 0.0001981607900607267,
      "loss": 0.5645030975341797,
      "mean_token_accuracy": 0.8293847769498826,
      "num_tokens": 3604480.0,
      "step": 220
    },
    {
      "entropy": 0.5451171875,
      "epoch": 0.24533333333333332,
      "grad_norm": 1.1714539527893066,
      "learning_rate": 0.0001979374658652496,
      "loss": 0.559196949005127,
      "mean_token_accuracy": 0.829152399301529,
      "num_tokens": 3768320.0,
      "step": 230
    },
    {
      "entropy": 0.54228515625,
      "epoch": 0.256,
      "grad_norm": 1.3494971990585327,
      "learning_rate": 0.00019770148941714345,
      "loss": 0.5582159519195556,
      "mean_token_accuracy": 0.8287977010011673,
      "num_tokens": 3932160.0,
      "step": 240
    },
    {
      "entropy": 0.523828125,
      "epoch": 0.26666666666666666,
      "grad_norm": 1.4074219465255737,
      "learning_rate": 0.00019745289120151033,
      "loss": 0.5372745990753174,
      "mean_token_accuracy": 0.8322529286146164,
      "num_tokens": 4096000.0,
      "step": 250
    },
    {
      "entropy": 0.52431640625,
      "epoch": 0.2773333333333333,
      "grad_norm": 1.2933133840560913,
      "learning_rate": 0.00019719170333402114,
      "loss": 0.5422052860260009,
      "mean_token_accuracy": 0.8313172668218612,
      "num_tokens": 4259840.0,
      "step": 260
    },
    {
      "entropy": 0.541796875,
      "epoch": 0.288,
      "grad_norm": 1.7657071352005005,
      "learning_rate": 0.00019691795955676687,
      "loss": 0.5548649311065674,
      "mean_token_accuracy": 0.8282839983701706,
      "num_tokens": 4423680.0,
      "step": 270
    },
    {
      "entropy": 0.52177734375,
      "epoch": 0.2986666666666667,
      "grad_norm": 1.9909300804138184,
      "learning_rate": 0.00019663169523389954,
      "loss": 0.5387121677398682,
      "mean_token_accuracy": 0.832240703701973,
      "num_tokens": 4587520.0,
      "step": 280
    },
    {
      "entropy": 0.53408203125,
      "epoch": 0.30933333333333335,
      "grad_norm": 1.5377999544143677,
      "learning_rate": 0.0001963329473470634,
      "loss": 0.5521369457244873,
      "mean_token_accuracy": 0.8280088067054748,
      "num_tokens": 4751360.0,
      "step": 290
    },
    {
      "entropy": 0.54208984375,
      "epoch": 0.32,
      "grad_norm": 1.7751295566558838,
      "learning_rate": 0.00019602175449061766,
      "loss": 0.561640453338623,
      "mean_token_accuracy": 0.8274950981140137,
      "num_tokens": 4915200.0,
      "step": 300
    },
    {
      "entropy": 0.521484375,
      "epoch": 0.33066666666666666,
      "grad_norm": 1.3905848264694214,
      "learning_rate": 0.00019569815686665054,
      "loss": 0.5343613147735595,
      "mean_token_accuracy": 0.8321978926658631,
      "num_tokens": 5079040.0,
      "step": 310
    },
    {
      "entropy": 0.52978515625,
      "epoch": 0.3413333333333333,
      "grad_norm": 1.3625966310501099,
      "learning_rate": 0.00019536219627978553,
      "loss": 0.5453921318054199,
      "mean_token_accuracy": 0.8305222570896149,
      "num_tokens": 5242880.0,
      "step": 320
    },
    {
      "entropy": 0.53154296875,
      "epoch": 0.352,
      "grad_norm": 1.37050199508667,
      "learning_rate": 0.00019501391613178085,
      "loss": 0.5461536407470703,
      "mean_token_accuracy": 0.8296477526426316,
      "num_tokens": 5406720.0,
      "step": 330
    },
    {
      "entropy": 0.51630859375,
      "epoch": 0.3626666666666667,
      "grad_norm": 1.1072008609771729,
      "learning_rate": 0.00019465336141592264,
      "loss": 0.5351454734802246,
      "mean_token_accuracy": 0.8314090102910996,
      "num_tokens": 5570560.0,
      "step": 340
    },
    {
      "entropy": 0.5201171875,
      "epoch": 0.37333333333333335,
      "grad_norm": 1.4493062496185303,
      "learning_rate": 0.0001942805787112122,
      "loss": 0.5382786750793457,
      "mean_token_accuracy": 0.8328033238649368,
      "num_tokens": 5734400.0,
      "step": 350
    },
    {
      "entropy": 0.5103515625,
      "epoch": 0.384,
      "grad_norm": 1.4085346460342407,
      "learning_rate": 0.00019389561617634863,
      "loss": 0.5266407489776611,
      "mean_token_accuracy": 0.8348397731781005,
      "num_tokens": 5898240.0,
      "step": 360
    },
    {
      "entropy": 0.51552734375,
      "epoch": 0.39466666666666667,
      "grad_norm": 1.1408005952835083,
      "learning_rate": 0.00019349852354350752,
      "loss": 0.5307485103607178,
      "mean_token_accuracy": 0.8337695717811584,
      "num_tokens": 6062080.0,
      "step": 370
    },
    {
      "entropy": 0.50908203125,
      "epoch": 0.4053333333333333,
      "grad_norm": 1.6698675155639648,
      "learning_rate": 0.000193089352111916,
      "loss": 0.5267929077148438,
      "mean_token_accuracy": 0.8337451100349427,
      "num_tokens": 6225920.0,
      "step": 380
    },
    {
      "entropy": 0.50888671875,
      "epoch": 0.416,
      "grad_norm": 1.5624690055847168,
      "learning_rate": 0.00019266815474122567,
      "loss": 0.5275313854217529,
      "mean_token_accuracy": 0.8345706999301911,
      "num_tokens": 6389760.0,
      "step": 390
    },
    {
      "entropy": 0.5220703125,
      "epoch": 0.4266666666666667,
      "grad_norm": 0.9785416126251221,
      "learning_rate": 0.00019223498584468372,
      "loss": 0.5390635967254639,
      "mean_token_accuracy": 0.8315252065658569,
      "num_tokens": 6553600.0,
      "step": 400
    },
    {
      "entropy": 0.5146484375,
      "epoch": 0.43733333333333335,
      "grad_norm": 1.1690549850463867,
      "learning_rate": 0.0001917899013821035,
      "loss": 0.5308643341064453,
      "mean_token_accuracy": 0.8331274539232254,
      "num_tokens": 6717440.0,
      "step": 410
    },
    {
      "entropy": 0.51552734375,
      "epoch": 0.448,
      "grad_norm": 1.4404877424240112,
      "learning_rate": 0.00019133295885263518,
      "loss": 0.5314573287963867,
      "mean_token_accuracy": 0.833072406053543,
      "num_tokens": 6881280.0,
      "step": 420
    },
    {
      "entropy": 0.50634765625,
      "epoch": 0.45866666666666667,
      "grad_norm": 1.785623550415039,
      "learning_rate": 0.00019086421728733776,
      "loss": 0.5268424034118653,
      "mean_token_accuracy": 0.8328461319208145,
      "num_tokens": 7045120.0,
      "step": 430
    },
    {
      "entropy": 0.5060546875,
      "epoch": 0.4693333333333333,
      "grad_norm": 1.5035761594772339,
      "learning_rate": 0.00019038373724155282,
      "loss": 0.5267580032348633,
      "mean_token_accuracy": 0.8329317480325699,
      "num_tokens": 7208960.0,
      "step": 440
    },
    {
      "entropy": 0.5150390625,
      "epoch": 0.48,
      "grad_norm": 1.8456891775131226,
      "learning_rate": 0.0001898915807870817,
      "loss": 0.5315303802490234,
      "mean_token_accuracy": 0.8317453533411026,
      "num_tokens": 7372800.0,
      "step": 450
    },
    {
      "entropy": 0.5091796875,
      "epoch": 0.49066666666666664,
      "grad_norm": 1.7735804319381714,
      "learning_rate": 0.00018938781150416653,
      "loss": 0.5288398742675782,
      "mean_token_accuracy": 0.8307301878929139,
      "num_tokens": 7536640.0,
      "step": 460
    },
    {
      "entropy": 0.5044921875,
      "epoch": 0.5013333333333333,
      "grad_norm": 1.2228153944015503,
      "learning_rate": 0.00018887249447327656,
      "loss": 0.5237016201019287,
      "mean_token_accuracy": 0.8311582684516907,
      "num_tokens": 7700480.0,
      "step": 470
    },
    {
      "entropy": 0.5087890625,
      "epoch": 0.512,
      "grad_norm": 1.093058466911316,
      "learning_rate": 0.00018834569626670052,
      "loss": 0.5251482009887696,
      "mean_token_accuracy": 0.8328522473573685,
      "num_tokens": 7864320.0,
      "step": 480
    },
    {
      "entropy": 0.5080078125,
      "epoch": 0.5226666666666666,
      "grad_norm": 1.5129916667938232,
      "learning_rate": 0.00018780748493994647,
      "loss": 0.5237375736236572,
      "mean_token_accuracy": 0.8322835147380829,
      "num_tokens": 8028160.0,
      "step": 490
    },
    {
      "entropy": 0.5044921875,
      "epoch": 0.5333333333333333,
      "grad_norm": 1.8124247789382935,
      "learning_rate": 0.0001872579300229498,
      "loss": 0.5259128093719483,
      "mean_token_accuracy": 0.8323507875204086,
      "num_tokens": 8192000.0,
      "step": 500
    },
    {
      "entropy": 0.49658203125,
      "epoch": 0.544,
      "grad_norm": 1.221025824546814,
      "learning_rate": 0.00018669710251109086,
      "loss": 0.5104663848876954,
      "mean_token_accuracy": 0.8357142895460129,
      "num_tokens": 8355840.0,
      "step": 510
    },
    {
      "entropy": 0.4939453125,
      "epoch": 0.5546666666666666,
      "grad_norm": 1.2056242227554321,
      "learning_rate": 0.0001861250748560234,
      "loss": 0.510359525680542,
      "mean_token_accuracy": 0.8362646758556366,
      "num_tokens": 8519680.0,
      "step": 520
    },
    {
      "entropy": 0.50458984375,
      "epoch": 0.5653333333333334,
      "grad_norm": 1.4202454090118408,
      "learning_rate": 0.0001855419209563146,
      "loss": 0.5239694118499756,
      "mean_token_accuracy": 0.833494371175766,
      "num_tokens": 8683520.0,
      "step": 530
    },
    {
      "entropy": 0.50126953125,
      "epoch": 0.576,
      "grad_norm": 1.414934754371643,
      "learning_rate": 0.00018494771614789842,
      "loss": 0.5178774833679199,
      "mean_token_accuracy": 0.8347663879394531,
      "num_tokens": 8847360.0,
      "step": 540
    },
    {
      "entropy": 0.5041015625,
      "epoch": 0.5866666666666667,
      "grad_norm": 1.9108942747116089,
      "learning_rate": 0.00018434253719434314,
      "loss": 0.5239945411682129,
      "mean_token_accuracy": 0.8318370848894119,
      "num_tokens": 9011200.0,
      "step": 550
    },
    {
      "entropy": 0.49814453125,
      "epoch": 0.5973333333333334,
      "grad_norm": 1.40981924533844,
      "learning_rate": 0.00018372646227693458,
      "loss": 0.517290735244751,
      "mean_token_accuracy": 0.8336961805820465,
      "num_tokens": 9175040.0,
      "step": 560
    },
    {
      "entropy": 0.4943359375,
      "epoch": 0.608,
      "grad_norm": 1.2730116844177246,
      "learning_rate": 0.00018309957098457588,
      "loss": 0.5129383087158204,
      "mean_token_accuracy": 0.835084393620491,
      "num_tokens": 9338880.0,
      "step": 570
    },
    {
      "entropy": 0.51123046875,
      "epoch": 0.6186666666666667,
      "grad_norm": 1.4962691068649292,
      "learning_rate": 0.000182461944303506,
      "loss": 0.5291899681091309,
      "mean_token_accuracy": 0.831928813457489,
      "num_tokens": 9502720.0,
      "step": 580
    },
    {
      "entropy": 0.49619140625,
      "epoch": 0.6293333333333333,
      "grad_norm": 1.6723508834838867,
      "learning_rate": 0.0001818136646068369,
      "loss": 0.513413667678833,
      "mean_token_accuracy": 0.8344361543655395,
      "num_tokens": 9666560.0,
      "step": 590
    },
    {
      "entropy": 0.49931640625,
      "epoch": 0.64,
      "grad_norm": 1.1433684825897217,
      "learning_rate": 0.0001811548156439124,
      "loss": 0.5187834739685059,
      "mean_token_accuracy": 0.8341364979743957,
      "num_tokens": 9830400.0,
      "step": 600
    },
    {
      "entropy": 0.49462890625,
      "epoch": 0.6506666666666666,
      "grad_norm": 1.3142673969268799,
      "learning_rate": 0.0001804854825294886,
      "loss": 0.5118935585021973,
      "mean_token_accuracy": 0.8363135993480683,
      "num_tokens": 9994240.0,
      "step": 610
    },
    {
      "entropy": 0.4966796875,
      "epoch": 0.6613333333333333,
      "grad_norm": 1.1120491027832031,
      "learning_rate": 0.00017980575173273832,
      "loss": 0.5137100219726562,
      "mean_token_accuracy": 0.8352128237485885,
      "num_tokens": 10158080.0,
      "step": 620
    },
    {
      "entropy": 0.507421875,
      "epoch": 0.672,
      "grad_norm": 1.510169506072998,
      "learning_rate": 0.00017911571106608024,
      "loss": 0.5237502574920654,
      "mean_token_accuracy": 0.8310909926891327,
      "num_tokens": 10321920.0,
      "step": 630
    },
    {
      "entropy": 0.5048828125,
      "epoch": 0.6826666666666666,
      "grad_norm": 1.3611756563186646,
      "learning_rate": 0.00017841544967383484,
      "loss": 0.5235665321350098,
      "mean_token_accuracy": 0.8321245074272156,
      "num_tokens": 10485760.0,
      "step": 640
    },
    {
      "entropy": 0.48955078125,
      "epoch": 0.6933333333333334,
      "grad_norm": 1.0056926012039185,
      "learning_rate": 0.00017770505802070785,
      "loss": 0.5044590950012207,
      "mean_token_accuracy": 0.8368945747613907,
      "num_tokens": 10649600.0,
      "step": 650
    },
    {
      "entropy": 0.49072265625,
      "epoch": 0.704,
      "grad_norm": 1.4617265462875366,
      "learning_rate": 0.0001769846278801038,
      "loss": 0.5156108856201171,
      "mean_token_accuracy": 0.8338123828172683,
      "num_tokens": 10813440.0,
      "step": 660
    },
    {
      "entropy": 0.488671875,
      "epoch": 0.7146666666666667,
      "grad_norm": 1.039659023284912,
      "learning_rate": 0.00017625425232226962,
      "loss": 0.5084824562072754,
      "mean_token_accuracy": 0.8362035185098649,
      "num_tokens": 10977280.0,
      "step": 670
    },
    {
      "entropy": 0.49892578125,
      "epoch": 0.7253333333333334,
      "grad_norm": 1.234175443649292,
      "learning_rate": 0.00017551402570227152,
      "loss": 0.5152301788330078,
      "mean_token_accuracy": 0.8343383014202118,
      "num_tokens": 11141120.0,
      "step": 680
    },
    {
      "entropy": 0.49443359375,
      "epoch": 0.736,
      "grad_norm": 1.011959195137024,
      "learning_rate": 0.00017476404364780534,
      "loss": 0.5114828109741211,
      "mean_token_accuracy": 0.8354452073574066,
      "num_tokens": 11304960.0,
      "step": 690
    },
    {
      "entropy": 0.4984375,
      "epoch": 0.7466666666666667,
      "grad_norm": 1.2436904907226562,
      "learning_rate": 0.0001740044030468428,
      "loss": 0.5167118072509765,
      "mean_token_accuracy": 0.8338246047496796,
      "num_tokens": 11468800.0,
      "step": 700
    },
    {
      "entropy": 0.4958984375,
      "epoch": 0.7573333333333333,
      "grad_norm": 1.080377221107483,
      "learning_rate": 0.00017323520203511473,
      "loss": 0.5108581066131592,
      "mean_token_accuracy": 0.8361974090337754,
      "num_tokens": 11632640.0,
      "step": 710
    },
    {
      "entropy": 0.494921875,
      "epoch": 0.768,
      "grad_norm": 1.806570291519165,
      "learning_rate": 0.00017245653998343324,
      "loss": 0.5138883113861084,
      "mean_token_accuracy": 0.8344850718975068,
      "num_tokens": 11796480.0,
      "step": 720
    },
    {
      "entropy": 0.49306640625,
      "epoch": 0.7786666666666666,
      "grad_norm": 1.0864297151565552,
      "learning_rate": 0.00017166851748485443,
      "loss": 0.5095850944519043,
      "mean_token_accuracy": 0.8356408983469009,
      "num_tokens": 11960320.0,
      "step": 730
    },
    {
      "entropy": 0.49755859375,
      "epoch": 0.7893333333333333,
      "grad_norm": 1.3445836305618286,
      "learning_rate": 0.00017087123634168283,
      "loss": 0.5140289306640625,
      "mean_token_accuracy": 0.8345584601163865,
      "num_tokens": 12124160.0,
      "step": 740
    },
    {
      "entropy": 0.49453125,
      "epoch": 0.8,
      "grad_norm": 1.293155550956726,
      "learning_rate": 0.00017006479955232004,
      "loss": 0.5134669780731201,
      "mean_token_accuracy": 0.8346379637718201,
      "num_tokens": 12288000.0,
      "step": 750
    },
    {
      "entropy": 0.48623046875,
      "epoch": 0.8106666666666666,
      "grad_norm": 1.5039390325546265,
      "learning_rate": 0.0001692493112979586,
      "loss": 0.5030923366546631,
      "mean_token_accuracy": 0.8371147215366364,
      "num_tokens": 12451840.0,
      "step": 760
    },
    {
      "entropy": 0.49404296875,
      "epoch": 0.8213333333333334,
      "grad_norm": 1.6980180740356445,
      "learning_rate": 0.00016842487692912312,
      "loss": 0.5080769538879395,
      "mean_token_accuracy": 0.8368639916181564,
      "num_tokens": 12615680.0,
      "step": 770
    },
    {
      "entropy": 0.50126953125,
      "epoch": 0.832,
      "grad_norm": 0.9878014922142029,
      "learning_rate": 0.00016759160295206043,
      "loss": 0.5175663471221924,
      "mean_token_accuracy": 0.8337145239114762,
      "num_tokens": 12779520.0,
      "step": 780
    },
    {
      "entropy": 0.49521484375,
      "epoch": 0.8426666666666667,
      "grad_norm": 1.7671765089035034,
      "learning_rate": 0.00016674959701498028,
      "loss": 0.5180065631866455,
      "mean_token_accuracy": 0.8333904147148132,
      "num_tokens": 12943360.0,
      "step": 790
    },
    {
      "entropy": 0.49931640625,
      "epoch": 0.8533333333333334,
      "grad_norm": 1.4861191511154175,
      "learning_rate": 0.00016589896789414858,
      "loss": 0.5191779136657715,
      "mean_token_accuracy": 0.8338796466588974,
      "num_tokens": 13107200.0,
      "step": 800
    },
    {
      "entropy": 0.47958984375,
      "epoch": 0.864,
      "grad_norm": 1.0632413625717163,
      "learning_rate": 0.00016503982547983498,
      "loss": 0.49591312408447263,
      "mean_token_accuracy": 0.8385518550872803,
      "num_tokens": 13271040.0,
      "step": 810
    },
    {
      "entropy": 0.4859375,
      "epoch": 0.8746666666666667,
      "grad_norm": 1.2972675561904907,
      "learning_rate": 0.0001641722807621163,
      "loss": 0.5045878410339355,
      "mean_token_accuracy": 0.8368395328521728,
      "num_tokens": 13434880.0,
      "step": 820
    },
    {
      "entropy": 0.4853515625,
      "epoch": 0.8853333333333333,
      "grad_norm": 1.279313087463379,
      "learning_rate": 0.00016329644581653842,
      "loss": 0.5012047290802002,
      "mean_token_accuracy": 0.8373409926891326,
      "num_tokens": 13598720.0,
      "step": 830
    },
    {
      "entropy": 0.48681640625,
      "epoch": 0.896,
      "grad_norm": 1.0833185911178589,
      "learning_rate": 0.00016241243378963712,
      "loss": 0.5019305229187012,
      "mean_token_accuracy": 0.8368517637252808,
      "num_tokens": 13762560.0,
      "step": 840
    },
    {
      "entropy": 0.48916015625,
      "epoch": 0.9066666666666666,
      "grad_norm": 1.8804922103881836,
      "learning_rate": 0.00016152035888432133,
      "loss": 0.5081352233886719,
      "mean_token_accuracy": 0.8345645725727081,
      "num_tokens": 13926400.0,
      "step": 850
    },
    {
      "entropy": 0.487109375,
      "epoch": 0.9173333333333333,
      "grad_norm": 1.2958416938781738,
      "learning_rate": 0.0001606203363451194,
      "loss": 0.5035964965820312,
      "mean_token_accuracy": 0.8362585574388504,
      "num_tokens": 14090240.0,
      "step": 860
    },
    {
      "entropy": 0.4892578125,
      "epoch": 0.928,
      "grad_norm": 1.2060233354568481,
      "learning_rate": 0.00015971248244329119,
      "loss": 0.5059124946594238,
      "mean_token_accuracy": 0.8352128177881241,
      "num_tokens": 14254080.0,
      "step": 870
    },
    {
      "entropy": 0.49677734375,
      "epoch": 0.9386666666666666,
      "grad_norm": 1.3199410438537598,
      "learning_rate": 0.00015879691446180706,
      "loss": 0.5118205547332764,
      "mean_token_accuracy": 0.8334148794412612,
      "num_tokens": 14417920.0,
      "step": 880
    },
    {
      "entropy": 0.48359375,
      "epoch": 0.9493333333333334,
      "grad_norm": 0.943783164024353,
      "learning_rate": 0.00015787375068019682,
      "loss": 0.5006434917449951,
      "mean_token_accuracy": 0.8380993157625198,
      "num_tokens": 14581760.0,
      "step": 890
    },
    {
      "entropy": 0.4830078125,
      "epoch": 0.96,
      "grad_norm": 1.180012583732605,
      "learning_rate": 0.00015694311035926908,
      "loss": 0.49756603240966796,
      "mean_token_accuracy": 0.8393468648195267,
      "num_tokens": 14745600.0,
      "step": 900
    },
    {
      "entropy": 0.48125,
      "epoch": 0.9706666666666667,
      "grad_norm": 2.581937789916992,
      "learning_rate": 0.0001560051137257047,
      "loss": 0.5028840065002441,
      "mean_token_accuracy": 0.8358732849359513,
      "num_tokens": 14909440.0,
      "step": 910
    },
    {
      "entropy": 0.4841796875,
      "epoch": 0.9813333333333333,
      "grad_norm": 1.5701502561569214,
      "learning_rate": 0.00015505988195652485,
      "loss": 0.5039067745208741,
      "mean_token_accuracy": 0.836588791012764,
      "num_tokens": 15073280.0,
      "step": 920
    },
    {
      "entropy": 0.4849609375,
      "epoch": 0.992,
      "grad_norm": 0.882295548915863,
      "learning_rate": 0.0001541075371634365,
      "loss": 0.500466775894165,
      "mean_token_accuracy": 0.8372186899185181,
      "num_tokens": 15237120.0,
      "step": 930
    },
    {
      "entropy": 0.49578536184210525,
      "epoch": 1.0021333333333333,
      "grad_norm": 1.2846187353134155,
      "learning_rate": 0.00015314820237705723,
      "loss": 0.5112793922424317,
      "mean_token_accuracy": 0.8342195372832449,
      "num_tokens": 15392768.0,
      "step": 940
    },
    {
      "entropy": 0.4796875,
      "epoch": 1.0128,
      "grad_norm": 1.2446873188018799,
      "learning_rate": 0.00015218200153102135,
      "loss": 0.4987667083740234,
      "mean_token_accuracy": 0.8377874255180359,
      "num_tokens": 15556608.0,
      "step": 950
    },
    {
      "entropy": 0.4880859375,
      "epoch": 1.0234666666666667,
      "grad_norm": 1.306470274925232,
      "learning_rate": 0.00015120905944596906,
      "loss": 0.5060878276824952,
      "mean_token_accuracy": 0.8348703563213349,
      "num_tokens": 15720448.0,
      "step": 960
    },
    {
      "entropy": 0.4810546875,
      "epoch": 1.0341333333333333,
      "grad_norm": 1.5865827798843384,
      "learning_rate": 0.00015022950181342142,
      "loss": 0.5032922744750976,
      "mean_token_accuracy": 0.834001961350441,
      "num_tokens": 15884288.0,
      "step": 970
    },
    {
      "entropy": 0.480859375,
      "epoch": 1.0448,
      "grad_norm": 1.107364535331726,
      "learning_rate": 0.00014924345517954265,
      "loss": 0.503816556930542,
      "mean_token_accuracy": 0.8340569913387299,
      "num_tokens": 16048128.0,
      "step": 980
    },
    {
      "entropy": 0.48271484375,
      "epoch": 1.0554666666666668,
      "grad_norm": 1.6236153841018677,
      "learning_rate": 0.00014825104692879179,
      "loss": 0.5053020000457764,
      "mean_token_accuracy": 0.8355124801397323,
      "num_tokens": 16211968.0,
      "step": 990
    },
    {
      "entropy": 0.49091796875,
      "epoch": 1.0661333333333334,
      "grad_norm": 1.2803524732589722,
      "learning_rate": 0.0001472524052674665,
      "loss": 0.5113526344299316,
      "mean_token_accuracy": 0.8327849775552749,
      "num_tokens": 16375808.0,
      "step": 1000
    },
    {
      "entropy": 0.48671875,
      "epoch": 1.0768,
      "grad_norm": 1.121848464012146,
      "learning_rate": 0.0001462476592071404,
      "loss": 0.5078666687011719,
      "mean_token_accuracy": 0.834668543934822,
      "num_tokens": 16539648.0,
      "step": 1010
    },
    {
      "entropy": 0.4919921875,
      "epoch": 1.0874666666666666,
      "grad_norm": 0.9084911942481995,
      "learning_rate": 0.00014523693854799635,
      "loss": 0.5169054508209229,
      "mean_token_accuracy": 0.8306139886379242,
      "num_tokens": 16703488.0,
      "step": 1020
    },
    {
      "entropy": 0.47607421875,
      "epoch": 1.0981333333333334,
      "grad_norm": 1.3610546588897705,
      "learning_rate": 0.00014422037386205814,
      "loss": 0.49955267906188966,
      "mean_token_accuracy": 0.8367477983236313,
      "num_tokens": 16867328.0,
      "step": 1030
    },
    {
      "entropy": 0.47939453125,
      "epoch": 1.1088,
      "grad_norm": 1.9202488660812378,
      "learning_rate": 0.00014319809647632206,
      "loss": 0.5038884162902832,
      "mean_token_accuracy": 0.8346930027008057,
      "num_tokens": 17031168.0,
      "step": 1040
    },
    {
      "entropy": 0.48017578125,
      "epoch": 1.1194666666666666,
      "grad_norm": 1.6467069387435913,
      "learning_rate": 0.00014217023845579128,
      "loss": 0.503753662109375,
      "mean_token_accuracy": 0.8339469105005264,
      "num_tokens": 17195008.0,
      "step": 1050
    },
    {
      "entropy": 0.4796875,
      "epoch": 1.1301333333333332,
      "grad_norm": 1.5664154291152954,
      "learning_rate": 0.00014113693258641482,
      "loss": 0.5029225826263428,
      "mean_token_accuracy": 0.8353718101978302,
      "num_tokens": 17358848.0,
      "step": 1060
    },
    {
      "entropy": 0.4837890625,
      "epoch": 1.1408,
      "grad_norm": 1.5026112794876099,
      "learning_rate": 0.00014009831235793307,
      "loss": 0.5091967105865478,
      "mean_token_accuracy": 0.8342465668916702,
      "num_tokens": 17522688.0,
      "step": 1070
    },
    {
      "entropy": 0.48330078125,
      "epoch": 1.1514666666666666,
      "grad_norm": 0.952322781085968,
      "learning_rate": 0.00013905451194663298,
      "loss": 0.5070674896240235,
      "mean_token_accuracy": 0.8336044579744339,
      "num_tokens": 17686528.0,
      "step": 1080
    },
    {
      "entropy": 0.4685546875,
      "epoch": 1.1621333333333332,
      "grad_norm": 1.1793233156204224,
      "learning_rate": 0.00013800566619801393,
      "loss": 0.49231753349304197,
      "mean_token_accuracy": 0.8383989781141281,
      "num_tokens": 17850368.0,
      "step": 1090
    },
    {
      "entropy": 0.47802734375,
      "epoch": 1.1728,
      "grad_norm": 1.865923523902893,
      "learning_rate": 0.00013695191060936768,
      "loss": 0.5056027889251709,
      "mean_token_accuracy": 0.8337389916181565,
      "num_tokens": 18014208.0,
      "step": 1100
    },
    {
      "entropy": 0.48125,
      "epoch": 1.1834666666666667,
      "grad_norm": 1.0916545391082764,
      "learning_rate": 0.00013589338131227366,
      "loss": 0.5030971050262452,
      "mean_token_accuracy": 0.8342526853084564,
      "num_tokens": 18178048.0,
      "step": 1110
    },
    {
      "entropy": 0.471484375,
      "epoch": 1.1941333333333333,
      "grad_norm": 1.2349437475204468,
      "learning_rate": 0.00013483021505501268,
      "loss": 0.4925968647003174,
      "mean_token_accuracy": 0.838105434179306,
      "num_tokens": 18341888.0,
      "step": 1120
    },
    {
      "entropy": 0.47158203125,
      "epoch": 1.2048,
      "grad_norm": 1.2658876180648804,
      "learning_rate": 0.00013376254918490073,
      "loss": 0.4975261211395264,
      "mean_token_accuracy": 0.8357999056577683,
      "num_tokens": 18505728.0,
      "step": 1130
    },
    {
      "entropy": 0.480078125,
      "epoch": 1.2154666666666667,
      "grad_norm": 1.882321834564209,
      "learning_rate": 0.00013269052163054554,
      "loss": 0.5023324489593506,
      "mean_token_accuracy": 0.835053813457489,
      "num_tokens": 18669568.0,
      "step": 1140
    },
    {
      "entropy": 0.48173828125,
      "epoch": 1.2261333333333333,
      "grad_norm": 1.280403733253479,
      "learning_rate": 0.00013161427088402797,
      "loss": 0.5054220199584961,
      "mean_token_accuracy": 0.8327544003725051,
      "num_tokens": 18833408.0,
      "step": 1150
    },
    {
      "entropy": 0.47958984375,
      "epoch": 1.2368000000000001,
      "grad_norm": 1.6637307405471802,
      "learning_rate": 0.0001305339359830107,
      "loss": 0.5027641296386719,
      "mean_token_accuracy": 0.8343933433294296,
      "num_tokens": 18997248.0,
      "step": 1160
    },
    {
      "entropy": 0.4892578125,
      "epoch": 1.2474666666666667,
      "grad_norm": 1.661028504371643,
      "learning_rate": 0.00012944965649277617,
      "loss": 0.5107007026672363,
      "mean_token_accuracy": 0.8318676620721817,
      "num_tokens": 19161088.0,
      "step": 1170
    },
    {
      "entropy": 0.48037109375,
      "epoch": 1.2581333333333333,
      "grad_norm": 1.2537710666656494,
      "learning_rate": 0.0001283615724881968,
      "loss": 0.5039165496826172,
      "mean_token_accuracy": 0.8336533725261688,
      "num_tokens": 19324928.0,
      "step": 1180
    },
    {
      "entropy": 0.478515625,
      "epoch": 1.2688,
      "grad_norm": 1.3916021585464478,
      "learning_rate": 0.00012726982453563893,
      "loss": 0.5020300865173339,
      "mean_token_accuracy": 0.8348581165075302,
      "num_tokens": 19488768.0,
      "step": 1190
    },
    {
      "entropy": 0.48427734375,
      "epoch": 1.2794666666666665,
      "grad_norm": 1.2187550067901611,
      "learning_rate": 0.00012617455367480357,
      "loss": 0.5073482036590576,
      "mean_token_accuracy": 0.8328828305006027,
      "num_tokens": 19652608.0,
      "step": 1200
    },
    {
      "entropy": 0.47958984375,
      "epoch": 1.2901333333333334,
      "grad_norm": 1.4132243394851685,
      "learning_rate": 0.00012507590140050588,
      "loss": 0.5043100357055664,
      "mean_token_accuracy": 0.8334943771362304,
      "num_tokens": 19816448.0,
      "step": 1210
    },
    {
      "entropy": 0.48134765625,
      "epoch": 1.3008,
      "grad_norm": 0.8808680772781372,
      "learning_rate": 0.0001239740096443958,
      "loss": 0.5028908252716064,
      "mean_token_accuracy": 0.8350476950407029,
      "num_tokens": 19980288.0,
      "step": 1220
    },
    {
      "entropy": 0.4775390625,
      "epoch": 1.3114666666666666,
      "grad_norm": 1.1168277263641357,
      "learning_rate": 0.00012286902075662247,
      "loss": 0.501319408416748,
      "mean_token_accuracy": 0.8339897215366363,
      "num_tokens": 20144128.0,
      "step": 1230
    },
    {
      "entropy": 0.47177734375,
      "epoch": 1.3221333333333334,
      "grad_norm": 1.7778539657592773,
      "learning_rate": 0.00012176107748744429,
      "loss": 0.497831392288208,
      "mean_token_accuracy": 0.8362524420022964,
      "num_tokens": 20307968.0,
      "step": 1240
    },
    {
      "entropy": 0.47451171875,
      "epoch": 1.3328,
      "grad_norm": 1.7540314197540283,
      "learning_rate": 0.00012065032296878752,
      "loss": 0.4981066703796387,
      "mean_token_accuracy": 0.8352862119674682,
      "num_tokens": 20471808.0,
      "step": 1250
    },
    {
      "entropy": 0.475390625,
      "epoch": 1.3434666666666666,
      "grad_norm": 1.1379261016845703,
      "learning_rate": 0.00011953690069575538,
      "loss": 0.5007460117340088,
      "mean_token_accuracy": 0.8357570946216584,
      "num_tokens": 20635648.0,
      "step": 1260
    },
    {
      "entropy": 0.48349609375,
      "epoch": 1.3541333333333334,
      "grad_norm": 1.474172592163086,
      "learning_rate": 0.0001184209545080905,
      "loss": 0.5084107398986817,
      "mean_token_accuracy": 0.8328216701745987,
      "num_tokens": 20799488.0,
      "step": 1270
    },
    {
      "entropy": 0.47880859375,
      "epoch": 1.3648,
      "grad_norm": 1.821547269821167,
      "learning_rate": 0.0001173026285715926,
      "loss": 0.5033222675323487,
      "mean_token_accuracy": 0.8338123798370362,
      "num_tokens": 20963328.0,
      "step": 1280
    },
    {
      "entropy": 0.484765625,
      "epoch": 1.3754666666666666,
      "grad_norm": 1.3738772869110107,
      "learning_rate": 0.00011618206735949413,
      "loss": 0.5067636013031006,
      "mean_token_accuracy": 0.8332619935274124,
      "num_tokens": 21127168.0,
      "step": 1290
    },
    {
      "entropy": 0.47705078125,
      "epoch": 1.3861333333333334,
      "grad_norm": 1.250972867012024,
      "learning_rate": 0.00011505941563379612,
      "loss": 0.504629373550415,
      "mean_token_accuracy": 0.8350660473108291,
      "num_tokens": 21291008.0,
      "step": 1300
    },
    {
      "entropy": 0.47412109375,
      "epoch": 1.3968,
      "grad_norm": 1.3033738136291504,
      "learning_rate": 0.00011393481842656699,
      "loss": 0.49842138290405275,
      "mean_token_accuracy": 0.8350538223981857,
      "num_tokens": 21454848.0,
      "step": 1310
    },
    {
      "entropy": 0.48876953125,
      "epoch": 1.4074666666666666,
      "grad_norm": 1.451621651649475,
      "learning_rate": 0.00011280842102120621,
      "loss": 0.5103711128234864,
      "mean_token_accuracy": 0.8314579278230667,
      "num_tokens": 21618688.0,
      "step": 1320
    },
    {
      "entropy": 0.48017578125,
      "epoch": 1.4181333333333335,
      "grad_norm": 2.0090088844299316,
      "learning_rate": 0.00011168036893367547,
      "loss": 0.5066225051879882,
      "mean_token_accuracy": 0.8330785214900971,
      "num_tokens": 21782528.0,
      "step": 1330
    },
    {
      "entropy": 0.46640625,
      "epoch": 1.4288,
      "grad_norm": 0.9762557744979858,
      "learning_rate": 0.00011055080789369995,
      "loss": 0.4896556377410889,
      "mean_token_accuracy": 0.8376590073108673,
      "num_tokens": 21946368.0,
      "step": 1340
    },
    {
      "entropy": 0.47177734375,
      "epoch": 1.4394666666666667,
      "grad_norm": 1.1043437719345093,
      "learning_rate": 0.00010941988382594208,
      "loss": 0.4966303825378418,
      "mean_token_accuracy": 0.8361240267753601,
      "num_tokens": 22110208.0,
      "step": 1350
    },
    {
      "entropy": 0.4779296875,
      "epoch": 1.4501333333333333,
      "grad_norm": 1.2414591312408447,
      "learning_rate": 0.0001082877428311497,
      "loss": 0.5036239147186279,
      "mean_token_accuracy": 0.8347052365541459,
      "num_tokens": 22274048.0,
      "step": 1360
    },
    {
      "entropy": 0.47666015625,
      "epoch": 1.4607999999999999,
      "grad_norm": 1.0108938217163086,
      "learning_rate": 0.00010715453116728196,
      "loss": 0.502647590637207,
      "mean_token_accuracy": 0.8352984338998795,
      "num_tokens": 22437888.0,
      "step": 1370
    },
    {
      "entropy": 0.480859375,
      "epoch": 1.4714666666666667,
      "grad_norm": 1.8987362384796143,
      "learning_rate": 0.00010602039523061449,
      "loss": 0.5048245429992676,
      "mean_token_accuracy": 0.8335371851921082,
      "num_tokens": 22601728.0,
      "step": 1380
    },
    {
      "entropy": 0.47705078125,
      "epoch": 1.4821333333333333,
      "grad_norm": 1.1474896669387817,
      "learning_rate": 0.00010488548153682709,
      "loss": 0.5044803142547607,
      "mean_token_accuracy": 0.8330540537834168,
      "num_tokens": 22765568.0,
      "step": 1390
    },
    {
      "entropy": 0.46787109375,
      "epoch": 1.4928,
      "grad_norm": 0.9750197529792786,
      "learning_rate": 0.00010374993670207571,
      "loss": 0.4909402370452881,
      "mean_token_accuracy": 0.8376895815134049,
      "num_tokens": 22929408.0,
      "step": 1400
    },
    {
      "entropy": 0.46650390625,
      "epoch": 1.5034666666666667,
      "grad_norm": 1.0499016046524048,
      "learning_rate": 0.00010261390742405148,
      "loss": 0.488748836517334,
      "mean_token_accuracy": 0.8390839040279389,
      "num_tokens": 23093248.0,
      "step": 1410
    },
    {
      "entropy": 0.4669921875,
      "epoch": 1.5141333333333333,
      "grad_norm": 1.0453500747680664,
      "learning_rate": 0.00010147754046302936,
      "loss": 0.48976974487304686,
      "mean_token_accuracy": 0.8387903690338134,
      "num_tokens": 23257088.0,
      "step": 1420
    },
    {
      "entropy": 0.47421875,
      "epoch": 1.5248,
      "grad_norm": 1.061371088027954,
      "learning_rate": 0.00010034098262290861,
      "loss": 0.4976058483123779,
      "mean_token_accuracy": 0.8357570976018905,
      "num_tokens": 23420928.0,
      "step": 1430
    },
    {
      "entropy": 0.46953125,
      "epoch": 1.5354666666666668,
      "grad_norm": 1.151824712753296,
      "learning_rate": 9.920438073224758e-05,
      "loss": 0.49523177146911623,
      "mean_token_accuracy": 0.8371819972991943,
      "num_tokens": 23584768.0,
      "step": 1440
    },
    {
      "entropy": 0.4828125,
      "epoch": 1.5461333333333334,
      "grad_norm": 0.9126304388046265,
      "learning_rate": 9.806788162529532e-05,
      "loss": 0.5036550521850586,
      "mean_token_accuracy": 0.8335371851921082,
      "num_tokens": 23748608.0,
      "step": 1450
    },
    {
      "entropy": 0.4779296875,
      "epoch": 1.5568,
      "grad_norm": 1.0665663480758667,
      "learning_rate": 9.693163212302264e-05,
      "loss": 0.5045021533966064,
      "mean_token_accuracy": 0.8340141922235489,
      "num_tokens": 23912448.0,
      "step": 1460
    },
    {
      "entropy": 0.4783203125,
      "epoch": 1.5674666666666668,
      "grad_norm": 1.468713402748108,
      "learning_rate": 9.579577901415464e-05,
      "loss": 0.5028573036193847,
      "mean_token_accuracy": 0.8337328761816025,
      "num_tokens": 24076288.0,
      "step": 1470
    },
    {
      "entropy": 0.47255859375,
      "epoch": 1.5781333333333334,
      "grad_norm": 1.177609920501709,
      "learning_rate": 9.466046903620747e-05,
      "loss": 0.4984735012054443,
      "mean_token_accuracy": 0.8356164395809174,
      "num_tokens": 24240128.0,
      "step": 1480
    },
    {
      "entropy": 0.4791015625,
      "epoch": 1.5888,
      "grad_norm": 1.4302760362625122,
      "learning_rate": 9.352584885653186e-05,
      "loss": 0.504204797744751,
      "mean_token_accuracy": 0.8336594879627228,
      "num_tokens": 24403968.0,
      "step": 1490
    },
    {
      "entropy": 0.48037109375,
      "epoch": 1.5994666666666668,
      "grad_norm": 1.147282361984253,
      "learning_rate": 9.239206505336549e-05,
      "loss": 0.5026605129241943,
      "mean_token_accuracy": 0.8338674128055572,
      "num_tokens": 24567808.0,
      "step": 1500
    },
    {
      "entropy": 0.4828125,
      "epoch": 1.6101333333333332,
      "grad_norm": 1.3384449481964111,
      "learning_rate": 9.125926409689706e-05,
      "loss": 0.5098608493804931,
      "mean_token_accuracy": 0.8324486285448074,
      "num_tokens": 24731648.0,
      "step": 1510
    },
    {
      "entropy": 0.479296875,
      "epoch": 1.6208,
      "grad_norm": 1.1918610334396362,
      "learning_rate": 9.01275923303442e-05,
      "loss": 0.5018500328063965,
      "mean_token_accuracy": 0.8343627750873566,
      "num_tokens": 24895488.0,
      "step": 1520
    },
    {
      "entropy": 0.47333984375,
      "epoch": 1.6314666666666666,
      "grad_norm": 1.1267058849334717,
      "learning_rate": 8.899719595104776e-05,
      "loss": 0.49539833068847655,
      "mean_token_accuracy": 0.8360628664493561,
      "num_tokens": 25059328.0,
      "step": 1530
    },
    {
      "entropy": 0.4708984375,
      "epoch": 1.6421333333333332,
      "grad_norm": 1.7237401008605957,
      "learning_rate": 8.786822099158513e-05,
      "loss": 0.4935861110687256,
      "mean_token_accuracy": 0.8363564103841782,
      "num_tokens": 25223168.0,
      "step": 1540
    },
    {
      "entropy": 0.46123046875,
      "epoch": 1.6528,
      "grad_norm": 1.3290094137191772,
      "learning_rate": 8.674081330090468e-05,
      "loss": 0.48729572296142576,
      "mean_token_accuracy": 0.8386741638183594,
      "num_tokens": 25387008.0,
      "step": 1550
    },
    {
      "entropy": 0.471875,
      "epoch": 1.6634666666666666,
      "grad_norm": 1.008961796760559,
      "learning_rate": 8.561511852548389e-05,
      "loss": 0.4935816764831543,
      "mean_token_accuracy": 0.8359099805355072,
      "num_tokens": 25550848.0,
      "step": 1560
    },
    {
      "entropy": 0.4775390625,
      "epoch": 1.6741333333333333,
      "grad_norm": 1.1814944744110107,
      "learning_rate": 8.449128209051374e-05,
      "loss": 0.5017025470733643,
      "mean_token_accuracy": 0.8335249453783036,
      "num_tokens": 25714688.0,
      "step": 1570
    },
    {
      "entropy": 0.46611328125,
      "epoch": 1.6848,
      "grad_norm": 1.0657734870910645,
      "learning_rate": 8.336944918111176e-05,
      "loss": 0.4896141529083252,
      "mean_token_accuracy": 0.8381115436553955,
      "num_tokens": 25878528.0,
      "step": 1580
    },
    {
      "entropy": 0.47578125,
      "epoch": 1.6954666666666667,
      "grad_norm": 1.2163207530975342,
      "learning_rate": 8.224976472356589e-05,
      "loss": 0.49829697608947754,
      "mean_token_accuracy": 0.8355491697788239,
      "num_tokens": 26042368.0,
      "step": 1590
    },
    {
      "entropy": 0.4728515625,
      "epoch": 1.7061333333333333,
      "grad_norm": 1.1504528522491455,
      "learning_rate": 8.113237336661187e-05,
      "loss": 0.4982324123382568,
      "mean_token_accuracy": 0.8357081681489944,
      "num_tokens": 26206208.0,
      "step": 1600
    },
    {
      "entropy": 0.46806640625,
      "epoch": 1.7168,
      "grad_norm": 1.690124273300171,
      "learning_rate": 8.001741946274659e-05,
      "loss": 0.4925436496734619,
      "mean_token_accuracy": 0.8375366896390914,
      "num_tokens": 26370048.0,
      "step": 1610
    },
    {
      "entropy": 0.4697265625,
      "epoch": 1.7274666666666667,
      "grad_norm": 1.8209350109100342,
      "learning_rate": 7.890504704957951e-05,
      "loss": 0.4934333324432373,
      "mean_token_accuracy": 0.8374755412340165,
      "num_tokens": 26533888.0,
      "step": 1620
    },
    {
      "entropy": 0.464453125,
      "epoch": 1.7381333333333333,
      "grad_norm": 1.8102431297302246,
      "learning_rate": 7.779539983122494e-05,
      "loss": 0.48685126304626464,
      "mean_token_accuracy": 0.8384478956460952,
      "num_tokens": 26697728.0,
      "step": 1630
    },
    {
      "entropy": 0.4732421875,
      "epoch": 1.7488000000000001,
      "grad_norm": 1.1146517992019653,
      "learning_rate": 7.668862115973729e-05,
      "loss": 0.49297099113464354,
      "mean_token_accuracy": 0.8367783665657044,
      "num_tokens": 26861568.0,
      "step": 1640
    },
    {
      "entropy": 0.47109375,
      "epoch": 1.7594666666666665,
      "grad_norm": 1.3856486082077026,
      "learning_rate": 7.558485401659183e-05,
      "loss": 0.49691157341003417,
      "mean_token_accuracy": 0.8358182519674301,
      "num_tokens": 27025408.0,
      "step": 1650
    },
    {
      "entropy": 0.48076171875,
      "epoch": 1.7701333333333333,
      "grad_norm": 0.9621296525001526,
      "learning_rate": 7.448424099421341e-05,
      "loss": 0.5016973972320556,
      "mean_token_accuracy": 0.8323079705238342,
      "num_tokens": 27189248.0,
      "step": 1660
    },
    {
      "entropy": 0.4794921875,
      "epoch": 1.7808000000000002,
      "grad_norm": 0.9972234964370728,
      "learning_rate": 7.338692427755527e-05,
      "loss": 0.5010671615600586,
      "mean_token_accuracy": 0.835506358742714,
      "num_tokens": 27353088.0,
      "step": 1670
    },
    {
      "entropy": 0.4755859375,
      "epoch": 1.7914666666666665,
      "grad_norm": 1.3987704515457153,
      "learning_rate": 7.229304562573063e-05,
      "loss": 0.49901595115661623,
      "mean_token_accuracy": 0.8355185896158218,
      "num_tokens": 27516928.0,
      "step": 1680
    },
    {
      "entropy": 0.4751953125,
      "epoch": 1.8021333333333334,
      "grad_norm": 1.2872964143753052,
      "learning_rate": 7.120274635369929e-05,
      "loss": 0.49885988235473633,
      "mean_token_accuracy": 0.8337145268917083,
      "num_tokens": 27680768.0,
      "step": 1690
    },
    {
      "entropy": 0.46708984375,
      "epoch": 1.8128,
      "grad_norm": 1.153632640838623,
      "learning_rate": 7.011616731401146e-05,
      "loss": 0.48921732902526854,
      "mean_token_accuracy": 0.8374633073806763,
      "num_tokens": 27844608.0,
      "step": 1700
    },
    {
      "entropy": 0.4791015625,
      "epoch": 1.8234666666666666,
      "grad_norm": 0.9770835638046265,
      "learning_rate": 6.903344887861154e-05,
      "loss": 0.49947776794433596,
      "mean_token_accuracy": 0.8343321949243545,
      "num_tokens": 28008448.0,
      "step": 1710
    },
    {
      "entropy": 0.46650390625,
      "epoch": 1.8341333333333334,
      "grad_norm": 1.6191648244857788,
      "learning_rate": 6.795473092070384e-05,
      "loss": 0.48804388046264646,
      "mean_token_accuracy": 0.8371819943189621,
      "num_tokens": 28172288.0,
      "step": 1720
    },
    {
      "entropy": 0.47890625,
      "epoch": 1.8448,
      "grad_norm": 1.3073537349700928,
      "learning_rate": 6.688015279668269e-05,
      "loss": 0.5010708808898926,
      "mean_token_accuracy": 0.8348153114318848,
      "num_tokens": 28336128.0,
      "step": 1730
    },
    {
      "entropy": 0.474609375,
      "epoch": 1.8554666666666666,
      "grad_norm": 1.0956038236618042,
      "learning_rate": 6.580985332812959e-05,
      "loss": 0.4915802478790283,
      "mean_token_accuracy": 0.8358610540628433,
      "num_tokens": 28499968.0,
      "step": 1740
    },
    {
      "entropy": 0.47578125,
      "epoch": 1.8661333333333334,
      "grad_norm": 1.0503036975860596,
      "learning_rate": 6.474397078387908e-05,
      "loss": 0.49879956245422363,
      "mean_token_accuracy": 0.8352800875902175,
      "num_tokens": 28663808.0,
      "step": 1750
    },
    {
      "entropy": 0.47998046875,
      "epoch": 1.8768,
      "grad_norm": 1.2566874027252197,
      "learning_rate": 6.36826428621563e-05,
      "loss": 0.5033798694610596,
      "mean_token_accuracy": 0.8327788650989533,
      "num_tokens": 28827648.0,
      "step": 1760
    },
    {
      "entropy": 0.46640625,
      "epoch": 1.8874666666666666,
      "grad_norm": 1.6809600591659546,
      "learning_rate": 6.262600667278809e-05,
      "loss": 0.48870315551757815,
      "mean_token_accuracy": 0.8386435896158219,
      "num_tokens": 28991488.0,
      "step": 1770
    },
    {
      "entropy": 0.4693359375,
      "epoch": 1.8981333333333335,
      "grad_norm": 1.9163674116134644,
      "learning_rate": 6.157419871949023e-05,
      "loss": 0.4921398162841797,
      "mean_token_accuracy": 0.8370963782072067,
      "num_tokens": 29155328.0,
      "step": 1780
    },
    {
      "entropy": 0.47666015625,
      "epoch": 1.9088,
      "grad_norm": 1.2937995195388794,
      "learning_rate": 6.0527354882232946e-05,
      "loss": 0.49634556770324706,
      "mean_token_accuracy": 0.8360934436321259,
      "num_tokens": 29319168.0,
      "step": 1790
    },
    {
      "entropy": 0.47255859375,
      "epoch": 1.9194666666666667,
      "grad_norm": 1.222242832183838,
      "learning_rate": 5.948561039968689e-05,
      "loss": 0.49498662948608396,
      "mean_token_accuracy": 0.8368028432130814,
      "num_tokens": 29483008.0,
      "step": 1800
    },
    {
      "entropy": 0.474609375,
      "epoch": 1.9301333333333335,
      "grad_norm": 1.0627124309539795,
      "learning_rate": 5.8449099851752086e-05,
      "loss": 0.4943253040313721,
      "mean_token_accuracy": 0.8358549416065216,
      "num_tokens": 29646848.0,
      "step": 1810
    },
    {
      "entropy": 0.47802734375,
      "epoch": 1.9407999999999999,
      "grad_norm": 1.2573727369308472,
      "learning_rate": 5.741795714217202e-05,
      "loss": 0.5012723922729492,
      "mean_token_accuracy": 0.8338001430034637,
      "num_tokens": 29810688.0,
      "step": 1820
    },
    {
      "entropy": 0.46943359375,
      "epoch": 1.9514666666666667,
      "grad_norm": 1.429948329925537,
      "learning_rate": 5.639231548123496e-05,
      "loss": 0.49042453765869143,
      "mean_token_accuracy": 0.8363380581140518,
      "num_tokens": 29974528.0,
      "step": 1830
    },
    {
      "entropy": 0.46767578125,
      "epoch": 1.9621333333333333,
      "grad_norm": 1.869755744934082,
      "learning_rate": 5.5372307368564856e-05,
      "loss": 0.4879509925842285,
      "mean_token_accuracy": 0.838087084889412,
      "num_tokens": 30138368.0,
      "step": 1840
    },
    {
      "entropy": 0.47265625,
      "epoch": 1.9727999999999999,
      "grad_norm": 1.6679859161376953,
      "learning_rate": 5.435806457600422e-05,
      "loss": 0.49459309577941896,
      "mean_token_accuracy": 0.8351271986961365,
      "num_tokens": 30302208.0,
      "step": 1850
    },
    {
      "entropy": 0.47958984375,
      "epoch": 1.9834666666666667,
      "grad_norm": 1.1695197820663452,
      "learning_rate": 5.33497181305908e-05,
      "loss": 0.5006305694580078,
      "mean_token_accuracy": 0.8341915398836136,
      "num_tokens": 30466048.0,
      "step": 1860
    },
    {
      "entropy": 0.48408203125,
      "epoch": 1.9941333333333333,
      "grad_norm": 1.1044217348098755,
      "learning_rate": 5.2347398297630655e-05,
      "loss": 0.5035054206848144,
      "mean_token_accuracy": 0.8321673184633255,
      "num_tokens": 30629888.0,
      "step": 1870
    },
    {
      "entropy": 0.4712171052631579,
      "epoch": 2.0042666666666666,
      "grad_norm": 1.168285846710205,
      "learning_rate": 5.135123456386953e-05,
      "loss": 0.4956371784210205,
      "mean_token_accuracy": 0.8343740325225028,
      "num_tokens": 30785536.0,
      "step": 1880
    },
    {
      "entropy": 0.47255859375,
      "epoch": 2.0149333333333335,
      "grad_norm": 1.8216763734817505,
      "learning_rate": 5.036135562076487e-05,
      "loss": 0.4909085750579834,
      "mean_token_accuracy": 0.8365459889173508,
      "num_tokens": 30949376.0,
      "step": 1890
    },
    {
      "entropy": 0.479296875,
      "epoch": 2.0256,
      "grad_norm": 1.2574148178100586,
      "learning_rate": 4.93778893478605e-05,
      "loss": 0.5017770767211914,
      "mean_token_accuracy": 0.8333659470081329,
      "num_tokens": 31113216.0,
      "step": 1900
    },
    {
      "entropy": 0.4658203125,
      "epoch": 2.0362666666666667,
      "grad_norm": 1.2370630502700806,
      "learning_rate": 4.8400962796266315e-05,
      "loss": 0.4847989559173584,
      "mean_token_accuracy": 0.8377874284982681,
      "num_tokens": 31277056.0,
      "step": 1910
    },
    {
      "entropy": 0.4619140625,
      "epoch": 2.0469333333333335,
      "grad_norm": 1.2173320055007935,
      "learning_rate": 4.743070217224485e-05,
      "loss": 0.4812936305999756,
      "mean_token_accuracy": 0.8395242154598236,
      "num_tokens": 31440896.0,
      "step": 1920
    },
    {
      "entropy": 0.47158203125,
      "epoch": 2.0576,
      "grad_norm": 1.1087722778320312,
      "learning_rate": 4.646723282090718e-05,
      "loss": 0.4904578685760498,
      "mean_token_accuracy": 0.835769322514534,
      "num_tokens": 31604736.0,
      "step": 1930
    },
    {
      "entropy": 0.4712890625,
      "epoch": 2.0682666666666667,
      "grad_norm": 1.1680694818496704,
      "learning_rate": 4.551067921001971e-05,
      "loss": 0.4919876575469971,
      "mean_token_accuracy": 0.8373899251222611,
      "num_tokens": 31768576.0,
      "step": 1940
    },
    {
      "entropy": 0.4740234375,
      "epoch": 2.0789333333333335,
      "grad_norm": 1.1964699029922485,
      "learning_rate": 4.456116491392487e-05,
      "loss": 0.49324817657470704,
      "mean_token_accuracy": 0.8353718191385269,
      "num_tokens": 31932416.0,
      "step": 1950
    },
    {
      "entropy": 0.4693359375,
      "epoch": 2.0896,
      "grad_norm": 0.8004054427146912,
      "learning_rate": 4.3618812597576685e-05,
      "loss": 0.4894967555999756,
      "mean_token_accuracy": 0.836307480931282,
      "num_tokens": 32096256.0,
      "step": 1960
    },
    {
      "entropy": 0.4740234375,
      "epoch": 2.1002666666666667,
      "grad_norm": 1.4918841123580933,
      "learning_rate": 4.2683744000694147e-05,
      "loss": 0.49335317611694335,
      "mean_token_accuracy": 0.8369557231664657,
      "num_tokens": 32260096.0,
      "step": 1970
    },
    {
      "entropy": 0.47314453125,
      "epoch": 2.1109333333333336,
      "grad_norm": 1.1185866594314575,
      "learning_rate": 4.175607992203396e-05,
      "loss": 0.49172520637512207,
      "mean_token_accuracy": 0.8364726036787034,
      "num_tokens": 32423936.0,
      "step": 1980
    },
    {
      "entropy": 0.47802734375,
      "epoch": 2.1216,
      "grad_norm": 1.0982848405838013,
      "learning_rate": 4.083594020378498e-05,
      "loss": 0.49456014633178713,
      "mean_token_accuracy": 0.8352066993713378,
      "num_tokens": 32587776.0,
      "step": 1990
    },
    {
      "entropy": 0.47646484375,
      "epoch": 2.1322666666666668,
      "grad_norm": 1.2837402820587158,
      "learning_rate": 3.992344371608608e-05,
      "loss": 0.49712090492248534,
      "mean_token_accuracy": 0.8346196204423905,
      "num_tokens": 32751616.0,
      "step": 2000
    },
    {
      "entropy": 0.46259765625,
      "epoch": 2.142933333333333,
      "grad_norm": 0.9957435131072998,
      "learning_rate": 3.901870834166972e-05,
      "loss": 0.48122057914733884,
      "mean_token_accuracy": 0.8398483335971832,
      "num_tokens": 32915456.0,
      "step": 2010
    },
    {
      "entropy": 0.4755859375,
      "epoch": 2.1536,
      "grad_norm": 0.9987120032310486,
      "learning_rate": 3.8121850960632976e-05,
      "loss": 0.49598493576049807,
      "mean_token_accuracy": 0.8347541570663453,
      "num_tokens": 33079296.0,
      "step": 2020
    },
    {
      "entropy": 0.4705078125,
      "epoch": 2.164266666666667,
      "grad_norm": 1.0633126497268677,
      "learning_rate": 3.723298743533819e-05,
      "loss": 0.48839988708496096,
      "mean_token_accuracy": 0.8377568483352661,
      "num_tokens": 33243136.0,
      "step": 2030
    },
    {
      "entropy": 0.47294921875,
      "epoch": 2.174933333333333,
      "grad_norm": 1.2127145528793335,
      "learning_rate": 3.635223259544507e-05,
      "loss": 0.49268903732299807,
      "mean_token_accuracy": 0.8365398734807968,
      "num_tokens": 33406976.0,
      "step": 2040
    },
    {
      "entropy": 0.46630859375,
      "epoch": 2.1856,
      "grad_norm": 1.1930609941482544,
      "learning_rate": 3.5479700223076086e-05,
      "loss": 0.4841958522796631,
      "mean_token_accuracy": 0.8371269524097442,
      "num_tokens": 33570816.0,
      "step": 2050
    },
    {
      "entropy": 0.4693359375,
      "epoch": 2.196266666666667,
      "grad_norm": 1.1601330041885376,
      "learning_rate": 3.461550303811739e-05,
      "loss": 0.4883225917816162,
      "mean_token_accuracy": 0.8362585604190826,
      "num_tokens": 33734656.0,
      "step": 2060
    },
    {
      "entropy": 0.47255859375,
      "epoch": 2.206933333333333,
      "grad_norm": 1.0405131578445435,
      "learning_rate": 3.375975268365684e-05,
      "loss": 0.4926749229431152,
      "mean_token_accuracy": 0.8348642379045487,
      "num_tokens": 33898496.0,
      "step": 2070
    },
    {
      "entropy": 0.46875,
      "epoch": 2.2176,
      "grad_norm": 1.0318982601165771,
      "learning_rate": 3.291255971156114e-05,
      "loss": 0.4867087364196777,
      "mean_token_accuracy": 0.8380259215831757,
      "num_tokens": 34062336.0,
      "step": 2080
    },
    {
      "entropy": 0.4755859375,
      "epoch": 2.228266666666667,
      "grad_norm": 1.0624244213104248,
      "learning_rate": 3.2074033568194026e-05,
      "loss": 0.49181084632873534,
      "mean_token_accuracy": 0.836545991897583,
      "num_tokens": 34226176.0,
      "step": 2090
    },
    {
      "entropy": 0.48369140625,
      "epoch": 2.238933333333333,
      "grad_norm": 1.1065040826797485,
      "learning_rate": 3.124428258027705e-05,
      "loss": 0.5018001556396484,
      "mean_token_accuracy": 0.8328277885913848,
      "num_tokens": 34390016.0,
      "step": 2100
    },
    {
      "entropy": 0.46767578125,
      "epoch": 2.2496,
      "grad_norm": 1.0899173021316528,
      "learning_rate": 3.0423413940895405e-05,
      "loss": 0.48531055450439453,
      "mean_token_accuracy": 0.837334880232811,
      "num_tokens": 34553856.0,
      "step": 2110
    },
    {
      "entropy": 0.467578125,
      "epoch": 2.2602666666666664,
      "grad_norm": 0.9600430727005005,
      "learning_rate": 2.961153369564982e-05,
      "loss": 0.48721961975097655,
      "mean_token_accuracy": 0.835671478509903,
      "num_tokens": 34717696.0,
      "step": 2120
    },
    {
      "entropy": 0.47216796875,
      "epoch": 2.2709333333333332,
      "grad_norm": 1.0006866455078125,
      "learning_rate": 2.8808746728956836e-05,
      "loss": 0.48862457275390625,
      "mean_token_accuracy": 0.8362707942724228,
      "num_tokens": 34881536.0,
      "step": 2130
    },
    {
      "entropy": 0.46796875,
      "epoch": 2.2816,
      "grad_norm": 0.9775232076644897,
      "learning_rate": 2.8015156750499183e-05,
      "loss": 0.4871671676635742,
      "mean_token_accuracy": 0.8365949153900146,
      "num_tokens": 35045376.0,
      "step": 2140
    },
    {
      "entropy": 0.46806640625,
      "epoch": 2.2922666666666665,
      "grad_norm": 0.9388034343719482,
      "learning_rate": 2.723086628182776e-05,
      "loss": 0.48509922027587893,
      "mean_token_accuracy": 0.8376651138067246,
      "num_tokens": 35209216.0,
      "step": 2150
    },
    {
      "entropy": 0.469140625,
      "epoch": 2.3029333333333333,
      "grad_norm": 0.8595200181007385,
      "learning_rate": 2.6455976643117175e-05,
      "loss": 0.48534379005432127,
      "mean_token_accuracy": 0.8374388456344605,
      "num_tokens": 35373056.0,
      "step": 2160
    },
    {
      "entropy": 0.47470703125,
      "epoch": 2.3136,
      "grad_norm": 1.1569092273712158,
      "learning_rate": 2.5690587940076537e-05,
      "loss": 0.49245386123657225,
      "mean_token_accuracy": 0.8359222143888474,
      "num_tokens": 35536896.0,
      "step": 2170
    },
    {
      "entropy": 0.47392578125,
      "epoch": 2.3242666666666665,
      "grad_norm": 1.12228524684906,
      "learning_rate": 2.4934799051017087e-05,
      "loss": 0.491815185546875,
      "mean_token_accuracy": 0.8360750943422317,
      "num_tokens": 35700736.0,
      "step": 2180
    },
    {
      "entropy": 0.47197265625,
      "epoch": 2.3349333333333333,
      "grad_norm": 1.4032238721847534,
      "learning_rate": 2.4188707614078355e-05,
      "loss": 0.4876112461090088,
      "mean_token_accuracy": 0.8367172181606293,
      "num_tokens": 35864576.0,
      "step": 2190
    },
    {
      "entropy": 0.46875,
      "epoch": 2.3456,
      "grad_norm": 1.238250970840454,
      "learning_rate": 2.345241001461461e-05,
      "loss": 0.4860090255737305,
      "mean_token_accuracy": 0.837163645029068,
      "num_tokens": 36028416.0,
      "step": 2200
    },
    {
      "entropy": 0.4697265625,
      "epoch": 2.3562666666666665,
      "grad_norm": 1.1286371946334839,
      "learning_rate": 2.2726001372743155e-05,
      "loss": 0.48883981704711915,
      "mean_token_accuracy": 0.8358855217695236,
      "num_tokens": 36192256.0,
      "step": 2210
    },
    {
      "entropy": 0.473828125,
      "epoch": 2.3669333333333333,
      "grad_norm": 1.4201050996780396,
      "learning_rate": 2.2009575531055938e-05,
      "loss": 0.48892917633056643,
      "mean_token_accuracy": 0.8367049872875214,
      "num_tokens": 36356096.0,
      "step": 2220
    },
    {
      "entropy": 0.47509765625,
      "epoch": 2.3776,
      "grad_norm": 1.5284488201141357,
      "learning_rate": 2.130322504249641e-05,
      "loss": 0.4901691436767578,
      "mean_token_accuracy": 0.8368272989988327,
      "num_tokens": 36519936.0,
      "step": 2230
    },
    {
      "entropy": 0.47939453125,
      "epoch": 2.3882666666666665,
      "grad_norm": 1.0703142881393433,
      "learning_rate": 2.060704115840284e-05,
      "loss": 0.49634685516357424,
      "mean_token_accuracy": 0.8333353757858276,
      "num_tokens": 36683776.0,
      "step": 2240
    },
    {
      "entropy": 0.47177734375,
      "epoch": 2.3989333333333334,
      "grad_norm": 1.7773879766464233,
      "learning_rate": 1.9921113816719807e-05,
      "loss": 0.4879495620727539,
      "mean_token_accuracy": 0.8364664912223816,
      "num_tokens": 36847616.0,
      "step": 2250
    },
    {
      "entropy": 0.484375,
      "epoch": 2.4096,
      "grad_norm": 1.668304443359375,
      "learning_rate": 1.924553163037943e-05,
      "loss": 0.49918017387390134,
      "mean_token_accuracy": 0.8331763684749603,
      "num_tokens": 37011456.0,
      "step": 2260
    },
    {
      "entropy": 0.4681640625,
      "epoch": 2.4202666666666666,
      "grad_norm": 1.2085943222045898,
      "learning_rate": 1.8580381875853613e-05,
      "loss": 0.4853365421295166,
      "mean_token_accuracy": 0.8376895785331726,
      "num_tokens": 37175296.0,
      "step": 2270
    },
    {
      "entropy": 0.47275390625,
      "epoch": 2.4309333333333334,
      "grad_norm": 0.9767484664916992,
      "learning_rate": 1.7925750481879212e-05,
      "loss": 0.4887202262878418,
      "mean_token_accuracy": 0.8367477983236313,
      "num_tokens": 37339136.0,
      "step": 2280
    },
    {
      "entropy": 0.47861328125,
      "epoch": 2.4416,
      "grad_norm": 1.106047511100769,
      "learning_rate": 1.7281722018357062e-05,
      "loss": 0.49437708854675294,
      "mean_token_accuracy": 0.8345156580209732,
      "num_tokens": 37502976.0,
      "step": 2290
    },
    {
      "entropy": 0.46845703125,
      "epoch": 2.4522666666666666,
      "grad_norm": 1.0685312747955322,
      "learning_rate": 1.664837968542664e-05,
      "loss": 0.48371171951293945,
      "mean_token_accuracy": 0.8376345455646514,
      "num_tokens": 37666816.0,
      "step": 2300
    },
    {
      "entropy": 0.4732421875,
      "epoch": 2.4629333333333334,
      "grad_norm": 0.7760265469551086,
      "learning_rate": 1.6025805302717722e-05,
      "loss": 0.4884315490722656,
      "mean_token_accuracy": 0.8364236801862717,
      "num_tokens": 37830656.0,
      "step": 2310
    },
    {
      "entropy": 0.46748046875,
      "epoch": 2.4736000000000002,
      "grad_norm": 1.0381231307983398,
      "learning_rate": 1.5414079298780282e-05,
      "loss": 0.47982187271118165,
      "mean_token_accuracy": 0.8393774449825286,
      "num_tokens": 37994496.0,
      "step": 2320
    },
    {
      "entropy": 0.47353515625,
      "epoch": 2.4842666666666666,
      "grad_norm": 0.9242449998855591,
      "learning_rate": 1.4813280700694265e-05,
      "loss": 0.4877339839935303,
      "mean_token_accuracy": 0.8362952619791031,
      "num_tokens": 38158336.0,
      "step": 2330
    },
    {
      "entropy": 0.4779296875,
      "epoch": 2.4949333333333334,
      "grad_norm": 0.924400806427002,
      "learning_rate": 1.4223487123860201e-05,
      "loss": 0.49462409019470216,
      "mean_token_accuracy": 0.8345645815134048,
      "num_tokens": 38322176.0,
      "step": 2340
    },
    {
      "entropy": 0.47822265625,
      "epoch": 2.5056000000000003,
      "grad_norm": 0.862738847732544,
      "learning_rate": 1.3644774761972378e-05,
      "loss": 0.49219579696655275,
      "mean_token_accuracy": 0.8350293546915054,
      "num_tokens": 38486016.0,
      "step": 2350
    },
    {
      "entropy": 0.47236328125,
      "epoch": 2.5162666666666667,
      "grad_norm": 1.0803108215332031,
      "learning_rate": 1.3077218377175582e-05,
      "loss": 0.4883579730987549,
      "mean_token_accuracy": 0.836478728055954,
      "num_tokens": 38649856.0,
      "step": 2360
    },
    {
      "entropy": 0.474609375,
      "epoch": 2.5269333333333335,
      "grad_norm": 1.1649197340011597,
      "learning_rate": 1.2520891290406767e-05,
      "loss": 0.48967976570129396,
      "mean_token_accuracy": 0.835971137881279,
      "num_tokens": 38813696.0,
      "step": 2370
    },
    {
      "entropy": 0.46376953125,
      "epoch": 2.5376,
      "grad_norm": 1.4534000158309937,
      "learning_rate": 1.1975865371923001e-05,
      "loss": 0.47943758964538574,
      "mean_token_accuracy": 0.8381971627473831,
      "num_tokens": 38977536.0,
      "step": 2380
    },
    {
      "entropy": 0.4658203125,
      "epoch": 2.5482666666666667,
      "grad_norm": 0.7989732623100281,
      "learning_rate": 1.144221103201668e-05,
      "loss": 0.48253560066223145,
      "mean_token_accuracy": 0.8389554798603058,
      "num_tokens": 39141376.0,
      "step": 2390
    },
    {
      "entropy": 0.4740234375,
      "epoch": 2.558933333333333,
      "grad_norm": 0.8184325098991394,
      "learning_rate": 1.0919997211919497e-05,
      "loss": 0.48844327926635744,
      "mean_token_accuracy": 0.8358121305704117,
      "num_tokens": 39305216.0,
      "step": 2400
    },
    {
      "entropy": 0.47119140625,
      "epoch": 2.5696,
      "grad_norm": 1.3882791996002197,
      "learning_rate": 1.0409291374896102e-05,
      "loss": 0.4853825569152832,
      "mean_token_accuracy": 0.8377446174621582,
      "num_tokens": 39469056.0,
      "step": 2410
    },
    {
      "entropy": 0.467578125,
      "epoch": 2.5802666666666667,
      "grad_norm": 0.9312089681625366,
      "learning_rate": 9.910159497528715e-06,
      "loss": 0.48127002716064454,
      "mean_token_accuracy": 0.8390105187892913,
      "num_tokens": 39632896.0,
      "step": 2420
    },
    {
      "entropy": 0.47314453125,
      "epoch": 2.590933333333333,
      "grad_norm": 1.0183887481689453,
      "learning_rate": 9.422666061193808e-06,
      "loss": 0.4885664939880371,
      "mean_token_accuracy": 0.83658267557621,
      "num_tokens": 39796736.0,
      "step": 2430
    },
    {
      "entropy": 0.464453125,
      "epoch": 2.6016,
      "grad_norm": 1.141377329826355,
      "learning_rate": 8.946874043731945e-06,
      "loss": 0.4782118320465088,
      "mean_token_accuracy": 0.8394569456577301,
      "num_tokens": 39960576.0,
      "step": 2440
    },
    {
      "entropy": 0.476953125,
      "epoch": 2.6122666666666667,
      "grad_norm": 1.2677037715911865,
      "learning_rate": 8.482844911311882e-06,
      "loss": 0.49211993217468264,
      "mean_token_accuracy": 0.8359527885913849,
      "num_tokens": 40124416.0,
      "step": 2450
    },
    {
      "entropy": 0.4744140625,
      "epoch": 2.622933333333333,
      "grad_norm": 0.9096052050590515,
      "learning_rate": 8.030638610489916e-06,
      "loss": 0.4872893810272217,
      "mean_token_accuracy": 0.8361607193946838,
      "num_tokens": 40288256.0,
      "step": 2460
    },
    {
      "entropy": 0.4744140625,
      "epoch": 2.6336,
      "grad_norm": 1.0933566093444824,
      "learning_rate": 7.5903135604655675e-06,
      "loss": 0.49154157638549806,
      "mean_token_accuracy": 0.8348948121070862,
      "num_tokens": 40452096.0,
      "step": 2470
    },
    {
      "entropy": 0.469140625,
      "epoch": 2.6442666666666668,
      "grad_norm": 0.9961123466491699,
      "learning_rate": 7.161926645534567e-06,
      "loss": 0.486204195022583,
      "mean_token_accuracy": 0.8363441795110702,
      "num_tokens": 40615936.0,
      "step": 2480
    },
    {
      "entropy": 0.47548828125,
      "epoch": 2.654933333333333,
      "grad_norm": 0.880841851234436,
      "learning_rate": 6.74553320774014e-06,
      "loss": 0.4878545761108398,
      "mean_token_accuracy": 0.8350966244935989,
      "num_tokens": 40779776.0,
      "step": 2490
    },
    {
      "entropy": 0.4728515625,
      "epoch": 2.6656,
      "grad_norm": 0.7303209900856018,
      "learning_rate": 6.3411870397235724e-06,
      "loss": 0.48897600173950195,
      "mean_token_accuracy": 0.8364603728055954,
      "num_tokens": 40943616.0,
      "step": 2500
    },
    {
      "entropy": 0.4759765625,
      "epoch": 2.676266666666667,
      "grad_norm": 0.7380245923995972,
      "learning_rate": 5.948940377774847e-06,
      "loss": 0.4888805866241455,
      "mean_token_accuracy": 0.837230920791626,
      "num_tokens": 41107456.0,
      "step": 2510
    },
    {
      "entropy": 0.47177734375,
      "epoch": 2.686933333333333,
      "grad_norm": 0.9122079610824585,
      "learning_rate": 5.568843895084408e-06,
      "loss": 0.4892979621887207,
      "mean_token_accuracy": 0.8357204049825668,
      "num_tokens": 41271296.0,
      "step": 2520
    },
    {
      "entropy": 0.47890625,
      "epoch": 2.6976,
      "grad_norm": 1.0076168775558472,
      "learning_rate": 5.2009466951968514e-06,
      "loss": 0.4887696266174316,
      "mean_token_accuracy": 0.8355063557624817,
      "num_tokens": 41435136.0,
      "step": 2530
    },
    {
      "entropy": 0.46748046875,
      "epoch": 2.708266666666667,
      "grad_norm": 0.8389441967010498,
      "learning_rate": 4.845296305667357e-06,
      "loss": 0.483320951461792,
      "mean_token_accuracy": 0.8383255779743195,
      "num_tokens": 41598976.0,
      "step": 2540
    },
    {
      "entropy": 0.4779296875,
      "epoch": 2.718933333333333,
      "grad_norm": 0.8513816595077515,
      "learning_rate": 4.501938671921791e-06,
      "loss": 0.4910125732421875,
      "mean_token_accuracy": 0.8348458915948868,
      "num_tokens": 41762816.0,
      "step": 2550
    },
    {
      "entropy": 0.47373046875,
      "epoch": 2.7296,
      "grad_norm": 0.7910478711128235,
      "learning_rate": 4.170918151321112e-06,
      "loss": 0.48728270530700685,
      "mean_token_accuracy": 0.8364603728055954,
      "num_tokens": 41926656.0,
      "step": 2560
    },
    {
      "entropy": 0.4810546875,
      "epoch": 2.740266666666667,
      "grad_norm": 0.7113332152366638,
      "learning_rate": 3.852277507430957e-06,
      "loss": 0.4973765850067139,
      "mean_token_accuracy": 0.8330051332712174,
      "num_tokens": 42090496.0,
      "step": 2570
    },
    {
      "entropy": 0.4654296875,
      "epoch": 2.7509333333333332,
      "grad_norm": 0.9979612827301025,
      "learning_rate": 3.546057904497213e-06,
      "loss": 0.4806512832641602,
      "mean_token_accuracy": 0.8394141376018525,
      "num_tokens": 42254336.0,
      "step": 2580
    },
    {
      "entropy": 0.47373046875,
      "epoch": 2.7616,
      "grad_norm": 0.7355391979217529,
      "learning_rate": 3.2522989021280703e-06,
      "loss": 0.4865226745605469,
      "mean_token_accuracy": 0.8360995650291443,
      "num_tokens": 42418176.0,
      "step": 2590
    },
    {
      "entropy": 0.47666015625,
      "epoch": 2.772266666666667,
      "grad_norm": 0.9505990147590637,
      "learning_rate": 2.9710384501834165e-06,
      "loss": 0.4901712894439697,
      "mean_token_accuracy": 0.8354329764842987,
      "num_tokens": 42582016.0,
      "step": 2600
    },
    {
      "entropy": 0.46328125,
      "epoch": 2.7829333333333333,
      "grad_norm": 0.7489116787910461,
      "learning_rate": 2.7023128838722823e-06,
      "loss": 0.47481794357299806,
      "mean_token_accuracy": 0.8400623738765717,
      "num_tokens": 42745856.0,
      "step": 2610
    },
    {
      "entropy": 0.4830078125,
      "epoch": 2.7936,
      "grad_norm": 0.8920544385910034,
      "learning_rate": 2.4461569190587397e-06,
      "loss": 0.49599614143371584,
      "mean_token_accuracy": 0.8342588126659394,
      "num_tokens": 42909696.0,
      "step": 2620
    },
    {
      "entropy": 0.4705078125,
      "epoch": 2.804266666666667,
      "grad_norm": 0.8419931530952454,
      "learning_rate": 2.2026036477770796e-06,
      "loss": 0.4835965633392334,
      "mean_token_accuracy": 0.8371881127357483,
      "num_tokens": 43073536.0,
      "step": 2630
    },
    {
      "entropy": 0.470703125,
      "epoch": 2.8149333333333333,
      "grad_norm": 0.8855007290840149,
      "learning_rate": 1.9716845339567638e-06,
      "loss": 0.48393888473510743,
      "mean_token_accuracy": 0.8386863946914673,
      "num_tokens": 43237376.0,
      "step": 2640
    },
    {
      "entropy": 0.47646484375,
      "epoch": 2.8256,
      "grad_norm": 0.9602571129798889,
      "learning_rate": 1.7534294093576765e-06,
      "loss": 0.4905832290649414,
      "mean_token_accuracy": 0.8360934495925904,
      "num_tokens": 43401216.0,
      "step": 2650
    },
    {
      "entropy": 0.46689453125,
      "epoch": 2.836266666666667,
      "grad_norm": 0.9157773852348328,
      "learning_rate": 1.5478664697162304e-06,
      "loss": 0.481394100189209,
      "mean_token_accuracy": 0.838087084889412,
      "num_tokens": 43565056.0,
      "step": 2660
    },
    {
      "entropy": 0.474609375,
      "epoch": 2.8469333333333333,
      "grad_norm": 0.7914666533470154,
      "learning_rate": 1.3550222711028903e-06,
      "loss": 0.4890142440795898,
      "mean_token_accuracy": 0.8363808691501617,
      "num_tokens": 43728896.0,
      "step": 2670
    },
    {
      "entropy": 0.4810546875,
      "epoch": 2.8576,
      "grad_norm": 0.8364726305007935,
      "learning_rate": 1.1749217264914403e-06,
      "loss": 0.4919880867004395,
      "mean_token_accuracy": 0.835255628824234,
      "num_tokens": 43892736.0,
      "step": 2680
    },
    {
      "entropy": 0.47255859375,
      "epoch": 2.8682666666666665,
      "grad_norm": 0.9721719622612,
      "learning_rate": 1.0075881025405243e-06,
      "loss": 0.48676671981811526,
      "mean_token_accuracy": 0.8372309237718583,
      "num_tokens": 44056576.0,
      "step": 2690
    },
    {
      "entropy": 0.47021484375,
      "epoch": 2.8789333333333333,
      "grad_norm": 0.8158939480781555,
      "learning_rate": 8.530430165879732e-07,
      "loss": 0.4834254264831543,
      "mean_token_accuracy": 0.837714034318924,
      "num_tokens": 44220416.0,
      "step": 2700
    },
    {
      "entropy": 0.46416015625,
      "epoch": 2.8895999999999997,
      "grad_norm": 0.7893140912055969,
      "learning_rate": 7.113064338580278e-07,
      "loss": 0.4783947944641113,
      "mean_token_accuracy": 0.8397627234458923,
      "num_tokens": 44384256.0,
      "step": 2710
    },
    {
      "entropy": 0.4751953125,
      "epoch": 2.9002666666666665,
      "grad_norm": 1.0580861568450928,
      "learning_rate": 5.823966648821899e-07,
      "loss": 0.48727917671203613,
      "mean_token_accuracy": 0.8356775939464569,
      "num_tokens": 44548096.0,
      "step": 2720
    },
    {
      "entropy": 0.47333984375,
      "epoch": 2.9109333333333334,
      "grad_norm": 0.7311145663261414,
      "learning_rate": 4.6633036313366017e-07,
      "loss": 0.4856112957000732,
      "mean_token_accuracy": 0.8366438329219819,
      "num_tokens": 44711936.0,
      "step": 2730
    },
    {
      "entropy": 0.46865234375,
      "epoch": 2.9215999999999998,
      "grad_norm": 0.7444504499435425,
      "learning_rate": 3.6312252287593607e-07,
      "loss": 0.4816608905792236,
      "mean_token_accuracy": 0.838044273853302,
      "num_tokens": 44875776.0,
      "step": 2740
    },
    {
      "entropy": 0.4763671875,
      "epoch": 2.9322666666666666,
      "grad_norm": 0.6784510016441345,
      "learning_rate": 2.72786477225806e-07,
      "loss": 0.4874776840209961,
      "mean_token_accuracy": 0.8354880213737488,
      "num_tokens": 45039616.0,
      "step": 2750
    },
    {
      "entropy": 0.4708984375,
      "epoch": 2.9429333333333334,
      "grad_norm": 0.8049107193946838,
      "learning_rate": 1.9533389643080491e-07,
      "loss": 0.4837294578552246,
      "mean_token_accuracy": 0.8365337491035462,
      "num_tokens": 45203456.0,
      "step": 2760
    },
    {
      "entropy": 0.476171875,
      "epoch": 2.9536,
      "grad_norm": 0.720976710319519,
      "learning_rate": 1.3077478636163153e-07,
      "loss": 0.4869108200073242,
      "mean_token_accuracy": 0.8359160989522934,
      "num_tokens": 45367296.0,
      "step": 2770
    },
    {
      "entropy": 0.47490234375,
      "epoch": 2.9642666666666666,
      "grad_norm": 0.7947031855583191,
      "learning_rate": 7.91174872194822e-08,
      "loss": 0.4873377323150635,
      "mean_token_accuracy": 0.8360322952270508,
      "num_tokens": 45531136.0,
      "step": 2780
    },
    {
      "entropy": 0.466015625,
      "epoch": 2.9749333333333334,
      "grad_norm": 0.9816582202911377,
      "learning_rate": 4.036867245861276e-08,
      "loss": 0.47879443168640134,
      "mean_token_accuracy": 0.8388453930616379,
      "num_tokens": 45694976.0,
      "step": 2790
    },
    {
      "entropy": 0.47294921875,
      "epoch": 2.9856,
      "grad_norm": 1.0455915927886963,
      "learning_rate": 1.4533347924217122e-08,
      "loss": 0.48594970703125,
      "mean_token_accuracy": 0.8365826785564423,
      "num_tokens": 45858816.0,
      "step": 2800
    },
    {
      "entropy": 0.47431640625,
      "epoch": 2.9962666666666666,
      "grad_norm": 0.8308309316635132,
      "learning_rate": 1.6148512057667654e-09,
      "loss": 0.4854851722717285,
      "mean_token_accuracy": 0.8368272960186005,
      "num_tokens": 46022656.0,
      "step": 2810
    },
    {
      "entropy": 0.47628348214285715,
      "epoch": 3.0,
      "mean_token_accuracy": 0.8349175197737557,
      "num_tokens": 46080000.0,
      "step": 2814,
      "total_flos": 2.878792925184e+16,
      "train_loss": 0.5595098287820308,
      "train_runtime": 8090.8158,
      "train_samples_per_second": 11.124,
      "train_steps_per_second": 0.348
    }
  ],
  "logging_steps": 10,
  "max_steps": 2814,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.878792925184e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}