{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.952755905511811,
  "eval_steps": 50,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.437505340576172,
      "epoch": 0.01968503937007874,
      "grad_norm": 0.63671875,
      "learning_rate": 3.91304347826087e-05,
      "loss": 1.657536506652832,
      "mean_token_accuracy": 0.6597658038139343,
      "num_tokens": 115439.0,
      "step": 10
    },
    {
      "entropy": 1.5124324679374694,
      "epoch": 0.03937007874015748,
      "grad_norm": 0.4453125,
      "learning_rate": 8.260869565217392e-05,
      "loss": 1.4648273468017579,
      "mean_token_accuracy": 0.683241218328476,
      "num_tokens": 225264.0,
      "step": 20
    },
    {
      "entropy": 1.2266616582870484,
      "epoch": 0.05905511811023622,
      "grad_norm": 0.6796875,
      "learning_rate": 0.00012608695652173915,
      "loss": 1.113858699798584,
      "mean_token_accuracy": 0.7449506461620331,
      "num_tokens": 330923.0,
      "step": 30
    },
    {
      "entropy": 1.05804543197155,
      "epoch": 0.07874015748031496,
      "grad_norm": 0.47265625,
      "learning_rate": 0.00016956521739130436,
      "loss": 1.076363468170166,
      "mean_token_accuracy": 0.7455369174480438,
      "num_tokens": 437622.0,
      "step": 40
    },
    {
      "entropy": 0.9502384155988693,
      "epoch": 0.0984251968503937,
      "grad_norm": 0.5,
      "learning_rate": 0.00019999796688511753,
      "loss": 0.9587571144104003,
      "mean_token_accuracy": 0.7683711260557174,
      "num_tokens": 551175.0,
      "step": 50
    },
    {
      "epoch": 0.0984251968503937,
      "eval_entropy": 1.0894114132280703,
      "eval_loss": 1.0480613708496094,
      "eval_mean_token_accuracy": 0.745736007337217,
      "eval_num_tokens": 551175.0,
      "eval_runtime": 15.9708,
      "eval_samples_per_second": 26.736,
      "eval_steps_per_second": 3.381,
      "step": 50
    },
    {
      "entropy": 0.9499408066272735,
      "epoch": 0.11811023622047244,
      "grad_norm": 0.5859375,
      "learning_rate": 0.0001999618249203163,
      "loss": 0.9569917678833008,
      "mean_token_accuracy": 0.7686831414699554,
      "num_tokens": 658006.0,
      "step": 60
    },
    {
      "entropy": 1.0497896373271942,
      "epoch": 0.1377952755905512,
      "grad_norm": 0.52734375,
      "learning_rate": 0.00019988052141977201,
      "loss": 0.9973426818847656,
      "mean_token_accuracy": 0.7564341962337494,
      "num_tokens": 771278.0,
      "step": 70
    },
    {
      "entropy": 0.8503248453140259,
      "epoch": 0.15748031496062992,
      "grad_norm": 0.5,
      "learning_rate": 0.0001997540931154166,
      "loss": 0.839632225036621,
      "mean_token_accuracy": 0.7869657546281814,
      "num_tokens": 886071.0,
      "step": 80
    },
    {
      "entropy": 0.9232195496559144,
      "epoch": 0.17716535433070865,
      "grad_norm": 0.482421875,
      "learning_rate": 0.00019958259712601903,
      "loss": 0.8961009025573731,
      "mean_token_accuracy": 0.7736886650323868,
      "num_tokens": 996361.0,
      "step": 90
    },
    {
      "entropy": 0.8409020006656647,
      "epoch": 0.1968503937007874,
      "grad_norm": 0.515625,
      "learning_rate": 0.00019936611093137983,
      "loss": 0.839316463470459,
      "mean_token_accuracy": 0.787284591794014,
      "num_tokens": 1103999.0,
      "step": 100
    },
    {
      "epoch": 0.1968503937007874,
      "eval_entropy": 0.988924115896225,
      "eval_loss": 0.9491098523139954,
      "eval_mean_token_accuracy": 0.7609603978969433,
      "eval_num_tokens": 1103999.0,
      "eval_runtime": 15.8986,
      "eval_samples_per_second": 26.858,
      "eval_steps_per_second": 3.397,
      "step": 100
    },
    {
      "entropy": 0.8768177807331086,
      "epoch": 0.21653543307086615,
      "grad_norm": 0.41015625,
      "learning_rate": 0.00019910473233732668,
      "loss": 0.8348051071166992,
      "mean_token_accuracy": 0.7883853554725647,
      "num_tokens": 1216586.0,
      "step": 110
    },
    {
      "entropy": 0.8739962369203568,
      "epoch": 0.23622047244094488,
      "grad_norm": 0.59375,
      "learning_rate": 0.00019879857943152674,
      "loss": 0.8352558135986328,
      "mean_token_accuracy": 0.7789304822683334,
      "num_tokens": 1324767.0,
      "step": 120
    },
    {
      "entropy": 0.8777758985757828,
      "epoch": 0.2559055118110236,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00019844779053013637,
      "loss": 0.8331022262573242,
      "mean_token_accuracy": 0.7784219890832901,
      "num_tokens": 1432356.0,
      "step": 130
    },
    {
      "entropy": 0.7877890735864639,
      "epoch": 0.2755905511811024,
      "grad_norm": 0.55078125,
      "learning_rate": 0.00019805252411531126,
      "loss": 0.7960677146911621,
      "mean_token_accuracy": 0.7926767230033874,
      "num_tokens": 1542545.0,
      "step": 140
    },
    {
      "entropy": 0.8658760845661163,
      "epoch": 0.2952755905511811,
      "grad_norm": 0.5625,
      "learning_rate": 0.0001976129587636064,
      "loss": 0.8095477104187012,
      "mean_token_accuracy": 0.7904855847358704,
      "num_tokens": 1654462.0,
      "step": 150
    },
    {
      "epoch": 0.2952755905511811,
      "eval_entropy": 0.9384135085123556,
      "eval_loss": 0.9150573015213013,
      "eval_mean_token_accuracy": 0.7675120333830515,
      "eval_num_tokens": 1654462.0,
      "eval_runtime": 15.91,
      "eval_samples_per_second": 26.839,
      "eval_steps_per_second": 3.394,
      "step": 150
    },
    {
      "entropy": 0.771521957218647,
      "epoch": 0.31496062992125984,
      "grad_norm": 0.470703125,
      "learning_rate": 0.00019712929306529723,
      "loss": 0.7662525177001953,
      "mean_token_accuracy": 0.8057839542627334,
      "num_tokens": 1760181.0,
      "step": 160
    },
    {
      "entropy": 0.8810683578252793,
      "epoch": 0.3346456692913386,
      "grad_norm": 0.625,
      "learning_rate": 0.000196601745534659,
      "loss": 0.8402512550354004,
      "mean_token_accuracy": 0.7819562494754791,
      "num_tokens": 1864683.0,
      "step": 170
    },
    {
      "entropy": 0.8320274949073792,
      "epoch": 0.3543307086614173,
      "grad_norm": 0.474609375,
      "learning_rate": 0.0001960305545112445,
      "loss": 0.8194713592529297,
      "mean_token_accuracy": 0.7880559414625168,
      "num_tokens": 1978641.0,
      "step": 180
    },
    {
      "entropy": 0.8050856739282608,
      "epoch": 0.37401574803149606,
      "grad_norm": 0.4921875,
      "learning_rate": 0.0001954159780522056,
      "loss": 0.8002090454101562,
      "mean_token_accuracy": 0.791279673576355,
      "num_tokens": 2087989.0,
      "step": 190
    },
    {
      "entropy": 0.8930813580751419,
      "epoch": 0.3937007874015748,
      "grad_norm": 0.6171875,
      "learning_rate": 0.00019475829381570566,
      "loss": 0.8300093650817871,
      "mean_token_accuracy": 0.7832900315523148,
      "num_tokens": 2195494.0,
      "step": 200
    },
    {
      "epoch": 0.3937007874015748,
      "eval_entropy": 0.9506367754053187,
      "eval_loss": 0.8978024125099182,
      "eval_mean_token_accuracy": 0.7694783839914534,
      "eval_num_tokens": 2195494.0,
      "eval_runtime": 15.9025,
      "eval_samples_per_second": 26.851,
      "eval_steps_per_second": 3.396,
      "step": 200
    },
    {
      "entropy": 0.7895667225122451,
      "epoch": 0.41338582677165353,
      "grad_norm": 0.41015625,
      "learning_rate": 0.0001940577989354773,
      "loss": 0.7946149826049804,
      "mean_token_accuracy": 0.7988882720470428,
      "num_tokens": 2314228.0,
      "step": 210
    },
    {
      "entropy": 0.8575397372245789,
      "epoch": 0.4330708661417323,
      "grad_norm": 0.6640625,
      "learning_rate": 0.00019331480988658083,
      "loss": 0.812718391418457,
      "mean_token_accuracy": 0.7836957842111587,
      "num_tokens": 2419122.0,
      "step": 220
    },
    {
      "entropy": 0.8544128388166428,
      "epoch": 0.452755905511811,
      "grad_norm": 0.462890625,
      "learning_rate": 0.00019252966234242468,
      "loss": 0.8198953628540039,
      "mean_token_accuracy": 0.7827243804931641,
      "num_tokens": 2524623.0,
      "step": 230
    },
    {
      "entropy": 0.9001998215913772,
      "epoch": 0.47244094488188976,
      "grad_norm": 0.59375,
      "learning_rate": 0.00019170271102311238,
      "loss": 0.9015755653381348,
      "mean_token_accuracy": 0.7709000498056412,
      "num_tokens": 2631031.0,
      "step": 240
    },
    {
      "entropy": 0.7926925078034401,
      "epoch": 0.4921259842519685,
      "grad_norm": 0.49609375,
      "learning_rate": 0.00019083432953518418,
      "loss": 0.8107254028320312,
      "mean_token_accuracy": 0.7990583658218384,
      "num_tokens": 2741031.0,
      "step": 250
    },
    {
      "epoch": 0.4921259842519685,
      "eval_entropy": 0.9023421671655443,
      "eval_loss": 0.884621262550354,
      "eval_mean_token_accuracy": 0.7719593346118927,
      "eval_num_tokens": 2741031.0,
      "eval_runtime": 15.8929,
      "eval_samples_per_second": 26.867,
      "eval_steps_per_second": 3.398,
      "step": 250
    },
    {
      "entropy": 0.7738933145999909,
      "epoch": 0.5118110236220472,
      "grad_norm": 0.58203125,
      "learning_rate": 0.00018992491020282618,
      "loss": 0.7572973251342774,
      "mean_token_accuracy": 0.7949966192245483,
      "num_tokens": 2853547.0,
      "step": 260
    },
    {
      "entropy": 0.8569424450397491,
      "epoch": 0.531496062992126,
      "grad_norm": 1.375,
      "learning_rate": 0.0001889748638906231,
      "loss": 0.8083616256713867,
      "mean_token_accuracy": 0.791676989197731,
      "num_tokens": 2968367.0,
      "step": 270
    },
    {
      "entropy": 0.889449554681778,
      "epoch": 0.5511811023622047,
      "grad_norm": 0.44921875,
      "learning_rate": 0.00018798461981793462,
      "loss": 0.8472497940063477,
      "mean_token_accuracy": 0.7779456943273544,
      "num_tokens": 3075963.0,
      "step": 280
    },
    {
      "entropy": 0.8320091620087624,
      "epoch": 0.5708661417322834,
      "grad_norm": 0.453125,
      "learning_rate": 0.00018695462536497934,
      "loss": 0.8240900993347168,
      "mean_token_accuracy": 0.7846117675304413,
      "num_tokens": 3185899.0,
      "step": 290
    },
    {
      "entropy": 0.8333972141146659,
      "epoch": 0.5905511811023622,
      "grad_norm": 0.486328125,
      "learning_rate": 0.00018588534587071385,
      "loss": 0.8117186546325683,
      "mean_token_accuracy": 0.7908233135938645,
      "num_tokens": 3297319.0,
      "step": 300
    },
    {
      "epoch": 0.5905511811023622,
      "eval_entropy": 0.8452967874429844,
      "eval_loss": 0.8709425926208496,
      "eval_mean_token_accuracy": 0.7752477537702631,
      "eval_num_tokens": 3297319.0,
      "eval_runtime": 15.8962,
      "eval_samples_per_second": 26.862,
      "eval_steps_per_second": 3.397,
      "step": 300
    },
    {
      "entropy": 0.8324314504861832,
      "epoch": 0.610236220472441,
      "grad_norm": 0.39453125,
      "learning_rate": 0.00018477726442259806,
      "loss": 0.801210880279541,
      "mean_token_accuracy": 0.7885541886091232,
      "num_tokens": 3415477.0,
      "step": 310
    },
    {
      "entropy": 0.8624329552054405,
      "epoch": 0.6299212598425197,
      "grad_norm": 0.400390625,
      "learning_rate": 0.00018363088163834252,
      "loss": 0.8126995086669921,
      "mean_token_accuracy": 0.785911500453949,
      "num_tokens": 3524759.0,
      "step": 320
    },
    {
      "entropy": 0.8507865130901336,
      "epoch": 0.6496062992125984,
      "grad_norm": 0.458984375,
      "learning_rate": 0.00018244671543973513,
      "loss": 0.8206624031066895,
      "mean_token_accuracy": 0.7833890795707703,
      "num_tokens": 3640838.0,
      "step": 330
    },
    {
      "entropy": 0.7785286724567413,
      "epoch": 0.6692913385826772,
      "grad_norm": 0.51953125,
      "learning_rate": 0.0001812253008186505,
      "loss": 0.753394889831543,
      "mean_token_accuracy": 0.8014622688293457,
      "num_tokens": 3756735.0,
      "step": 340
    },
    {
      "entropy": 0.8389557063579559,
      "epoch": 0.6889763779527559,
      "grad_norm": 0.50390625,
      "learning_rate": 0.0001799671895953471,
      "loss": 0.8116916656494141,
      "mean_token_accuracy": 0.7909021496772766,
      "num_tokens": 3870377.0,
      "step": 350
    },
    {
      "epoch": 0.6889763779527559,
      "eval_entropy": 0.8793082656683745,
      "eval_loss": 0.8570566177368164,
      "eval_mean_token_accuracy": 0.7760916120476193,
      "eval_num_tokens": 3870377.0,
      "eval_runtime": 15.9217,
      "eval_samples_per_second": 26.819,
      "eval_steps_per_second": 3.392,
      "step": 350
    },
    {
      "entropy": 0.7998805820941925,
      "epoch": 0.7086614173228346,
      "grad_norm": 0.494140625,
      "learning_rate": 0.00017867295016916167,
      "loss": 0.7578703880310058,
      "mean_token_accuracy": 0.7953435629606247,
      "num_tokens": 3975692.0,
      "step": 360
    },
    {
      "entropy": 0.9054980486631393,
      "epoch": 0.7283464566929134,
      "grad_norm": 0.4765625,
      "learning_rate": 0.0001773431672617132,
      "loss": 0.8652525901794433,
      "mean_token_accuracy": 0.7738145262002945,
      "num_tokens": 4087256.0,
      "step": 370
    },
    {
      "entropy": 0.7607795268297195,
      "epoch": 0.7480314960629921,
      "grad_norm": 0.498046875,
      "learning_rate": 0.00017597844165273286,
      "loss": 0.7365350723266602,
      "mean_token_accuracy": 0.7966376423835755,
      "num_tokens": 4192629.0,
      "step": 380
    },
    {
      "entropy": 0.8066679537296295,
      "epoch": 0.7677165354330708,
      "grad_norm": 0.443359375,
      "learning_rate": 0.00017457938990863908,
      "loss": 0.7610409736633301,
      "mean_token_accuracy": 0.8000116467475891,
      "num_tokens": 4303838.0,
      "step": 390
    },
    {
      "entropy": 0.8927708208560944,
      "epoch": 0.7874015748031497,
      "grad_norm": 0.494140625,
      "learning_rate": 0.00017314664410398033,
      "loss": 0.8361794471740722,
      "mean_token_accuracy": 0.7784365385770797,
      "num_tokens": 4414002.0,
      "step": 400
    },
    {
      "epoch": 0.7874015748031497,
      "eval_entropy": 0.8607326823252218,
      "eval_loss": 0.8459087610244751,
      "eval_mean_token_accuracy": 0.7791097947844753,
      "eval_num_tokens": 4414002.0,
      "eval_runtime": 16.176,
      "eval_samples_per_second": 26.397,
      "eval_steps_per_second": 3.338,
      "step": 400
    },
    {
      "entropy": 0.7851713225245476,
      "epoch": 0.8070866141732284,
      "grad_norm": 0.56640625,
      "learning_rate": 0.0001716808515358714,
      "loss": 0.8022769927978516,
      "mean_token_accuracy": 0.7959282606840133,
      "num_tokens": 4525758.0,
      "step": 410
    },
    {
      "entropy": 0.7688719958066941,
      "epoch": 0.8267716535433071,
      "grad_norm": 0.447265625,
      "learning_rate": 0.00017018267443155268,
      "loss": 0.7182153224945068,
      "mean_token_accuracy": 0.8076630532741547,
      "num_tokens": 4640742.0,
      "step": 420
    },
    {
      "entropy": 0.8512372523546219,
      "epoch": 0.8464566929133859,
      "grad_norm": 0.455078125,
      "learning_rate": 0.00016865278964920378,
      "loss": 0.8275189399719238,
      "mean_token_accuracy": 0.7865870356559753,
      "num_tokens": 4754244.0,
      "step": 430
    },
    {
      "entropy": 0.7933165490627289,
      "epoch": 0.8661417322834646,
      "grad_norm": 0.34765625,
      "learning_rate": 0.00016709188837214732,
      "loss": 0.7769423484802246,
      "mean_token_accuracy": 0.7905291795730591,
      "num_tokens": 4862576.0,
      "step": 440
    },
    {
      "entropy": 0.7952004507184028,
      "epoch": 0.8858267716535433,
      "grad_norm": 0.375,
      "learning_rate": 0.0001655006757965806,
      "loss": 0.755205774307251,
      "mean_token_accuracy": 0.7940250873565674,
      "num_tokens": 4984410.0,
      "step": 450
    },
    {
      "epoch": 0.8858267716535433,
      "eval_entropy": 0.8390876281040686,
      "eval_loss": 0.8377042412757874,
      "eval_mean_token_accuracy": 0.7800426229282662,
      "eval_num_tokens": 4984410.0,
      "eval_runtime": 15.9314,
      "eval_samples_per_second": 26.802,
      "eval_steps_per_second": 3.39,
      "step": 450
    },
    {
      "entropy": 0.8776025518774986,
      "epoch": 0.905511811023622,
      "grad_norm": 0.5703125,
      "learning_rate": 0.0001638798708129768,
      "loss": 0.8327735900878906,
      "mean_token_accuracy": 0.7833347082138061,
      "num_tokens": 5094029.0,
      "step": 460
    },
    {
      "entropy": 0.8570419639348984,
      "epoch": 0.9251968503937008,
      "grad_norm": 0.38671875,
      "learning_rate": 0.00016223020568129854,
      "loss": 0.830419921875,
      "mean_token_accuracy": 0.7789777547121048,
      "num_tokens": 5210613.0,
      "step": 470
    },
    {
      "entropy": 0.8276598662137985,
      "epoch": 0.9448818897637795,
      "grad_norm": 0.59375,
      "learning_rate": 0.00016055242570017244,
      "loss": 0.787315034866333,
      "mean_token_accuracy": 0.7914238631725311,
      "num_tokens": 5322546.0,
      "step": 480
    },
    {
      "entropy": 0.8109192818403244,
      "epoch": 0.9645669291338582,
      "grad_norm": 0.4375,
      "learning_rate": 0.00015884728887017162,
      "loss": 0.7857238292694092,
      "mean_token_accuracy": 0.7958390027284622,
      "num_tokens": 5434008.0,
      "step": 490
    },
    {
      "entropy": 0.6643441513180732,
      "epoch": 0.984251968503937,
      "grad_norm": 0.419921875,
      "learning_rate": 0.0001571155655513606,
      "loss": 0.6719898700714111,
      "mean_token_accuracy": 0.8172835201025009,
      "num_tokens": 5544086.0,
      "step": 500
    },
    {
      "epoch": 0.984251968503937,
      "eval_entropy": 0.8121603627999624,
      "eval_loss": 0.8310334086418152,
      "eval_mean_token_accuracy": 0.7817157308260599,
      "eval_num_tokens": 5544086.0,
      "eval_runtime": 15.9111,
      "eval_samples_per_second": 26.837,
      "eval_steps_per_second": 3.394,
      "step": 500
    },
    {
      "entropy": 0.7330836996436119,
      "epoch": 1.0039370078740157,
      "grad_norm": 0.38671875,
      "learning_rate": 0.0001553580381152559,
      "loss": 0.6914227962493896,
      "mean_token_accuracy": 0.8179270803928376,
      "num_tokens": 5657246.0,
      "step": 510
    },
    {
      "entropy": 0.7551421344280242,
      "epoch": 1.0236220472440944,
      "grad_norm": 0.486328125,
      "learning_rate": 0.0001535755005913602,
      "loss": 0.6854944705963135,
      "mean_token_accuracy": 0.8148082345724106,
      "num_tokens": 5768184.0,
      "step": 520
    },
    {
      "entropy": 0.7464401632547378,
      "epoch": 1.0433070866141732,
      "grad_norm": 0.6015625,
      "learning_rate": 0.00015176875830842938,
      "loss": 0.7031197071075439,
      "mean_token_accuracy": 0.814039769768715,
      "num_tokens": 5885598.0,
      "step": 530
    },
    {
      "entropy": 0.6804196655750274,
      "epoch": 1.0629921259842519,
      "grad_norm": 0.435546875,
      "learning_rate": 0.0001499386275306348,
      "loss": 0.6404387950897217,
      "mean_token_accuracy": 0.8235772669315338,
      "num_tokens": 6002120.0,
      "step": 540
    },
    {
      "entropy": 0.6835484370589257,
      "epoch": 1.0826771653543308,
      "grad_norm": 0.416015625,
      "learning_rate": 0.0001480859350887852,
      "loss": 0.6457780361175537,
      "mean_token_accuracy": 0.8242624312639236,
      "num_tokens": 6112085.0,
      "step": 550
    },
    {
      "epoch": 1.0826771653543308,
      "eval_entropy": 0.7967574977212482,
      "eval_loss": 0.835203230381012,
      "eval_mean_token_accuracy": 0.7809422369356509,
      "eval_num_tokens": 6112085.0,
      "eval_runtime": 15.9214,
      "eval_samples_per_second": 26.819,
      "eval_steps_per_second": 3.392,
      "step": 550
    },
    {
      "entropy": 0.7008466422557831,
      "epoch": 1.1023622047244095,
      "grad_norm": 0.48046875,
      "learning_rate": 0.00014621151800677434,
      "loss": 0.6664916515350342,
      "mean_token_accuracy": 0.8154505103826523,
      "num_tokens": 6224645.0,
      "step": 560
    },
    {
      "entropy": 0.6077279314398766,
      "epoch": 1.1220472440944882,
      "grad_norm": 0.37890625,
      "learning_rate": 0.0001443162231234242,
      "loss": 0.5706528663635254,
      "mean_token_accuracy": 0.8392371147871017,
      "num_tokens": 6336790.0,
      "step": 570
    },
    {
      "entropy": 0.683767618238926,
      "epoch": 1.141732283464567,
      "grad_norm": 0.55078125,
      "learning_rate": 0.0001424009067098931,
      "loss": 0.6489217758178711,
      "mean_token_accuracy": 0.8194159388542175,
      "num_tokens": 6450761.0,
      "step": 580
    },
    {
      "entropy": 0.6302935466170311,
      "epoch": 1.1614173228346456,
      "grad_norm": 0.486328125,
      "learning_rate": 0.00014046643408282305,
      "loss": 0.5900890350341796,
      "mean_token_accuracy": 0.8340148836374283,
      "num_tokens": 6562060.0,
      "step": 590
    },
    {
      "entropy": 0.6183161422610283,
      "epoch": 1.1811023622047245,
      "grad_norm": 0.455078125,
      "learning_rate": 0.00013851367921340017,
      "loss": 0.5828392982482911,
      "mean_token_accuracy": 0.835411223769188,
      "num_tokens": 6676922.0,
      "step": 600
    },
    {
      "epoch": 1.1811023622047245,
      "eval_entropy": 0.7808757076660792,
      "eval_loss": 0.835366427898407,
      "eval_mean_token_accuracy": 0.7813083474282865,
      "eval_num_tokens": 6676922.0,
      "eval_runtime": 15.8899,
      "eval_samples_per_second": 26.872,
      "eval_steps_per_second": 3.398,
      "step": 600
    },
    {
      "entropy": 0.6484389528632164,
      "epoch": 1.2007874015748032,
      "grad_norm": 0.4921875,
      "learning_rate": 0.0001365435243325051,
      "loss": 0.6237234592437744,
      "mean_token_accuracy": 0.8273959398269654,
      "num_tokens": 6783497.0,
      "step": 610
    },
    {
      "entropy": 0.6731782004237175,
      "epoch": 1.220472440944882,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00013455685953213194,
      "loss": 0.6462783336639404,
      "mean_token_accuracy": 0.8242785036563873,
      "num_tokens": 6894014.0,
      "step": 620
    },
    {
      "entropy": 0.6746662899851799,
      "epoch": 1.2401574803149606,
      "grad_norm": 0.484375,
      "learning_rate": 0.0001325545823632554,
      "loss": 0.6090981006622315,
      "mean_token_accuracy": 0.8258850336074829,
      "num_tokens": 6997350.0,
      "step": 630
    },
    {
      "entropy": 0.6707195937633514,
      "epoch": 1.2598425196850394,
      "grad_norm": 0.451171875,
      "learning_rate": 0.00013053759743032823,
      "loss": 0.6500792026519775,
      "mean_token_accuracy": 0.8198623776435852,
      "num_tokens": 7109067.0,
      "step": 640
    },
    {
      "entropy": 0.7742280542850495,
      "epoch": 1.279527559055118,
      "grad_norm": 0.55859375,
      "learning_rate": 0.00012850681598259195,
      "loss": 0.717244005203247,
      "mean_token_accuracy": 0.7994646310806275,
      "num_tokens": 7218739.0,
      "step": 650
    },
    {
      "epoch": 1.279527559055118,
      "eval_entropy": 0.7477047746932065,
      "eval_loss": 0.8298369646072388,
      "eval_mean_token_accuracy": 0.7822379061469326,
      "eval_num_tokens": 7218739.0,
      "eval_runtime": 15.8838,
      "eval_samples_per_second": 26.883,
      "eval_steps_per_second": 3.4,
      "step": 650
    },
    {
      "entropy": 0.7066994801163673,
      "epoch": 1.2992125984251968,
      "grad_norm": 0.52734375,
      "learning_rate": 0.0001264631555023856,
      "loss": 0.6610036373138428,
      "mean_token_accuracy": 0.8158259451389313,
      "num_tokens": 7334561.0,
      "step": 660
    },
    {
      "entropy": 0.6165548652410507,
      "epoch": 1.3188976377952755,
      "grad_norm": 0.76953125,
      "learning_rate": 0.00012440753929063826,
      "loss": 0.6015829563140869,
      "mean_token_accuracy": 0.8280171245336533,
      "num_tokens": 7443871.0,
      "step": 670
    },
    {
      "entropy": 0.7171120464801788,
      "epoch": 1.3385826771653544,
      "grad_norm": 0.51171875,
      "learning_rate": 0.00012234089604973303,
      "loss": 0.6697704792022705,
      "mean_token_accuracy": 0.816604706645012,
      "num_tokens": 7555559.0,
      "step": 680
    },
    {
      "entropy": 0.6375228509306907,
      "epoch": 1.358267716535433,
      "grad_norm": 0.51171875,
      "learning_rate": 0.0001202641594639305,
      "loss": 0.6138493061065674,
      "mean_token_accuracy": 0.8299814134836196,
      "num_tokens": 7669861.0,
      "step": 690
    },
    {
      "entropy": 0.7527845561504364,
      "epoch": 1.3779527559055118,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00011817826777754171,
      "loss": 0.6848691940307617,
      "mean_token_accuracy": 0.8056917071342469,
      "num_tokens": 7782831.0,
      "step": 700
    },
    {
      "epoch": 1.3779527559055118,
      "eval_entropy": 0.7785312417480681,
      "eval_loss": 0.822371244430542,
      "eval_mean_token_accuracy": 0.7844836800186722,
      "eval_num_tokens": 7782831.0,
      "eval_runtime": 15.9232,
      "eval_samples_per_second": 26.816,
      "eval_steps_per_second": 3.391,
      "step": 700
    },
    {
      "entropy": 0.7150849819183349,
      "epoch": 1.3976377952755905,
      "grad_norm": 0.419921875,
      "learning_rate": 0.00011608416337104065,
      "loss": 0.6717889785766602,
      "mean_token_accuracy": 0.8194838553667069,
      "num_tokens": 7895511.0,
      "step": 710
    },
    {
      "entropy": 0.6417959302663803,
      "epoch": 1.4173228346456692,
      "grad_norm": 0.44140625,
      "learning_rate": 0.00011398279233530829,
      "loss": 0.6370071411132813,
      "mean_token_accuracy": 0.8262952893972397,
      "num_tokens": 8006186.0,
      "step": 720
    },
    {
      "entropy": 0.7074607491493226,
      "epoch": 1.4370078740157481,
      "grad_norm": 0.515625,
      "learning_rate": 0.00011187510404420028,
      "loss": 0.6439510822296143,
      "mean_token_accuracy": 0.8158842891454696,
      "num_tokens": 8110444.0,
      "step": 730
    },
    {
      "entropy": 0.7658031359314919,
      "epoch": 1.4566929133858268,
      "grad_norm": 0.7109375,
      "learning_rate": 0.00010976205072563132,
      "loss": 0.7316959857940674,
      "mean_token_accuracy": 0.8061666846275329,
      "num_tokens": 8220128.0,
      "step": 740
    },
    {
      "entropy": 0.6396550685167313,
      "epoch": 1.4763779527559056,
      "grad_norm": 0.5078125,
      "learning_rate": 0.00010764458703137024,
      "loss": 0.6169118881225586,
      "mean_token_accuracy": 0.8309739202260971,
      "num_tokens": 8327961.0,
      "step": 750
    },
    {
      "epoch": 1.4763779527559056,
      "eval_entropy": 0.7342568177867819,
      "eval_loss": 0.8272306323051453,
      "eval_mean_token_accuracy": 0.7827356656392416,
      "eval_num_tokens": 8327961.0,
      "eval_runtime": 15.911,
      "eval_samples_per_second": 26.837,
      "eval_steps_per_second": 3.394,
      "step": 750
    },
    {
      "entropy": 0.6999662488698959,
      "epoch": 1.4960629921259843,
      "grad_norm": 0.671875,
      "learning_rate": 0.00010552366960573985,
      "loss": 0.6717596054077148,
      "mean_token_accuracy": 0.8167526543140411,
      "num_tokens": 8432725.0,
      "step": 760
    },
    {
      "entropy": 0.6736484602093696,
      "epoch": 1.515748031496063,
      "grad_norm": 0.408203125,
      "learning_rate": 0.00010340025665341684,
      "loss": 0.6454797744750976,
      "mean_token_accuracy": 0.8241631388664246,
      "num_tokens": 8550042.0,
      "step": 770
    },
    {
      "entropy": 0.6505535006523132,
      "epoch": 1.5354330708661417,
      "grad_norm": 0.45703125,
      "learning_rate": 0.00010127530750652642,
      "loss": 0.6001406669616699,
      "mean_token_accuracy": 0.8359915494918824,
      "num_tokens": 8664499.0,
      "step": 780
    },
    {
      "entropy": 0.643044576048851,
      "epoch": 1.5551181102362204,
      "grad_norm": 0.6875,
      "learning_rate": 9.914978219122791e-05,
      "loss": 0.6050069808959961,
      "mean_token_accuracy": 0.8284690707921982,
      "num_tokens": 8778459.0,
      "step": 790
    },
    {
      "entropy": 0.6357371553778648,
      "epoch": 1.574803149606299,
      "grad_norm": 0.6484375,
      "learning_rate": 9.702464099398632e-05,
      "loss": 0.6026933193206787,
      "mean_token_accuracy": 0.8313924312591553,
      "num_tokens": 8886361.0,
      "step": 800
    },
    {
      "epoch": 1.574803149606299,
      "eval_entropy": 0.7611764465217237,
      "eval_loss": 0.8170427083969116,
      "eval_mean_token_accuracy": 0.7845815287695991,
      "eval_num_tokens": 8886361.0,
      "eval_runtime": 15.9048,
      "eval_samples_per_second": 26.847,
      "eval_steps_per_second": 3.395,
      "step": 800
    },
    {
      "entropy": 0.6775295034050941,
      "epoch": 1.594488188976378,
      "grad_norm": 0.515625,
      "learning_rate": 9.4900844027727e-05,
      "loss": 0.6273195743560791,
      "mean_token_accuracy": 0.8267990559339523,
      "num_tokens": 8995558.0,
      "step": 810
    },
    {
      "entropy": 0.7035147309303283,
      "epoch": 1.6141732283464567,
      "grad_norm": 0.53515625,
      "learning_rate": 9.277935079806795e-05,
      "loss": 0.6532655715942383,
      "mean_token_accuracy": 0.8209419757127762,
      "num_tokens": 9096245.0,
      "step": 820
    },
    {
      "entropy": 0.6083378851413727,
      "epoch": 1.6338582677165354,
      "grad_norm": 0.796875,
      "learning_rate": 9.066111976982712e-05,
      "loss": 0.5965664386749268,
      "mean_token_accuracy": 0.8344648003578186,
      "num_tokens": 9204264.0,
      "step": 830
    },
    {
      "entropy": 0.6942630469799042,
      "epoch": 1.6535433070866141,
      "grad_norm": 0.50390625,
      "learning_rate": 8.854710793399947e-05,
      "loss": 0.6506452560424805,
      "mean_token_accuracy": 0.8177429378032685,
      "num_tokens": 9314940.0,
      "step": 840
    },
    {
      "entropy": 0.6778671249747277,
      "epoch": 1.673228346456693,
      "grad_norm": 0.546875,
      "learning_rate": 8.64382703754003e-05,
      "loss": 0.6361007213592529,
      "mean_token_accuracy": 0.8213603168725967,
      "num_tokens": 9429496.0,
      "step": 850
    },
    {
      "epoch": 1.673228346456693,
      "eval_entropy": 0.7621777179064574,
      "eval_loss": 0.8149666786193848,
      "eval_mean_token_accuracy": 0.7850644290447235,
      "eval_num_tokens": 9429496.0,
      "eval_runtime": 15.9323,
      "eval_samples_per_second": 26.801,
      "eval_steps_per_second": 3.389,
      "step": 850
    },
    {
      "entropy": 0.6798793867230415,
      "epoch": 1.6929133858267718,
      "grad_norm": 0.46875,
      "learning_rate": 8.433555984116937e-05,
      "loss": 0.613386869430542,
      "mean_token_accuracy": 0.8275126874446869,
      "num_tokens": 9544270.0,
      "step": 860
    },
    {
      "entropy": 0.6023602589964867,
      "epoch": 1.7125984251968505,
      "grad_norm": 0.53125,
      "learning_rate": 8.22399263103314e-05,
      "loss": 0.6046566963195801,
      "mean_token_accuracy": 0.832113066315651,
      "num_tokens": 9654198.0,
      "step": 870
    },
    {
      "entropy": 0.640959857404232,
      "epoch": 1.7322834645669292,
      "grad_norm": 0.6875,
      "learning_rate": 8.015231656460712e-05,
      "loss": 0.6335668563842773,
      "mean_token_accuracy": 0.8338181763887406,
      "num_tokens": 9769457.0,
      "step": 880
    },
    {
      "entropy": 0.6831659495830535,
      "epoch": 1.7519685039370079,
      "grad_norm": 0.5390625,
      "learning_rate": 7.807367376066889e-05,
      "loss": 0.6274240970611572,
      "mean_token_accuracy": 0.8270208179950714,
      "num_tokens": 9875831.0,
      "step": 890
    },
    {
      "entropy": 0.691785454750061,
      "epoch": 1.7716535433070866,
      "grad_norm": 0.48828125,
      "learning_rate": 7.60049370040338e-05,
      "loss": 0.631044578552246,
      "mean_token_accuracy": 0.8213751494884491,
      "num_tokens": 9991758.0,
      "step": 900
    },
    {
      "epoch": 1.7716535433070866,
      "eval_entropy": 0.741917089731605,
      "eval_loss": 0.8126867413520813,
      "eval_mean_token_accuracy": 0.7860870074342798,
      "eval_num_tokens": 9991758.0,
      "eval_runtime": 15.9328,
      "eval_samples_per_second": 26.8,
      "eval_steps_per_second": 3.389,
      "step": 900
    },
    {
      "entropy": 0.7206486001610756,
      "epoch": 1.7913385826771653,
      "grad_norm": 0.51171875,
      "learning_rate": 7.394704092478742e-05,
      "loss": 0.685962438583374,
      "mean_token_accuracy": 0.8114332675933837,
      "num_tokens": 10106255.0,
      "step": 910
    },
    {
      "entropy": 0.6408105552196502,
      "epoch": 1.811023622047244,
      "grad_norm": 0.51953125,
      "learning_rate": 7.190091525532921e-05,
      "loss": 0.5981018543243408,
      "mean_token_accuracy": 0.8346310049295426,
      "num_tokens": 10218066.0,
      "step": 920
    },
    {
      "entropy": 0.6974071592092514,
      "epoch": 1.8307086614173227,
      "grad_norm": 0.62109375,
      "learning_rate": 6.98674844103308e-05,
      "loss": 0.6660090446472168,
      "mean_token_accuracy": 0.8166637867689133,
      "num_tokens": 10326991.0,
      "step": 930
    },
    {
      "entropy": 0.7043072924017906,
      "epoch": 1.8503937007874016,
      "grad_norm": 0.515625,
      "learning_rate": 6.784766706909662e-05,
      "loss": 0.6577100276947021,
      "mean_token_accuracy": 0.8193566143512726,
      "num_tokens": 10432558.0,
      "step": 940
    },
    {
      "entropy": 0.6612386792898178,
      "epoch": 1.8700787401574803,
      "grad_norm": 0.66015625,
      "learning_rate": 6.58423757605158e-05,
      "loss": 0.663103723526001,
      "mean_token_accuracy": 0.8252974808216095,
      "num_tokens": 10541702.0,
      "step": 950
    },
    {
      "epoch": 1.8700787401574803,
      "eval_entropy": 0.7594050356635341,
      "eval_loss": 0.809001624584198,
      "eval_mean_token_accuracy": 0.7858329779571958,
      "eval_num_tokens": 10541702.0,
      "eval_runtime": 15.9266,
      "eval_samples_per_second": 26.811,
      "eval_steps_per_second": 3.391,
      "step": 950
    },
    {
      "entropy": 0.7143705666065217,
      "epoch": 1.889763779527559,
      "grad_norm": 0.5234375,
      "learning_rate": 6.38525164507927e-05,
      "loss": 0.6485679149627686,
      "mean_token_accuracy": 0.8175264537334442,
      "num_tokens": 10652431.0,
      "step": 960
    },
    {
      "entropy": 0.681491743028164,
      "epoch": 1.909448818897638,
      "grad_norm": 0.5625,
      "learning_rate": 6.187898813414251e-05,
      "loss": 0.6528646945953369,
      "mean_token_accuracy": 0.82598916888237,
      "num_tokens": 10766036.0,
      "step": 970
    },
    {
      "entropy": 0.6429278805851937,
      "epoch": 1.9291338582677167,
      "grad_norm": 0.6484375,
      "learning_rate": 5.99226824266365e-05,
      "loss": 0.6115736961364746,
      "mean_token_accuracy": 0.8285941183567047,
      "num_tokens": 10869598.0,
      "step": 980
    },
    {
      "entropy": 0.6953536570072174,
      "epoch": 1.9488188976377954,
      "grad_norm": 0.640625,
      "learning_rate": 5.7984483163380764e-05,
      "loss": 0.6476809501647949,
      "mean_token_accuracy": 0.8186965942382812,
      "num_tokens": 10983085.0,
      "step": 990
    },
    {
      "entropy": 0.645970419049263,
      "epoch": 1.968503937007874,
      "grad_norm": 0.578125,
      "learning_rate": 5.606526599921026e-05,
      "loss": 0.6282810211181641,
      "mean_token_accuracy": 0.8330216020345688,
      "num_tokens": 11090966.0,
      "step": 1000
    },
    {
      "epoch": 1.968503937007874,
      "eval_entropy": 0.7374509738551246,
      "eval_loss": 0.80621737241745,
      "eval_mean_token_accuracy": 0.7872726199803529,
      "eval_num_tokens": 11090966.0,
      "eval_runtime": 15.9085,
      "eval_samples_per_second": 26.841,
      "eval_steps_per_second": 3.394,
      "step": 1000
    },
    {
      "entropy": 0.7159051194787025,
      "epoch": 1.9881889763779528,
      "grad_norm": 0.51953125,
      "learning_rate": 5.416589801307874e-05,
      "loss": 0.6746394634246826,
      "mean_token_accuracy": 0.816139143705368,
      "num_tokens": 11204584.0,
      "step": 1010
    },
    {
      "entropy": 0.6366253510117531,
      "epoch": 2.0078740157480315,
      "grad_norm": 0.380859375,
      "learning_rate": 5.2287237316322676e-05,
      "loss": 0.5599459171295166,
      "mean_token_accuracy": 0.8438727766275406,
      "num_tokens": 11314232.0,
      "step": 1020
    },
    {
      "entropy": 0.591021591424942,
      "epoch": 2.02755905511811,
      "grad_norm": 0.5234375,
      "learning_rate": 5.043013266497751e-05,
      "loss": 0.5247670173645019,
      "mean_token_accuracy": 0.8531373143196106,
      "num_tokens": 11422130.0,
      "step": 1030
    },
    {
      "entropy": 0.593072983622551,
      "epoch": 2.047244094488189,
      "grad_norm": 0.4609375,
      "learning_rate": 4.85954230763197e-05,
      "loss": 0.5415357112884521,
      "mean_token_accuracy": 0.8525791645050049,
      "num_tokens": 11538400.0,
      "step": 1040
    },
    {
      "entropy": 0.5583488404750824,
      "epoch": 2.0669291338582676,
      "grad_norm": 0.46875,
      "learning_rate": 4.6783937449808856e-05,
      "loss": 0.4957115173339844,
      "mean_token_accuracy": 0.8567668944597244,
      "num_tokens": 11645829.0,
      "step": 1050
    },
    {
      "epoch": 2.0669291338582676,
      "eval_entropy": 0.6570017006662157,
      "eval_loss": 0.8396118879318237,
      "eval_mean_token_accuracy": 0.7848087637512772,
      "eval_num_tokens": 11645829.0,
      "eval_runtime": 15.9273,
      "eval_samples_per_second": 26.809,
      "eval_steps_per_second": 3.39,
      "step": 1050
    },
    {
      "entropy": 0.5133202388882637,
      "epoch": 2.0866141732283463,
      "grad_norm": 0.5546875,
      "learning_rate": 4.499649419260133e-05,
      "loss": 0.46275715827941893,
      "mean_token_accuracy": 0.8645182639360428,
      "num_tokens": 11760076.0,
      "step": 1060
    },
    {
      "entropy": 0.5825602993369102,
      "epoch": 2.106299212598425,
      "grad_norm": 0.388671875,
      "learning_rate": 4.323390084980353e-05,
      "loss": 0.5249548435211182,
      "mean_token_accuracy": 0.8451484978199005,
      "num_tokens": 11875367.0,
      "step": 1070
    },
    {
      "entropy": 0.5683134436607361,
      "epoch": 2.1259842519685037,
      "grad_norm": 0.58203125,
      "learning_rate": 4.1496953739633124e-05,
      "loss": 0.49717249870300295,
      "mean_token_accuracy": 0.857735213637352,
      "num_tokens": 11983766.0,
      "step": 1080
    },
    {
      "entropy": 0.5722413748502732,
      "epoch": 2.145669291338583,
      "grad_norm": 0.61328125,
      "learning_rate": 3.978643759365202e-05,
      "loss": 0.5132757186889648,
      "mean_token_accuracy": 0.8525079637765884,
      "num_tokens": 12092559.0,
      "step": 1090
    },
    {
      "entropy": 0.5411732234060764,
      "epoch": 2.1653543307086616,
      "grad_norm": 0.52734375,
      "learning_rate": 3.810312520223419e-05,
      "loss": 0.5036151885986329,
      "mean_token_accuracy": 0.8593157321214676,
      "num_tokens": 12207307.0,
      "step": 1100
    },
    {
      "epoch": 2.1653543307086616,
      "eval_entropy": 0.6568675924230505,
      "eval_loss": 0.8406071662902832,
      "eval_mean_token_accuracy": 0.7840766156161273,
      "eval_num_tokens": 12207307.0,
      "eval_runtime": 15.9308,
      "eval_samples_per_second": 26.803,
      "eval_steps_per_second": 3.39,
      "step": 1100
    },
    {
      "entropy": 0.5992338545620441,
      "epoch": 2.1850393700787403,
      "grad_norm": 0.47265625,
      "learning_rate": 3.644777706542845e-05,
      "loss": 0.5292766571044922,
      "mean_token_accuracy": 0.8512318909168244,
      "num_tokens": 12319969.0,
      "step": 1110
    },
    {
      "entropy": 0.514266274124384,
      "epoch": 2.204724409448819,
      "grad_norm": 0.5,
      "learning_rate": 3.4821141049373804e-05,
      "loss": 0.4458413600921631,
      "mean_token_accuracy": 0.8724471956491471,
      "num_tokens": 12431027.0,
      "step": 1120
    },
    {
      "entropy": 0.5649555295705795,
      "epoch": 2.2244094488188977,
      "grad_norm": 0.52734375,
      "learning_rate": 3.3223952048422544e-05,
      "loss": 0.5298549652099609,
      "mean_token_accuracy": 0.8531913965940475,
      "num_tokens": 12546037.0,
      "step": 1130
    },
    {
      "entropy": 0.5983025059103966,
      "epoch": 2.2440944881889764,
      "grad_norm": 0.58203125,
      "learning_rate": 3.165693165312395e-05,
      "loss": 0.5455004215240479,
      "mean_token_accuracy": 0.8468506157398223,
      "num_tokens": 12656297.0,
      "step": 1140
    },
    {
      "entropy": 0.5571755915880203,
      "epoch": 2.263779527559055,
      "grad_norm": 0.55859375,
      "learning_rate": 3.0120787824218554e-05,
      "loss": 0.5133816719055175,
      "mean_token_accuracy": 0.8542670160531998,
      "num_tokens": 12763743.0,
      "step": 1150
    },
    {
      "epoch": 2.263779527559055,
      "eval_entropy": 0.647307668571119,
      "eval_loss": 0.844680666923523,
      "eval_mean_token_accuracy": 0.7836638799420109,
      "eval_num_tokens": 12763743.0,
      "eval_runtime": 15.9342,
      "eval_samples_per_second": 26.798,
      "eval_steps_per_second": 3.389,
      "step": 1150
    },
    {
      "entropy": 0.6052810281515122,
      "epoch": 2.283464566929134,
      "grad_norm": 0.482421875,
      "learning_rate": 2.8616214572789968e-05,
      "loss": 0.5399055004119873,
      "mean_token_accuracy": 0.8467662870883942,
      "num_tokens": 12874080.0,
      "step": 1160
    },
    {
      "entropy": 0.495437890291214,
      "epoch": 2.3031496062992125,
      "grad_norm": 0.57421875,
      "learning_rate": 2.7143891646719265e-05,
      "loss": 0.4656820297241211,
      "mean_token_accuracy": 0.8652756780385971,
      "num_tokens": 12986029.0,
      "step": 1170
    },
    {
      "entropy": 0.5415738135576248,
      "epoch": 2.322834645669291,
      "grad_norm": 0.59765625,
      "learning_rate": 2.570448422358297e-05,
      "loss": 0.5029004573822021,
      "mean_token_accuracy": 0.8594674915075302,
      "num_tokens": 13092818.0,
      "step": 1180
    },
    {
      "entropy": 0.5835595086216927,
      "epoch": 2.34251968503937,
      "grad_norm": 0.6484375,
      "learning_rate": 2.4298642610134105e-05,
      "loss": 0.502112865447998,
      "mean_token_accuracy": 0.8509185254573822,
      "num_tokens": 13206493.0,
      "step": 1190
    },
    {
      "entropy": 0.6709924265742302,
      "epoch": 2.362204724409449,
      "grad_norm": 0.6640625,
      "learning_rate": 2.292700194850147e-05,
      "loss": 0.5888309001922607,
      "mean_token_accuracy": 0.829518511891365,
      "num_tokens": 13314862.0,
      "step": 1200
    },
    {
      "epoch": 2.362204724409449,
      "eval_entropy": 0.6516638414727317,
      "eval_loss": 0.8389837741851807,
      "eval_mean_token_accuracy": 0.7852442959944407,
      "eval_num_tokens": 13314862.0,
      "eval_runtime": 15.9159,
      "eval_samples_per_second": 26.829,
      "eval_steps_per_second": 3.393,
      "step": 1200
    },
    {
      "entropy": 0.5645383015275002,
      "epoch": 2.3818897637795278,
      "grad_norm": 0.5390625,
      "learning_rate": 2.159018192924004e-05,
      "loss": 0.5259816169738769,
      "mean_token_accuracy": 0.8538117051124573,
      "num_tokens": 13423520.0,
      "step": 1210
    },
    {
      "entropy": 0.5880276575684548,
      "epoch": 2.4015748031496065,
      "grad_norm": 0.439453125,
      "learning_rate": 2.0288786511362525e-05,
      "loss": 0.5333644866943359,
      "mean_token_accuracy": 0.8482611387968063,
      "num_tokens": 13536197.0,
      "step": 1220
    },
    {
      "entropy": 0.5676271229982376,
      "epoch": 2.421259842519685,
      "grad_norm": 0.494140625,
      "learning_rate": 1.9023403649477813e-05,
      "loss": 0.5034989833831787,
      "mean_token_accuracy": 0.8573888152837753,
      "num_tokens": 13649116.0,
      "step": 1230
    },
    {
      "entropy": 0.5038574181497097,
      "epoch": 2.440944881889764,
      "grad_norm": 0.435546875,
      "learning_rate": 1.7794605028160462e-05,
      "loss": 0.45959224700927737,
      "mean_token_accuracy": 0.866667166352272,
      "num_tokens": 13759803.0,
      "step": 1240
    },
    {
      "entropy": 0.5891189277172089,
      "epoch": 2.4606299212598426,
      "grad_norm": 0.6875,
      "learning_rate": 1.6602945803670345e-05,
      "loss": 0.5333945274353027,
      "mean_token_accuracy": 0.8494920641183853,
      "num_tokens": 13872536.0,
      "step": 1250
    },
    {
      "epoch": 2.4606299212598426,
      "eval_entropy": 0.6562480860286288,
      "eval_loss": 0.8404139876365662,
      "eval_mean_token_accuracy": 0.7847034048151087,
      "eval_num_tokens": 13872536.0,
      "eval_runtime": 15.9225,
      "eval_samples_per_second": 26.817,
      "eval_steps_per_second": 3.391,
      "step": 1250
    },
    {
      "entropy": 0.5312069803476334,
      "epoch": 2.4803149606299213,
      "grad_norm": 0.67578125,
      "learning_rate": 1.544896435313997e-05,
      "loss": 0.48858180046081545,
      "mean_token_accuracy": 0.860382902622223,
      "num_tokens": 13982332.0,
      "step": 1260
    },
    {
      "entropy": 0.5351930364966393,
      "epoch": 2.5,
      "grad_norm": 0.53515625,
      "learning_rate": 1.4333182031342185e-05,
      "loss": 0.5056577682495117,
      "mean_token_accuracy": 0.857965275645256,
      "num_tokens": 14096535.0,
      "step": 1270
    },
    {
      "entropy": 0.6251816973090172,
      "epoch": 2.5196850393700787,
      "grad_norm": 0.4921875,
      "learning_rate": 1.3256102935148496e-05,
      "loss": 0.5645822048187256,
      "mean_token_accuracy": 0.8422666400671005,
      "num_tokens": 14214551.0,
      "step": 1280
    },
    {
      "entropy": 0.6227434769272804,
      "epoch": 2.5393700787401574,
      "grad_norm": 0.58984375,
      "learning_rate": 1.2218213675784063e-05,
      "loss": 0.5586906909942627,
      "mean_token_accuracy": 0.8418498873710633,
      "num_tokens": 14322085.0,
      "step": 1290
    },
    {
      "entropy": 0.5474544659256935,
      "epoch": 2.559055118110236,
      "grad_norm": 0.4375,
      "learning_rate": 1.1219983158982917e-05,
      "loss": 0.49657626152038575,
      "mean_token_accuracy": 0.8581330984830856,
      "num_tokens": 14431579.0,
      "step": 1300
    },
    {
      "epoch": 2.559055118110236,
      "eval_entropy": 0.6617519998991931,
      "eval_loss": 0.8383262157440186,
      "eval_mean_token_accuracy": 0.7846880113637006,
      "eval_num_tokens": 14431579.0,
      "eval_runtime": 15.9113,
      "eval_samples_per_second": 26.836,
      "eval_steps_per_second": 3.394,
      "step": 1300
    },
    {
      "entropy": 0.5216698408126831,
      "epoch": 2.578740157480315,
      "grad_norm": 0.6015625,
      "learning_rate": 1.0261862373141839e-05,
      "loss": 0.46427440643310547,
      "mean_token_accuracy": 0.8688691437244416,
      "num_tokens": 14536887.0,
      "step": 1310
    },
    {
      "entropy": 0.5210306510329247,
      "epoch": 2.5984251968503935,
      "grad_norm": 0.5625,
      "learning_rate": 9.344284185569463e-06,
      "loss": 0.4667550563812256,
      "mean_token_accuracy": 0.8664857149124146,
      "num_tokens": 14649040.0,
      "step": 1320
    },
    {
      "entropy": 0.5522045940160751,
      "epoch": 2.6181102362204722,
      "grad_norm": 0.494140625,
      "learning_rate": 8.467663146921979e-06,
      "loss": 0.4914377689361572,
      "mean_token_accuracy": 0.8588613420724869,
      "num_tokens": 14757590.0,
      "step": 1330
    },
    {
      "entropy": 0.5440365478396416,
      "epoch": 2.637795275590551,
      "grad_norm": 0.470703125,
      "learning_rate": 7.632395303914208e-06,
      "loss": 0.4774484634399414,
      "mean_token_accuracy": 0.8620693862438202,
      "num_tokens": 14870411.0,
      "step": 1340
    },
    {
      "entropy": 0.5705034703016281,
      "epoch": 2.65748031496063,
      "grad_norm": 0.515625,
      "learning_rate": 6.838858020390615e-06,
      "loss": 0.5196453094482422,
      "mean_token_accuracy": 0.8570811003446579,
      "num_tokens": 14986705.0,
      "step": 1350
    },
    {
      "epoch": 2.65748031496063,
      "eval_entropy": 0.6574608446271332,
      "eval_loss": 0.8387162685394287,
      "eval_mean_token_accuracy": 0.7850221616250498,
      "eval_num_tokens": 14986705.0,
      "eval_runtime": 15.9224,
      "eval_samples_per_second": 26.818,
      "eval_steps_per_second": 3.391,
      "step": 1350
    },
    {
      "entropy": 0.563168890029192,
      "epoch": 2.677165354330709,
      "grad_norm": 0.640625,
      "learning_rate": 6.087409806836752e-06,
      "loss": 0.5263133525848389,
      "mean_token_accuracy": 0.8563782393932342,
      "num_tokens": 15095469.0,
      "step": 1360
    },
    {
      "entropy": 0.5424789242446423,
      "epoch": 2.6968503937007875,
      "grad_norm": 0.5390625,
      "learning_rate": 5.378390158408731e-06,
      "loss": 0.5066217422485352,
      "mean_token_accuracy": 0.8590426325798035,
      "num_tokens": 15202248.0,
      "step": 1370
    },
    {
      "entropy": 0.5028943344950676,
      "epoch": 2.716535433070866,
      "grad_norm": 0.486328125,
      "learning_rate": 4.712119401553394e-06,
      "loss": 0.4527781009674072,
      "mean_token_accuracy": 0.8694937527179718,
      "num_tokens": 15312525.0,
      "step": 1380
    },
    {
      "entropy": 0.5716224625706673,
      "epoch": 2.736220472440945,
      "grad_norm": 0.427734375,
      "learning_rate": 4.088898549288689e-06,
      "loss": 0.5240719795227051,
      "mean_token_accuracy": 0.8516256272792816,
      "num_tokens": 15424901.0,
      "step": 1390
    },
    {
      "entropy": 0.5507901981472969,
      "epoch": 2.7559055118110236,
      "grad_norm": 0.5234375,
      "learning_rate": 3.5090091652096977e-06,
      "loss": 0.490921688079834,
      "mean_token_accuracy": 0.8620281994342804,
      "num_tokens": 15534177.0,
      "step": 1400
    },
    {
      "epoch": 2.7559055118110236,
      "eval_entropy": 0.659479867528986,
      "eval_loss": 0.8384642601013184,
      "eval_mean_token_accuracy": 0.7847910379921949,
      "eval_num_tokens": 15534177.0,
      "eval_runtime": 15.9298,
      "eval_samples_per_second": 26.805,
      "eval_steps_per_second": 3.39,
      "step": 1400
    },
    {
      "entropy": 0.6328877091407776,
      "epoch": 2.7755905511811023,
      "grad_norm": 0.62109375,
      "learning_rate": 2.972713236281632e-06,
      "loss": 0.5606705188751221,
      "mean_token_accuracy": 0.8396612524986267,
      "num_tokens": 15642725.0,
      "step": 1410
    },
    {
      "entropy": 0.5535789176821708,
      "epoch": 2.795275590551181,
      "grad_norm": 0.48046875,
      "learning_rate": 2.480253054477244e-06,
      "loss": 0.5025914669036865,
      "mean_token_accuracy": 0.8595594942569733,
      "num_tokens": 15749590.0,
      "step": 1420
    },
    {
      "entropy": 0.5887665987014771,
      "epoch": 2.8149606299212597,
      "grad_norm": 0.51171875,
      "learning_rate": 2.031851107312366e-06,
      "loss": 0.5465099334716796,
      "mean_token_accuracy": 0.8492792785167694,
      "num_tokens": 15865143.0,
      "step": 1430
    },
    {
      "entropy": 0.5139057531952858,
      "epoch": 2.8346456692913384,
      "grad_norm": 0.59765625,
      "learning_rate": 1.6277099773286951e-06,
      "loss": 0.4561887264251709,
      "mean_token_accuracy": 0.8726439267396927,
      "num_tokens": 15967667.0,
      "step": 1440
    },
    {
      "entropy": 0.5473496794700623,
      "epoch": 2.8543307086614176,
      "grad_norm": 0.5546875,
      "learning_rate": 1.2680122505694502e-06,
      "loss": 0.4959095001220703,
      "mean_token_accuracy": 0.858627387881279,
      "num_tokens": 16077301.0,
      "step": 1450
    },
    {
      "epoch": 2.8543307086614176,
      "eval_entropy": 0.6591992874940237,
      "eval_loss": 0.8384377956390381,
      "eval_mean_token_accuracy": 0.7848462451387335,
      "eval_num_tokens": 16077301.0,
      "eval_runtime": 15.8973,
      "eval_samples_per_second": 26.86,
      "eval_steps_per_second": 3.397,
      "step": 1450
    },
    {
      "entropy": 0.5839848071336746,
      "epoch": 2.8740157480314963,
      "grad_norm": 0.51953125,
      "learning_rate": 9.529204340892772e-07,
      "loss": 0.5406646251678466,
      "mean_token_accuracy": 0.8526917189359665,
      "num_tokens": 16190635.0,
      "step": 1460
    },
    {
      "entropy": 0.5787345126271248,
      "epoch": 2.893700787401575,
      "grad_norm": 0.796875,
      "learning_rate": 6.825768825355017e-07,
      "loss": 0.5195184230804444,
      "mean_token_accuracy": 0.8524530798196792,
      "num_tokens": 16304259.0,
      "step": 1470
    },
    {
      "entropy": 0.5098958224058151,
      "epoch": 2.9133858267716537,
      "grad_norm": 0.5390625,
      "learning_rate": 4.5710373383401895e-07,
      "loss": 0.4449178695678711,
      "mean_token_accuracy": 0.8701013952493668,
      "num_tokens": 16414959.0,
      "step": 1480
    },
    {
      "entropy": 0.5218460902571678,
      "epoch": 2.9330708661417324,
      "grad_norm": 0.53125,
      "learning_rate": 2.766028540089227e-07,
      "loss": 0.475307559967041,
      "mean_token_accuracy": 0.863593265414238,
      "num_tokens": 16531149.0,
      "step": 1490
    },
    {
      "entropy": 0.5244145601987839,
      "epoch": 2.952755905511811,
      "grad_norm": 0.640625,
      "learning_rate": 1.4115579116061873e-07,
      "loss": 0.4811119556427002,
      "mean_token_accuracy": 0.863358524441719,
      "num_tokens": 16634608.0,
      "step": 1500
    },
    {
      "epoch": 2.952755905511811,
      "eval_entropy": 0.6589832990257828,
      "eval_loss": 0.838596761226654,
      "eval_mean_token_accuracy": 0.7848243316014608,
      "eval_num_tokens": 16634608.0,
      "eval_runtime": 15.9196,
      "eval_samples_per_second": 26.822,
      "eval_steps_per_second": 3.392,
      "step": 1500
    }
  ],
  "logging_steps": 10,
  "max_steps": 1524,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.3507235758203085e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}