{
  "best_global_step": 2000,
  "best_metric": 0.44610196352005005,
  "best_model_checkpoint": "/data6/jongbinwon/evolving_MoE/checkpoints/sft_llama31_8b_lbox_statutory-element-matcher_fallback_ffn_only_5ep_eval500/checkpoint-2000",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 6480,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 2.256485325098038,
      "epoch": 0.007720517274657402,
      "grad_norm": 7.307734489440918,
      "learning_rate": 9.230769230769232e-07,
      "loss": 2.06158504486084,
      "mean_token_accuracy": 0.6090951159596443,
      "num_tokens": 32810.0,
      "step": 10
    },
    {
      "entropy": 2.284497153759003,
      "epoch": 0.015441034549314805,
      "grad_norm": 8.211675643920898,
      "learning_rate": 1.948717948717949e-06,
      "loss": 1.9772743225097655,
      "mean_token_accuracy": 0.6226175807416439,
      "num_tokens": 67674.0,
      "step": 20
    },
    {
      "entropy": 2.2925585120916367,
      "epoch": 0.023161551823972205,
      "grad_norm": 14.34852123260498,
      "learning_rate": 2.9743589743589746e-06,
      "loss": 1.8409126281738282,
      "mean_token_accuracy": 0.6243047766387463,
      "num_tokens": 105820.0,
      "step": 30
    },
    {
      "entropy": 2.257777214050293,
      "epoch": 0.03088206909862961,
      "grad_norm": 9.495259284973145,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.7386272430419922,
      "mean_token_accuracy": 0.6371752314269543,
      "num_tokens": 142765.0,
      "step": 40
    },
    {
      "entropy": 2.2558146893978117,
      "epoch": 0.038602586373287014,
      "grad_norm": 4.8525285720825195,
      "learning_rate": 5.025641025641026e-06,
      "loss": 1.5439723968505858,
      "mean_token_accuracy": 0.6804775930941105,
      "num_tokens": 179804.0,
      "step": 50
    },
    {
      "entropy": 2.3024860441684725,
      "epoch": 0.04632310364794441,
      "grad_norm": 4.64549446105957,
      "learning_rate": 6.051282051282051e-06,
      "loss": 1.251286220550537,
      "mean_token_accuracy": 0.7414289742708207,
      "num_tokens": 212054.0,
      "step": 60
    },
    {
      "entropy": 2.3543750554323197,
      "epoch": 0.054043620922601815,
      "grad_norm": 3.5757696628570557,
      "learning_rate": 7.076923076923078e-06,
      "loss": 0.9379256248474122,
      "mean_token_accuracy": 0.7693132214248181,
      "num_tokens": 245935.0,
      "step": 70
    },
    {
      "entropy": 2.3174787133932115,
      "epoch": 0.06176413819725922,
      "grad_norm": 4.677041530609131,
      "learning_rate": 8.102564102564103e-06,
      "loss": 1.0049633979797363,
      "mean_token_accuracy": 0.770514677464962,
      "num_tokens": 283556.0,
      "step": 80
    },
    {
      "entropy": 2.358958587050438,
      "epoch": 0.06948465547191662,
      "grad_norm": 3.2004008293151855,
      "learning_rate": 9.128205128205129e-06,
      "loss": 0.966558837890625,
      "mean_token_accuracy": 0.7766762919723987,
      "num_tokens": 319043.0,
      "step": 90
    },
    {
      "entropy": 2.3884228974580766,
      "epoch": 0.07720517274657403,
      "grad_norm": 4.562836647033691,
      "learning_rate": 1.0153846153846154e-05,
      "loss": 0.8094841003417969,
      "mean_token_accuracy": 0.8071015253663063,
      "num_tokens": 352565.0,
      "step": 100
    },
    {
      "entropy": 2.367259520292282,
      "epoch": 0.08492569002123142,
      "grad_norm": 4.451163291931152,
      "learning_rate": 1.117948717948718e-05,
      "loss": 0.6964176177978516,
      "mean_token_accuracy": 0.829041276872158,
      "num_tokens": 385068.0,
      "step": 110
    },
    {
      "entropy": 2.348944702744484,
      "epoch": 0.09264620729588882,
      "grad_norm": 3.5486667156219482,
      "learning_rate": 1.2205128205128208e-05,
      "loss": 0.5524073600769043,
      "mean_token_accuracy": 0.8610983595252037,
      "num_tokens": 417434.0,
      "step": 120
    },
    {
      "entropy": 2.362921339273453,
      "epoch": 0.10036672457054623,
      "grad_norm": 3.7314867973327637,
      "learning_rate": 1.3230769230769231e-05,
      "loss": 0.524355697631836,
      "mean_token_accuracy": 0.862143586575985,
      "num_tokens": 451513.0,
      "step": 130
    },
    {
      "entropy": 2.3315255403518678,
      "epoch": 0.10808724184520363,
      "grad_norm": 3.9957523345947266,
      "learning_rate": 1.4256410256410258e-05,
      "loss": 0.5264537334442139,
      "mean_token_accuracy": 0.8657179176807404,
      "num_tokens": 488960.0,
      "step": 140
    },
    {
      "entropy": 2.272639125585556,
      "epoch": 0.11580775911986103,
      "grad_norm": 6.005861759185791,
      "learning_rate": 1.5282051282051282e-05,
      "loss": 0.5440045833587647,
      "mean_token_accuracy": 0.846938057243824,
      "num_tokens": 522139.0,
      "step": 150
    },
    {
      "entropy": 2.355533537268639,
      "epoch": 0.12352827639451844,
      "grad_norm": 6.817739009857178,
      "learning_rate": 1.630769230769231e-05,
      "loss": 0.39336931705474854,
      "mean_token_accuracy": 0.9171057492494583,
      "num_tokens": 554659.0,
      "step": 160
    },
    {
      "entropy": 2.1675438672304153,
      "epoch": 0.13124879366917583,
      "grad_norm": 4.914283275604248,
      "learning_rate": 1.7333333333333336e-05,
      "loss": 0.49741554260253906,
      "mean_token_accuracy": 0.8793981537222862,
      "num_tokens": 595435.0,
      "step": 170
    },
    {
      "entropy": 2.2984329521656037,
      "epoch": 0.13896931094383325,
      "grad_norm": 6.025143623352051,
      "learning_rate": 1.835897435897436e-05,
      "loss": 0.3462223529815674,
      "mean_token_accuracy": 0.8944209352135658,
      "num_tokens": 628526.0,
      "step": 180
    },
    {
      "entropy": 2.3169893980026246,
      "epoch": 0.14668982821849064,
      "grad_norm": 5.91893196105957,
      "learning_rate": 1.9384615384615386e-05,
      "loss": 0.31524474620819093,
      "mean_token_accuracy": 0.9156093567609787,
      "num_tokens": 665610.0,
      "step": 190
    },
    {
      "entropy": 2.301328441500664,
      "epoch": 0.15441034549314805,
      "grad_norm": 3.231717109680176,
      "learning_rate": 1.9999980011554353e-05,
      "loss": 0.32752580642700196,
      "mean_token_accuracy": 0.910846720635891,
      "num_tokens": 706015.0,
      "step": 200
    },
    {
      "entropy": 2.273413100838661,
      "epoch": 0.16213086276780544,
      "grad_norm": 3.171372652053833,
      "learning_rate": 1.9999755142458468e-05,
      "loss": 0.37491621971130373,
      "mean_token_accuracy": 0.9038208425045013,
      "num_tokens": 742435.0,
      "step": 210
    },
    {
      "entropy": 2.2427696734666824,
      "epoch": 0.16985138004246284,
      "grad_norm": 6.08746337890625,
      "learning_rate": 1.9999280424346836e-05,
      "loss": 0.2458643436431885,
      "mean_token_accuracy": 0.9297083392739296,
      "num_tokens": 774452.0,
      "step": 220
    },
    {
      "entropy": 2.315064123272896,
      "epoch": 0.17757189731712025,
      "grad_norm": 6.792384147644043,
      "learning_rate": 1.9998555869080533e-05,
      "loss": 0.3388651132583618,
      "mean_token_accuracy": 0.9051751494407654,
      "num_tokens": 804925.0,
      "step": 230
    },
    {
      "entropy": 2.299090850353241,
      "epoch": 0.18529241459177764,
      "grad_norm": 6.4787468910217285,
      "learning_rate": 1.999758149476294e-05,
      "loss": 0.339692497253418,
      "mean_token_accuracy": 0.9080843642354012,
      "num_tokens": 845968.0,
      "step": 240
    },
    {
      "entropy": 2.280930459499359,
      "epoch": 0.19301293186643506,
      "grad_norm": 5.793071746826172,
      "learning_rate": 1.9996357325739306e-05,
      "loss": 0.27950329780578614,
      "mean_token_accuracy": 0.9254253804683685,
      "num_tokens": 879844.0,
      "step": 250
    },
    {
      "entropy": 2.2729754209518434,
      "epoch": 0.20073344914109245,
      "grad_norm": 7.555527687072754,
      "learning_rate": 1.999488339259612e-05,
      "loss": 0.30706956386566164,
      "mean_token_accuracy": 0.9272728309035301,
      "num_tokens": 913158.0,
      "step": 260
    },
    {
      "entropy": 2.267097595334053,
      "epoch": 0.20845396641574984,
      "grad_norm": 2.7754158973693848,
      "learning_rate": 1.9993159732160353e-05,
      "loss": 0.2901525735855103,
      "mean_token_accuracy": 0.9153242215514183,
      "num_tokens": 952998.0,
      "step": 270
    },
    {
      "entropy": 2.2886367619037626,
      "epoch": 0.21617448369040726,
      "grad_norm": 5.025292873382568,
      "learning_rate": 1.999118638749855e-05,
      "loss": 0.23909380435943603,
      "mean_token_accuracy": 0.9286381289362907,
      "num_tokens": 988029.0,
      "step": 280
    },
    {
      "entropy": 2.215249925851822,
      "epoch": 0.22389500096506465,
      "grad_norm": 2.3225855827331543,
      "learning_rate": 1.9988963407915743e-05,
      "loss": 0.24327101707458496,
      "mean_token_accuracy": 0.9352555975317955,
      "num_tokens": 1023017.0,
      "step": 290
    },
    {
      "entropy": 2.231352314352989,
      "epoch": 0.23161551823972207,
      "grad_norm": 3.9783694744110107,
      "learning_rate": 1.9986490848954213e-05,
      "loss": 0.28902668952941896,
      "mean_token_accuracy": 0.9245402902364731,
      "num_tokens": 1058144.0,
      "step": 300
    },
    {
      "entropy": 2.283156764507294,
      "epoch": 0.23933603551437946,
      "grad_norm": 4.462434768676758,
      "learning_rate": 1.9983768772392122e-05,
      "loss": 0.34264867305755614,
      "mean_token_accuracy": 0.9317045152187348,
      "num_tokens": 1089081.0,
      "step": 310
    },
    {
      "entropy": 2.184281289577484,
      "epoch": 0.24705655278903688,
      "grad_norm": 4.392584800720215,
      "learning_rate": 1.998079724624194e-05,
      "loss": 0.21626288890838624,
      "mean_token_accuracy": 0.9372277542948723,
      "num_tokens": 1128469.0,
      "step": 320
    },
    {
      "entropy": 2.2497308939695357,
      "epoch": 0.25477707006369427,
      "grad_norm": 4.71306848526001,
      "learning_rate": 1.997757634474878e-05,
      "loss": 0.18063234090805053,
      "mean_token_accuracy": 0.9538213908672333,
      "num_tokens": 1164636.0,
      "step": 330
    },
    {
      "entropy": 2.2387163996696473,
      "epoch": 0.26249758733835166,
      "grad_norm": 1.666062831878662,
      "learning_rate": 1.997410614838852e-05,
      "loss": 0.1772278666496277,
      "mean_token_accuracy": 0.9541213810443878,
      "num_tokens": 1198253.0,
      "step": 340
    },
    {
      "entropy": 2.2753040075302122,
      "epoch": 0.27021810461300905,
      "grad_norm": 3.6660280227661133,
      "learning_rate": 1.9970386743865794e-05,
      "loss": 0.191804039478302,
      "mean_token_accuracy": 0.9456240653991699,
      "num_tokens": 1230490.0,
      "step": 350
    },
    {
      "entropy": 2.2174918383359907,
      "epoch": 0.2779386218876665,
      "grad_norm": 6.83668851852417,
      "learning_rate": 1.9966418224111838e-05,
      "loss": 0.20989911556243895,
      "mean_token_accuracy": 0.9410885244607925,
      "num_tokens": 1267745.0,
      "step": 360
    },
    {
      "entropy": 2.219417518377304,
      "epoch": 0.2856591391623239,
      "grad_norm": 3.405336618423462,
      "learning_rate": 1.996220068828215e-05,
      "loss": 0.12719086408615113,
      "mean_token_accuracy": 0.9730818659067154,
      "num_tokens": 1303473.0,
      "step": 370
    },
    {
      "entropy": 2.2125363498926163,
      "epoch": 0.2933796564369813,
      "grad_norm": 4.977858066558838,
      "learning_rate": 1.995773424175403e-05,
      "loss": 0.23458333015441896,
      "mean_token_accuracy": 0.9394760414958,
      "num_tokens": 1338018.0,
      "step": 380
    },
    {
      "entropy": 2.2556005716323853,
      "epoch": 0.30110017371163866,
      "grad_norm": 4.019962787628174,
      "learning_rate": 1.9953018996123942e-05,
      "loss": 0.20530598163604735,
      "mean_token_accuracy": 0.9385380610823632,
      "num_tokens": 1373244.0,
      "step": 390
    },
    {
      "entropy": 2.2015073150396347,
      "epoch": 0.3088206909862961,
      "grad_norm": 2.728677749633789,
      "learning_rate": 1.9948055069204716e-05,
      "loss": 0.2082897901535034,
      "mean_token_accuracy": 0.9512322708964348,
      "num_tokens": 1407807.0,
      "step": 400
    },
    {
      "entropy": 2.273113363981247,
      "epoch": 0.3165412082609535,
      "grad_norm": 5.641965866088867,
      "learning_rate": 1.9942842585022607e-05,
      "loss": 0.16474014520645142,
      "mean_token_accuracy": 0.9578953236341476,
      "num_tokens": 1439941.0,
      "step": 410
    },
    {
      "entropy": 2.233418434858322,
      "epoch": 0.3242617255356109,
      "grad_norm": 1.6350418329238892,
      "learning_rate": 1.993738167381422e-05,
      "loss": 0.2152254104614258,
      "mean_token_accuracy": 0.9460450857877731,
      "num_tokens": 1476301.0,
      "step": 420
    },
    {
      "entropy": 2.2233593642711638,
      "epoch": 0.3319822428102683,
      "grad_norm": 5.32601261138916,
      "learning_rate": 1.9931672472023212e-05,
      "loss": 0.18065141439437865,
      "mean_token_accuracy": 0.9680908665060997,
      "num_tokens": 1514277.0,
      "step": 430
    },
    {
      "entropy": 2.13372061252594,
      "epoch": 0.33970276008492567,
      "grad_norm": 2.3526391983032227,
      "learning_rate": 1.992571512229693e-05,
      "loss": 0.19432415962219238,
      "mean_token_accuracy": 0.957370612025261,
      "num_tokens": 1552244.0,
      "step": 440
    },
    {
      "entropy": 2.2559954702854155,
      "epoch": 0.3474232773595831,
      "grad_norm": 4.275816440582275,
      "learning_rate": 1.9919509773482816e-05,
      "loss": 0.1578107237815857,
      "mean_token_accuracy": 0.9568435192108155,
      "num_tokens": 1589032.0,
      "step": 450
    },
    {
      "entropy": 2.1852520048618316,
      "epoch": 0.3551437946342405,
      "grad_norm": 4.293900489807129,
      "learning_rate": 1.991305658062469e-05,
      "loss": 0.23177709579467773,
      "mean_token_accuracy": 0.9536217465996742,
      "num_tokens": 1623926.0,
      "step": 460
    },
    {
      "entropy": 2.2606248050928115,
      "epoch": 0.3628643119088979,
      "grad_norm": 2.29001784324646,
      "learning_rate": 1.9906355704958895e-05,
      "loss": 0.17753545045852662,
      "mean_token_accuracy": 0.9497943684458733,
      "num_tokens": 1660932.0,
      "step": 470
    },
    {
      "entropy": 2.240851789712906,
      "epoch": 0.3705848291835553,
      "grad_norm": 6.091527462005615,
      "learning_rate": 1.989940731391024e-05,
      "loss": 0.1601572871208191,
      "mean_token_accuracy": 0.9648889541625977,
      "num_tokens": 1699168.0,
      "step": 480
    },
    {
      "entropy": 2.2124760150909424,
      "epoch": 0.3783053464582127,
      "grad_norm": 1.9111343622207642,
      "learning_rate": 1.9892211581087854e-05,
      "loss": 0.18850926160812378,
      "mean_token_accuracy": 0.950583329796791,
      "num_tokens": 1732169.0,
      "step": 490
    },
    {
      "entropy": 2.1586394786834715,
      "epoch": 0.3860258637328701,
      "grad_norm": 2.544804573059082,
      "learning_rate": 1.9884768686280807e-05,
      "loss": 0.16971189975738527,
      "mean_token_accuracy": 0.9592921108007431,
      "num_tokens": 1765074.0,
      "step": 500
    },
    {
      "epoch": 0.3860258637328701,
      "eval_entropy": 2.1075005949851486,
      "eval_loss": 0.5422846674919128,
      "eval_mean_token_accuracy": 0.8941449357922673,
      "eval_num_tokens": 1765074.0,
      "eval_runtime": 988.7029,
      "eval_samples_per_second": 7.738,
      "eval_steps_per_second": 0.968,
      "step": 500
    },
    {
      "entropy": 2.2212108463048934,
      "epoch": 0.3937463810075275,
      "grad_norm": 5.165001392364502,
      "learning_rate": 1.9877078815453648e-05,
      "loss": 0.2070312023162842,
      "mean_token_accuracy": 0.9447736650705337,
      "num_tokens": 1803147.0,
      "step": 510
    },
    {
      "entropy": 2.1793356448411942,
      "epoch": 0.4014668982821849,
      "grad_norm": 3.1133415699005127,
      "learning_rate": 1.9869142160741748e-05,
      "loss": 0.15652704238891602,
      "mean_token_accuracy": 0.9599319085478782,
      "num_tokens": 1836681.0,
      "step": 520
    },
    {
      "entropy": 2.1871384173631667,
      "epoch": 0.4091874155568423,
      "grad_norm": 2.828554153442383,
      "learning_rate": 1.9860958920446503e-05,
      "loss": 0.18446458578109742,
      "mean_token_accuracy": 0.9459314718842506,
      "num_tokens": 1872127.0,
      "step": 530
    },
    {
      "entropy": 2.2149875700473785,
      "epoch": 0.4169079328314997,
      "grad_norm": 1.8423798084259033,
      "learning_rate": 1.985252929903037e-05,
      "loss": 0.1715518832206726,
      "mean_token_accuracy": 0.9680211797356606,
      "num_tokens": 1905318.0,
      "step": 540
    },
    {
      "entropy": 2.2457988798618316,
      "epoch": 0.42462845010615713,
      "grad_norm": 0.8942133188247681,
      "learning_rate": 1.9843853507111764e-05,
      "loss": 0.19799450635910035,
      "mean_token_accuracy": 0.954279862344265,
      "num_tokens": 1942161.0,
      "step": 550
    },
    {
      "entropy": 2.2178206920623778,
      "epoch": 0.4323489673808145,
      "grad_norm": 4.0149455070495605,
      "learning_rate": 1.9834931761459804e-05,
      "loss": 0.16577671766281127,
      "mean_token_accuracy": 0.9597131043672562,
      "num_tokens": 1977752.0,
      "step": 560
    },
    {
      "entropy": 2.2375432044267654,
      "epoch": 0.4400694846554719,
      "grad_norm": 2.153841495513916,
      "learning_rate": 1.9825764284988884e-05,
      "loss": 0.17870255708694457,
      "mean_token_accuracy": 0.9546808168292046,
      "num_tokens": 2012513.0,
      "step": 570
    },
    {
      "entropy": 2.25352583527565,
      "epoch": 0.4477900019301293,
      "grad_norm": 5.3909592628479,
      "learning_rate": 1.9816351306753112e-05,
      "loss": 0.10405315160751342,
      "mean_token_accuracy": 0.965131339430809,
      "num_tokens": 2046420.0,
      "step": 580
    },
    {
      "entropy": 2.154633605480194,
      "epoch": 0.45551051920478675,
      "grad_norm": 3.521245002746582,
      "learning_rate": 1.980669306194058e-05,
      "loss": 0.20279982089996337,
      "mean_token_accuracy": 0.9519533723592758,
      "num_tokens": 2084016.0,
      "step": 590
    },
    {
      "entropy": 2.186835992336273,
      "epoch": 0.46323103647944414,
      "grad_norm": 2.632781982421875,
      "learning_rate": 1.979678979186749e-05,
      "loss": 0.18113456964492797,
      "mean_token_accuracy": 0.9502775952219963,
      "num_tokens": 2117364.0,
      "step": 600
    },
    {
      "entropy": 2.189032417535782,
      "epoch": 0.4709515537541015,
      "grad_norm": 3.7245187759399414,
      "learning_rate": 1.9786641743972135e-05,
      "loss": 0.17686980962753296,
      "mean_token_accuracy": 0.9488097980618477,
      "num_tokens": 2151634.0,
      "step": 610
    },
    {
      "entropy": 2.2037782222032547,
      "epoch": 0.4786720710287589,
      "grad_norm": 4.016031742095947,
      "learning_rate": 1.9776249171808693e-05,
      "loss": 0.15398112535476685,
      "mean_token_accuracy": 0.9585451945662499,
      "num_tokens": 2187502.0,
      "step": 620
    },
    {
      "entropy": 2.124007895588875,
      "epoch": 0.4863925883034163,
      "grad_norm": 0.7118439078330994,
      "learning_rate": 1.9765612335040923e-05,
      "loss": 0.1368114948272705,
      "mean_token_accuracy": 0.9596732050180435,
      "num_tokens": 2219123.0,
      "step": 630
    },
    {
      "entropy": 2.1975839883089066,
      "epoch": 0.49411310557807375,
      "grad_norm": 7.007303714752197,
      "learning_rate": 1.9754731499435648e-05,
      "loss": 0.1872836470603943,
      "mean_token_accuracy": 0.953962279856205,
      "num_tokens": 2249877.0,
      "step": 640
    },
    {
      "entropy": 2.152134519815445,
      "epoch": 0.5018336228527311,
      "grad_norm": 5.597184658050537,
      "learning_rate": 1.9743606936856134e-05,
      "loss": 0.12075231075286866,
      "mean_token_accuracy": 0.9612849146127701,
      "num_tokens": 2290176.0,
      "step": 650
    },
    {
      "entropy": 2.168921798467636,
      "epoch": 0.5095541401273885,
      "grad_norm": 2.507140636444092,
      "learning_rate": 1.9732238925255282e-05,
      "loss": 0.17506990432739258,
      "mean_token_accuracy": 0.9573684856295586,
      "num_tokens": 2326102.0,
      "step": 660
    },
    {
      "entropy": 2.2062420308589936,
      "epoch": 0.5172746574020459,
      "grad_norm": 0.6417449116706848,
      "learning_rate": 1.9720627748668705e-05,
      "loss": 0.13540934324264525,
      "mean_token_accuracy": 0.9623032554984092,
      "num_tokens": 2360241.0,
      "step": 670
    },
    {
      "entropy": 2.078236073255539,
      "epoch": 0.5249951746767033,
      "grad_norm": 6.3616623878479,
      "learning_rate": 1.9708773697207608e-05,
      "loss": 0.15398894548416137,
      "mean_token_accuracy": 0.9607418894767761,
      "num_tokens": 2404406.0,
      "step": 680
    },
    {
      "entropy": 2.111809679865837,
      "epoch": 0.5327156919513607,
      "grad_norm": 6.101365089416504,
      "learning_rate": 1.969667706705155e-05,
      "loss": 0.21354906558990477,
      "mean_token_accuracy": 0.9489550918340683,
      "num_tokens": 2440018.0,
      "step": 690
    },
    {
      "entropy": 2.1686588764190673,
      "epoch": 0.5404362092260181,
      "grad_norm": 3.270843982696533,
      "learning_rate": 1.9684338160441045e-05,
      "loss": 0.14644697904586793,
      "mean_token_accuracy": 0.9603226110339165,
      "num_tokens": 2476784.0,
      "step": 700
    },
    {
      "entropy": 2.166987511515617,
      "epoch": 0.5481567265006756,
      "grad_norm": 4.874166965484619,
      "learning_rate": 1.9671757285670013e-05,
      "loss": 0.16295208930969238,
      "mean_token_accuracy": 0.9569882333278656,
      "num_tokens": 2512350.0,
      "step": 710
    },
    {
      "entropy": 2.252730244398117,
      "epoch": 0.555877243775333,
      "grad_norm": 3.095005512237549,
      "learning_rate": 1.9658934757078067e-05,
      "loss": 0.11499756574630737,
      "mean_token_accuracy": 0.9637171044945717,
      "num_tokens": 2541581.0,
      "step": 720
    },
    {
      "entropy": 2.1925143748521805,
      "epoch": 0.5635977610499904,
      "grad_norm": 2.961524486541748,
      "learning_rate": 1.9645870895042665e-05,
      "loss": 0.0865030288696289,
      "mean_token_accuracy": 0.9704707160592079,
      "num_tokens": 2575015.0,
      "step": 730
    },
    {
      "entropy": 2.1793641477823256,
      "epoch": 0.5713182783246478,
      "grad_norm": 3.9880712032318115,
      "learning_rate": 1.9632566025971114e-05,
      "loss": 0.1488552212715149,
      "mean_token_accuracy": 0.9751747816801071,
      "num_tokens": 2610061.0,
      "step": 740
    },
    {
      "entropy": 2.172965335845947,
      "epoch": 0.5790387955993052,
      "grad_norm": 5.618796348571777,
      "learning_rate": 1.9619020482292386e-05,
      "loss": 0.11244801282882691,
      "mean_token_accuracy": 0.9690556541085243,
      "num_tokens": 2645126.0,
      "step": 750
    },
    {
      "entropy": 2.114529776573181,
      "epoch": 0.5867593128739625,
      "grad_norm": 4.01426362991333,
      "learning_rate": 1.960523460244885e-05,
      "loss": 0.14911041259765626,
      "mean_token_accuracy": 0.9594126790761948,
      "num_tokens": 2683118.0,
      "step": 760
    },
    {
      "entropy": 2.2817918568849564,
      "epoch": 0.5944798301486199,
      "grad_norm": 3.577343225479126,
      "learning_rate": 1.959120873088779e-05,
      "loss": 0.17379360198974608,
      "mean_token_accuracy": 0.958151726424694,
      "num_tokens": 2717012.0,
      "step": 770
    },
    {
      "entropy": 2.2415023863315584,
      "epoch": 0.6022003474232773,
      "grad_norm": 0.9728232622146606,
      "learning_rate": 1.9576943218052813e-05,
      "loss": 0.16758315563201903,
      "mean_token_accuracy": 0.9584022372961044,
      "num_tokens": 2750233.0,
      "step": 780
    },
    {
      "entropy": 2.2567327469587326,
      "epoch": 0.6099208646979347,
      "grad_norm": 4.933441638946533,
      "learning_rate": 1.956243842037507e-05,
      "loss": 0.1237065315246582,
      "mean_token_accuracy": 0.9687949195504189,
      "num_tokens": 2784593.0,
      "step": 790
    },
    {
      "entropy": 2.1688129603862762,
      "epoch": 0.6176413819725922,
      "grad_norm": 0.3439030349254608,
      "learning_rate": 1.9547694700264387e-05,
      "loss": 0.12909990549087524,
      "mean_token_accuracy": 0.963108229637146,
      "num_tokens": 2828912.0,
      "step": 800
    },
    {
      "entropy": 2.2464480668306352,
      "epoch": 0.6253618992472496,
      "grad_norm": 1.731982946395874,
      "learning_rate": 1.953271242610017e-05,
      "loss": 0.13596073389053345,
      "mean_token_accuracy": 0.974192063510418,
      "num_tokens": 2864199.0,
      "step": 810
    },
    {
      "entropy": 2.2575213581323625,
      "epoch": 0.633082416521907,
      "grad_norm": 2.722280263900757,
      "learning_rate": 1.951749197222224e-05,
      "loss": 0.12859855890274047,
      "mean_token_accuracy": 0.9683046102523803,
      "num_tokens": 2902141.0,
      "step": 820
    },
    {
      "entropy": 2.182758465409279,
      "epoch": 0.6408029337965644,
      "grad_norm": 2.9258344173431396,
      "learning_rate": 1.9502033718921444e-05,
      "loss": 0.09067170023918152,
      "mean_token_accuracy": 0.9700423076748848,
      "num_tokens": 2941709.0,
      "step": 830
    },
    {
      "entropy": 2.1869940221309663,
      "epoch": 0.6485234510712218,
      "grad_norm": 1.2507781982421875,
      "learning_rate": 1.948633805243018e-05,
      "loss": 0.06829038262367249,
      "mean_token_accuracy": 0.9780937045812607,
      "num_tokens": 2976754.0,
      "step": 840
    },
    {
      "entropy": 2.240604591369629,
      "epoch": 0.6562439683458792,
      "grad_norm": 6.191175937652588,
      "learning_rate": 1.9470405364912737e-05,
      "loss": 0.1681974411010742,
      "mean_token_accuracy": 0.962010458111763,
      "num_tokens": 3013223.0,
      "step": 850
    },
    {
      "entropy": 2.1933089971542357,
      "epoch": 0.6639644856205366,
      "grad_norm": 3.580299139022827,
      "learning_rate": 1.945423605445548e-05,
      "loss": 0.14260369539260864,
      "mean_token_accuracy": 0.9581344783306122,
      "num_tokens": 3051770.0,
      "step": 860
    },
    {
      "entropy": 2.2846630066633224,
      "epoch": 0.671685002895194,
      "grad_norm": 0.303085058927536,
      "learning_rate": 1.943783052505694e-05,
      "loss": 0.10166503190994262,
      "mean_token_accuracy": 0.9790826007723809,
      "num_tokens": 3086529.0,
      "step": 870
    },
    {
      "entropy": 2.229181852936745,
      "epoch": 0.6794055201698513,
      "grad_norm": 3.7699837684631348,
      "learning_rate": 1.9421189186617692e-05,
      "loss": 0.09227925539016724,
      "mean_token_accuracy": 0.9764371633529663,
      "num_tokens": 3132264.0,
      "step": 880
    },
    {
      "entropy": 2.2480290710926054,
      "epoch": 0.6871260374445087,
      "grad_norm": 1.3305270671844482,
      "learning_rate": 1.9404312454930116e-05,
      "loss": 0.13547370433807374,
      "mean_token_accuracy": 0.9761623114347457,
      "num_tokens": 3168569.0,
      "step": 890
    },
    {
      "entropy": 2.174534446001053,
      "epoch": 0.6948465547191662,
      "grad_norm": 0.882283627986908,
      "learning_rate": 1.9387200751668018e-05,
      "loss": 0.12927967309951782,
      "mean_token_accuracy": 0.9645377054810524,
      "num_tokens": 3203700.0,
      "step": 900
    },
    {
      "entropy": 2.1783421456813814,
      "epoch": 0.7025670719938236,
      "grad_norm": 3.659027099609375,
      "learning_rate": 1.9369854504376092e-05,
      "loss": 0.08977473378181458,
      "mean_token_accuracy": 0.9750412046909332,
      "num_tokens": 3244225.0,
      "step": 910
    },
    {
      "entropy": 2.2626183718442916,
      "epoch": 0.710287589268481,
      "grad_norm": 0.5882543921470642,
      "learning_rate": 1.9352274146459223e-05,
      "loss": 0.13526970148086548,
      "mean_token_accuracy": 0.9642464354634285,
      "num_tokens": 3282163.0,
      "step": 920
    },
    {
      "entropy": 2.2082873970270156,
      "epoch": 0.7180081065431384,
      "grad_norm": 3.8217687606811523,
      "learning_rate": 1.9334460117171687e-05,
      "loss": 0.10937002897262574,
      "mean_token_accuracy": 0.9754633396863938,
      "num_tokens": 3316191.0,
      "step": 930
    },
    {
      "entropy": 2.177019494771957,
      "epoch": 0.7257286238177958,
      "grad_norm": 0.9711639881134033,
      "learning_rate": 1.931641286160615e-05,
      "loss": 0.12394155263900757,
      "mean_token_accuracy": 0.9672280788421631,
      "num_tokens": 3353504.0,
      "step": 940
    },
    {
      "entropy": 2.159817245602608,
      "epoch": 0.7334491410924532,
      "grad_norm": 3.0002224445343018,
      "learning_rate": 1.9298132830682553e-05,
      "loss": 0.07563741207122802,
      "mean_token_accuracy": 0.9817924559116363,
      "num_tokens": 3390473.0,
      "step": 950
    },
    {
      "entropy": 2.1697775393724443,
      "epoch": 0.7411696583671106,
      "grad_norm": 4.114264965057373,
      "learning_rate": 1.9279620481136854e-05,
      "loss": 0.10978723764419555,
      "mean_token_accuracy": 0.9763004094362259,
      "num_tokens": 3429080.0,
      "step": 960
    },
    {
      "entropy": 2.154700428247452,
      "epoch": 0.748890175641768,
      "grad_norm": 2.0637450218200684,
      "learning_rate": 1.9260876275509614e-05,
      "loss": 0.12881894111633302,
      "mean_token_accuracy": 0.9691078916192055,
      "num_tokens": 3472222.0,
      "step": 970
    },
    {
      "entropy": 2.1561296582221985,
      "epoch": 0.7566106929164254,
      "grad_norm": 5.698066711425781,
      "learning_rate": 1.924190068213443e-05,
      "loss": 0.12960790395736693,
      "mean_token_accuracy": 0.97647725045681,
      "num_tokens": 3517841.0,
      "step": 980
    },
    {
      "entropy": 2.2234013736248017,
      "epoch": 0.7643312101910829,
      "grad_norm": 2.6206040382385254,
      "learning_rate": 1.9222694175126242e-05,
      "loss": 0.09993091225624084,
      "mean_token_accuracy": 0.9791779488325119,
      "num_tokens": 3554442.0,
      "step": 990
    },
    {
      "entropy": 2.180557644367218,
      "epoch": 0.7720517274657402,
      "grad_norm": 0.4288350045681,
      "learning_rate": 1.920325723436949e-05,
      "loss": 0.18079398870468139,
      "mean_token_accuracy": 0.9695220619440079,
      "num_tokens": 3589489.0,
      "step": 1000
    },
    {
      "epoch": 0.7720517274657402,
      "eval_entropy": 2.093520096975185,
      "eval_loss": 0.45434409379959106,
      "eval_mean_token_accuracy": 0.9182380637155929,
      "eval_num_tokens": 3589489.0,
      "eval_runtime": 990.2181,
      "eval_samples_per_second": 7.727,
      "eval_steps_per_second": 0.966,
      "step": 1000
    },
    {
      "entropy": 2.2155666559934617,
      "epoch": 0.7797722447403976,
      "grad_norm": 2.715181589126587,
      "learning_rate": 1.9183590345506115e-05,
      "loss": 0.07064727544784546,
      "mean_token_accuracy": 0.9841662392020225,
      "num_tokens": 3622362.0,
      "step": 1010
    },
    {
      "entropy": 2.1746787667274474,
      "epoch": 0.787492762015055,
      "grad_norm": 3.680882692337036,
      "learning_rate": 1.916369399992344e-05,
      "loss": 0.06015622615814209,
      "mean_token_accuracy": 0.9877562940120697,
      "num_tokens": 3656283.0,
      "step": 1020
    },
    {
      "entropy": 2.174808195233345,
      "epoch": 0.7952132792897124,
      "grad_norm": 0.3832246959209442,
      "learning_rate": 1.9143568694741858e-05,
      "loss": 0.11039963960647584,
      "mean_token_accuracy": 0.9725030601024628,
      "num_tokens": 3695042.0,
      "step": 1030
    },
    {
      "entropy": 2.2529222846031187,
      "epoch": 0.8029337965643698,
      "grad_norm": 3.5047266483306885,
      "learning_rate": 1.9123214932802465e-05,
      "loss": 0.11572520732879639,
      "mean_token_accuracy": 0.9776080340147019,
      "num_tokens": 3730250.0,
      "step": 1040
    },
    {
      "entropy": 2.1754327774047852,
      "epoch": 0.8106543138390272,
      "grad_norm": 3.506120204925537,
      "learning_rate": 1.9102633222654444e-05,
      "loss": 0.11624773740768432,
      "mean_token_accuracy": 0.9673858731985092,
      "num_tokens": 3767790.0,
      "step": 1050
    },
    {
      "entropy": 2.2163166791200637,
      "epoch": 0.8183748311136846,
      "grad_norm": 1.1102652549743652,
      "learning_rate": 1.9081824078542394e-05,
      "loss": 0.07758398056030273,
      "mean_token_accuracy": 0.9816681414842605,
      "num_tokens": 3803270.0,
      "step": 1060
    },
    {
      "entropy": 2.25885391831398,
      "epoch": 0.826095348388342,
      "grad_norm": 0.5661072134971619,
      "learning_rate": 1.9060788020393456e-05,
      "loss": 0.0615719735622406,
      "mean_token_accuracy": 0.9828296527266502,
      "num_tokens": 3834164.0,
      "step": 1070
    },
    {
      "entropy": 2.1803546130657194,
      "epoch": 0.8338158656629994,
      "grad_norm": 2.040249824523926,
      "learning_rate": 1.903952557380435e-05,
      "loss": 0.11473102569580078,
      "mean_token_accuracy": 0.9657544881105423,
      "num_tokens": 3871534.0,
      "step": 1080
    },
    {
      "entropy": 2.1811802864074705,
      "epoch": 0.8415363829376569,
      "grad_norm": 1.413288950920105,
      "learning_rate": 1.9018037270028213e-05,
      "loss": 0.07572046518325806,
      "mean_token_accuracy": 0.9782123982906341,
      "num_tokens": 3910887.0,
      "step": 1090
    },
    {
      "entropy": 2.23103586435318,
      "epoch": 0.8492569002123143,
      "grad_norm": 2.4706242084503174,
      "learning_rate": 1.8996323645961352e-05,
      "loss": 0.04151468873023987,
      "mean_token_accuracy": 0.9870179295539856,
      "num_tokens": 3943008.0,
      "step": 1100
    },
    {
      "entropy": 2.227116122841835,
      "epoch": 0.8569774174869716,
      "grad_norm": 2.6038529872894287,
      "learning_rate": 1.8974385244129805e-05,
      "loss": 0.0946826159954071,
      "mean_token_accuracy": 0.9796450510621071,
      "num_tokens": 3976668.0,
      "step": 1110
    },
    {
      "entropy": 2.208661425113678,
      "epoch": 0.864697934761629,
      "grad_norm": 5.780674457550049,
      "learning_rate": 1.8952222612675812e-05,
      "loss": 0.10974045991897582,
      "mean_token_accuracy": 0.9706292197108268,
      "num_tokens": 4012874.0,
      "step": 1120
    },
    {
      "entropy": 2.23996858894825,
      "epoch": 0.8724184520362864,
      "grad_norm": 1.2360124588012695,
      "learning_rate": 1.892983630534409e-05,
      "loss": 0.08555867075920105,
      "mean_token_accuracy": 0.9800435766577721,
      "num_tokens": 4044180.0,
      "step": 1130
    },
    {
      "entropy": 2.2110779762268065,
      "epoch": 0.8801389693109438,
      "grad_norm": 0.4163481891155243,
      "learning_rate": 1.890722688146802e-05,
      "loss": 0.07701975107192993,
      "mean_token_accuracy": 0.9780819907784462,
      "num_tokens": 4081151.0,
      "step": 1140
    },
    {
      "entropy": 2.226836398243904,
      "epoch": 0.8878594865856012,
      "grad_norm": 2.713832378387451,
      "learning_rate": 1.888439490595567e-05,
      "loss": 0.08010860681533813,
      "mean_token_accuracy": 0.9822189897298813,
      "num_tokens": 4116353.0,
      "step": 1150
    },
    {
      "entropy": 2.27404500246048,
      "epoch": 0.8955800038602586,
      "grad_norm": 0.15923215448856354,
      "learning_rate": 1.886134094927567e-05,
      "loss": 0.059002858400344846,
      "mean_token_accuracy": 0.9816316947340965,
      "num_tokens": 4148097.0,
      "step": 1160
    },
    {
      "entropy": 2.2260487526655197,
      "epoch": 0.903300521134916,
      "grad_norm": 3.0013537406921387,
      "learning_rate": 1.8838065587442956e-05,
      "loss": 0.08235616087913514,
      "mean_token_accuracy": 0.9715940162539483,
      "num_tokens": 4180489.0,
      "step": 1170
    },
    {
      "entropy": 2.124008280038834,
      "epoch": 0.9110210384095735,
      "grad_norm": 1.1767276525497437,
      "learning_rate": 1.8814569402004394e-05,
      "loss": 0.1106486201286316,
      "mean_token_accuracy": 0.9688979595899582,
      "num_tokens": 4220747.0,
      "step": 1180
    },
    {
      "entropy": 2.221278077363968,
      "epoch": 0.9187415556842309,
      "grad_norm": 1.546454906463623,
      "learning_rate": 1.8790852980024244e-05,
      "loss": 0.1707722783088684,
      "mean_token_accuracy": 0.9663840562105179,
      "num_tokens": 4259405.0,
      "step": 1190
    },
    {
      "entropy": 2.204825773835182,
      "epoch": 0.9264620729588883,
      "grad_norm": 2.0527737140655518,
      "learning_rate": 1.876691691406948e-05,
      "loss": 0.08684939742088318,
      "mean_token_accuracy": 0.9769035264849663,
      "num_tokens": 4292496.0,
      "step": 1200
    },
    {
      "entropy": 2.208803045749664,
      "epoch": 0.9341825902335457,
      "grad_norm": 4.592103481292725,
      "learning_rate": 1.8742761802194997e-05,
      "loss": 0.06823323369026184,
      "mean_token_accuracy": 0.9814566165208817,
      "num_tokens": 4325427.0,
      "step": 1210
    },
    {
      "entropy": 2.1777720004320145,
      "epoch": 0.941903107508203,
      "grad_norm": 3.140316963195801,
      "learning_rate": 1.871838824792867e-05,
      "loss": 0.16060715913772583,
      "mean_token_accuracy": 0.9660079538822174,
      "num_tokens": 4360480.0,
      "step": 1220
    },
    {
      "entropy": 2.2344942808151247,
      "epoch": 0.9496236247828604,
      "grad_norm": 4.9994001388549805,
      "learning_rate": 1.869379686025626e-05,
      "loss": 0.12226022481918335,
      "mean_token_accuracy": 0.9707063496112823,
      "num_tokens": 4392487.0,
      "step": 1230
    },
    {
      "entropy": 2.1433703660964967,
      "epoch": 0.9573441420575178,
      "grad_norm": 3.475524425506592,
      "learning_rate": 1.8668988253606212e-05,
      "loss": 0.12414617538452148,
      "mean_token_accuracy": 0.9741201639175415,
      "num_tokens": 4431701.0,
      "step": 1240
    },
    {
      "entropy": 2.190104368329048,
      "epoch": 0.9650646593321752,
      "grad_norm": 3.5840330123901367,
      "learning_rate": 1.8643963047834307e-05,
      "loss": 0.08782102465629578,
      "mean_token_accuracy": 0.9735233366489411,
      "num_tokens": 4467444.0,
      "step": 1250
    },
    {
      "entropy": 2.2168781042098997,
      "epoch": 0.9727851766068326,
      "grad_norm": 3.597426176071167,
      "learning_rate": 1.861872186820815e-05,
      "loss": 0.14496043920516968,
      "mean_token_accuracy": 0.9721528425812721,
      "num_tokens": 4496929.0,
      "step": 1260
    },
    {
      "entropy": 2.2057291209697723,
      "epoch": 0.9805056938814901,
      "grad_norm": 2.3206138610839844,
      "learning_rate": 1.8593265345391577e-05,
      "loss": 0.10822974443435669,
      "mean_token_accuracy": 0.9689827769994735,
      "num_tokens": 4531085.0,
      "step": 1270
    },
    {
      "entropy": 2.1946437865495683,
      "epoch": 0.9882262111561475,
      "grad_norm": 0.37915700674057007,
      "learning_rate": 1.8567594115428875e-05,
      "loss": 0.11415959596633911,
      "mean_token_accuracy": 0.9744870230555535,
      "num_tokens": 4569312.0,
      "step": 1280
    },
    {
      "entropy": 2.2317381650209427,
      "epoch": 0.9959467284308049,
      "grad_norm": 3.852874279022217,
      "learning_rate": 1.8541708819728902e-05,
      "loss": 0.06760208606719971,
      "mean_token_accuracy": 0.9818901315331459,
      "num_tokens": 4605253.0,
      "step": 1290
    },
    {
      "entropy": 2.1512465928051925,
      "epoch": 1.003088206909863,
      "grad_norm": 1.6707713603973389,
      "learning_rate": 1.8515610105049067e-05,
      "loss": 0.07693768739700317,
      "mean_token_accuracy": 0.9778838012669537,
      "num_tokens": 4643503.0,
      "step": 1300
    },
    {
      "entropy": 2.1755835592746733,
      "epoch": 1.0108087241845203,
      "grad_norm": 2.118502378463745,
      "learning_rate": 1.8489298623479147e-05,
      "loss": 0.07816660404205322,
      "mean_token_accuracy": 0.9813061475753784,
      "num_tokens": 4675662.0,
      "step": 1310
    },
    {
      "entropy": 2.1576342791318894,
      "epoch": 1.0185292414591778,
      "grad_norm": 5.421299934387207,
      "learning_rate": 1.846277503242502e-05,
      "loss": 0.09433794617652894,
      "mean_token_accuracy": 0.9725943058729172,
      "num_tokens": 4712787.0,
      "step": 1320
    },
    {
      "entropy": 2.1679455935955048,
      "epoch": 1.026249758733835,
      "grad_norm": 3.022178888320923,
      "learning_rate": 1.8436039994592224e-05,
      "loss": 0.08787925839424134,
      "mean_token_accuracy": 0.9729105263948441,
      "num_tokens": 4755527.0,
      "step": 1330
    },
    {
      "entropy": 2.146326667070389,
      "epoch": 1.0339702760084926,
      "grad_norm": 4.420295715332031,
      "learning_rate": 1.8409094177969408e-05,
      "loss": 0.09245921969413758,
      "mean_token_accuracy": 0.9803972214460372,
      "num_tokens": 4793692.0,
      "step": 1340
    },
    {
      "entropy": 2.125700297951698,
      "epoch": 1.04169079328315,
      "grad_norm": 0.7011611461639404,
      "learning_rate": 1.8381938255811626e-05,
      "loss": 0.04192873239517212,
      "mean_token_accuracy": 0.9881086781620979,
      "num_tokens": 4825139.0,
      "step": 1350
    },
    {
      "entropy": 2.265898513793945,
      "epoch": 1.0494113105578073,
      "grad_norm": 2.0361523628234863,
      "learning_rate": 1.8354572906623537e-05,
      "loss": 0.05326482057571411,
      "mean_token_accuracy": 0.9820634126663208,
      "num_tokens": 4860219.0,
      "step": 1360
    },
    {
      "entropy": 2.249096092581749,
      "epoch": 1.0571318278324648,
      "grad_norm": 2.0669565200805664,
      "learning_rate": 1.832699881414244e-05,
      "loss": 0.07189119458198548,
      "mean_token_accuracy": 0.9756469547748565,
      "num_tokens": 4895924.0,
      "step": 1370
    },
    {
      "entropy": 2.136953499913216,
      "epoch": 1.0648523451071221,
      "grad_norm": 3.7083239555358887,
      "learning_rate": 1.8299216667321192e-05,
      "loss": 0.0377763956785202,
      "mean_token_accuracy": 0.993205739557743,
      "num_tokens": 4932774.0,
      "step": 1380
    },
    {
      "entropy": 2.1725950062274935,
      "epoch": 1.0725728623817796,
      "grad_norm": 3.3318490982055664,
      "learning_rate": 1.827122716031099e-05,
      "loss": 0.03515214920043945,
      "mean_token_accuracy": 0.9864067286252975,
      "num_tokens": 4968312.0,
      "step": 1390
    },
    {
      "entropy": 2.171587583422661,
      "epoch": 1.080293379656437,
      "grad_norm": 1.3849668502807617,
      "learning_rate": 1.8243030992444042e-05,
      "loss": 0.033025556802749635,
      "mean_token_accuracy": 0.9898627385497093,
      "num_tokens": 5005630.0,
      "step": 1400
    },
    {
      "entropy": 2.123317489027977,
      "epoch": 1.0880138969310944,
      "grad_norm": 3.8918912410736084,
      "learning_rate": 1.821462886821607e-05,
      "loss": 0.06294504404067994,
      "mean_token_accuracy": 0.9815075367689132,
      "num_tokens": 5045736.0,
      "step": 1410
    },
    {
      "entropy": 2.1403660237789155,
      "epoch": 1.0957344142057517,
      "grad_norm": 3.85695219039917,
      "learning_rate": 1.8186021497268733e-05,
      "loss": 0.06816688179969788,
      "mean_token_accuracy": 0.9811367645859719,
      "num_tokens": 5087978.0,
      "step": 1420
    },
    {
      "entropy": 2.1448576182126997,
      "epoch": 1.1034549314804092,
      "grad_norm": 4.700116157531738,
      "learning_rate": 1.8157209594371873e-05,
      "loss": 0.06470519304275513,
      "mean_token_accuracy": 0.9794086501002311,
      "num_tokens": 5123611.0,
      "step": 1430
    },
    {
      "entropy": 2.246442288160324,
      "epoch": 1.1111754487550667,
      "grad_norm": 0.8788654208183289,
      "learning_rate": 1.8128193879405684e-05,
      "loss": 0.06699486970901489,
      "mean_token_accuracy": 0.9850307568907738,
      "num_tokens": 5156527.0,
      "step": 1440
    },
    {
      "entropy": 2.0466793209314345,
      "epoch": 1.118895966029724,
      "grad_norm": 2.6347248554229736,
      "learning_rate": 1.8098975077342696e-05,
      "loss": 0.05093771815299988,
      "mean_token_accuracy": 0.9866293117403984,
      "num_tokens": 5194973.0,
      "step": 1450
    },
    {
      "entropy": 2.175449812412262,
      "epoch": 1.1266164833043815,
      "grad_norm": 3.3113977909088135,
      "learning_rate": 1.806955391822968e-05,
      "loss": 0.07722722887992858,
      "mean_token_accuracy": 0.9812873587012291,
      "num_tokens": 5232238.0,
      "step": 1460
    },
    {
      "entropy": 2.1114370614290237,
      "epoch": 1.1343370005790387,
      "grad_norm": 6.0895538330078125,
      "learning_rate": 1.8039931137169398e-05,
      "loss": 0.10581094026565552,
      "mean_token_accuracy": 0.9739348217844963,
      "num_tokens": 5269462.0,
      "step": 1470
    },
    {
      "entropy": 2.057721531391144,
      "epoch": 1.1420575178536962,
      "grad_norm": 2.163757085800171,
      "learning_rate": 1.801010747430224e-05,
      "loss": 0.08237841129302978,
      "mean_token_accuracy": 0.9800081118941307,
      "num_tokens": 5312511.0,
      "step": 1480
    },
    {
      "entropy": 2.2099882423877717,
      "epoch": 1.1497780351283535,
      "grad_norm": 2.940668821334839,
      "learning_rate": 1.798008367478774e-05,
      "loss": 0.09443849325180054,
      "mean_token_accuracy": 0.9783109456300736,
      "num_tokens": 5347633.0,
      "step": 1490
    },
    {
      "entropy": 2.179814171791077,
      "epoch": 1.157498552403011,
      "grad_norm": 5.941656589508057,
      "learning_rate": 1.7949860488785935e-05,
      "loss": 0.11362334489822387,
      "mean_token_accuracy": 0.9784113153815269,
      "num_tokens": 5382716.0,
      "step": 1500
    },
    {
      "epoch": 1.157498552403011,
      "eval_entropy": 2.067308218003316,
      "eval_loss": 0.4690878093242645,
      "eval_mean_token_accuracy": 0.9239271352283633,
      "eval_num_tokens": 5382716.0,
      "eval_runtime": 987.4621,
      "eval_samples_per_second": 7.748,
      "eval_steps_per_second": 0.969,
      "step": 1500
    },
    {
      "entropy": 2.1574917674064635,
      "epoch": 1.1652190696776685,
      "grad_norm": 2.5012035369873047,
      "learning_rate": 1.7919438671438647e-05,
      "loss": 0.05635004043579102,
      "mean_token_accuracy": 0.9867865487933158,
      "num_tokens": 5427436.0,
      "step": 1510
    },
    {
      "entropy": 2.1504832983016966,
      "epoch": 1.1729395869523258,
      "grad_norm": 1.0461984872817993,
      "learning_rate": 1.78888189828506e-05,
      "loss": 0.06930508613586425,
      "mean_token_accuracy": 0.9815816029906272,
      "num_tokens": 5465521.0,
      "step": 1520
    },
    {
      "entropy": 2.201800489425659,
      "epoch": 1.180660104226983,
      "grad_norm": 3.5174059867858887,
      "learning_rate": 1.7858002188070428e-05,
      "loss": 0.06647626757621765,
      "mean_token_accuracy": 0.9884074360132218,
      "num_tokens": 5503228.0,
      "step": 1530
    },
    {
      "entropy": 2.1449740499258043,
      "epoch": 1.1883806215016406,
      "grad_norm": 2.1373608112335205,
      "learning_rate": 1.7826989057071576e-05,
      "loss": 0.08340185880661011,
      "mean_token_accuracy": 0.9833342641592026,
      "num_tokens": 5538689.0,
      "step": 1540
    },
    {
      "entropy": 2.1393818974494936,
      "epoch": 1.196101138776298,
      "grad_norm": 1.0357944965362549,
      "learning_rate": 1.779578036473304e-05,
      "loss": 0.06295200586318969,
      "mean_token_accuracy": 0.9850618243217468,
      "num_tokens": 5576456.0,
      "step": 1550
    },
    {
      "entropy": 2.1823107868433,
      "epoch": 1.2038216560509554,
      "grad_norm": 4.965859413146973,
      "learning_rate": 1.7764376890820014e-05,
      "loss": 0.09606855511665344,
      "mean_token_accuracy": 0.9736377954483032,
      "num_tokens": 5612953.0,
      "step": 1560
    },
    {
      "entropy": 2.1806869179010393,
      "epoch": 1.2115421733256129,
      "grad_norm": 0.8930599093437195,
      "learning_rate": 1.773277941996442e-05,
      "loss": 0.10148507356643677,
      "mean_token_accuracy": 0.9760855048894882,
      "num_tokens": 5649420.0,
      "step": 1570
    },
    {
      "entropy": 2.153961244225502,
      "epoch": 1.2192626906002701,
      "grad_norm": 5.579773426055908,
      "learning_rate": 1.770098874164529e-05,
      "loss": 0.0607684850692749,
      "mean_token_accuracy": 0.9863114863634109,
      "num_tokens": 5684144.0,
      "step": 1580
    },
    {
      "entropy": 2.1722445487976074,
      "epoch": 1.2269832078749277,
      "grad_norm": 6.017771244049072,
      "learning_rate": 1.7669005650169036e-05,
      "loss": 0.09015928506851197,
      "mean_token_accuracy": 0.9817434221506118,
      "num_tokens": 5720234.0,
      "step": 1590
    },
    {
      "entropy": 2.2104188591241836,
      "epoch": 1.234703725149585,
      "grad_norm": 3.4765541553497314,
      "learning_rate": 1.7636830944649628e-05,
      "loss": 0.06509234309196472,
      "mean_token_accuracy": 0.9830342516303062,
      "num_tokens": 5757637.0,
      "step": 1600
    },
    {
      "entropy": 2.198525631427765,
      "epoch": 1.2424242424242424,
      "grad_norm": 3.086829662322998,
      "learning_rate": 1.760446542898859e-05,
      "loss": 0.08481130599975586,
      "mean_token_accuracy": 0.9725503459572792,
      "num_tokens": 5792399.0,
      "step": 1610
    },
    {
      "entropy": 2.157542425394058,
      "epoch": 1.2501447596989,
      "grad_norm": 3.6862778663635254,
      "learning_rate": 1.757190991185495e-05,
      "loss": 0.02577916383743286,
      "mean_token_accuracy": 0.9937169313430786,
      "num_tokens": 5830379.0,
      "step": 1620
    },
    {
      "entropy": 2.2169549256563186,
      "epoch": 1.2578652769735572,
      "grad_norm": 6.9190192222595215,
      "learning_rate": 1.7539165206665018e-05,
      "loss": 0.11887587308883667,
      "mean_token_accuracy": 0.9760968968272209,
      "num_tokens": 5863578.0,
      "step": 1630
    },
    {
      "entropy": 2.173180091381073,
      "epoch": 1.2655857942482147,
      "grad_norm": 1.0279117822647095,
      "learning_rate": 1.750623213156206e-05,
      "loss": 0.03408839702606201,
      "mean_token_accuracy": 0.9902668550610543,
      "num_tokens": 5899214.0,
      "step": 1640
    },
    {
      "entropy": 2.155381426215172,
      "epoch": 1.273306311522872,
      "grad_norm": 0.8708764910697937,
      "learning_rate": 1.747311150939587e-05,
      "loss": 0.05587180852890015,
      "mean_token_accuracy": 0.9896471053361893,
      "num_tokens": 5935324.0,
      "step": 1650
    },
    {
      "entropy": 2.2335609972476957,
      "epoch": 1.2810268287975295,
      "grad_norm": 0.13011370599269867,
      "learning_rate": 1.743980416770219e-05,
      "loss": 0.035541835427284243,
      "mean_token_accuracy": 0.9919782534241677,
      "num_tokens": 5966942.0,
      "step": 1660
    },
    {
      "entropy": 2.1988327890634536,
      "epoch": 1.2887473460721868,
      "grad_norm": 0.10344131290912628,
      "learning_rate": 1.7406310938682048e-05,
      "loss": 0.046973693370819095,
      "mean_token_accuracy": 0.9925320342183113,
      "num_tokens": 6008187.0,
      "step": 1670
    },
    {
      "entropy": 2.23958740234375,
      "epoch": 1.2964678633468443,
      "grad_norm": 0.11602122336626053,
      "learning_rate": 1.7372632659180973e-05,
      "loss": 0.07079730033874512,
      "mean_token_accuracy": 0.9836244836449624,
      "num_tokens": 6041987.0,
      "step": 1680
    },
    {
      "entropy": 2.216384118795395,
      "epoch": 1.3041883806215018,
      "grad_norm": 2.2172930240631104,
      "learning_rate": 1.7338770170668067e-05,
      "loss": 0.03326311707496643,
      "mean_token_accuracy": 0.9883291959762573,
      "num_tokens": 6075157.0,
      "step": 1690
    },
    {
      "entropy": 2.1252345502376557,
      "epoch": 1.311908897896159,
      "grad_norm": 5.477461338043213,
      "learning_rate": 1.7304724319214984e-05,
      "loss": 0.06135511994361877,
      "mean_token_accuracy": 0.9845748677849769,
      "num_tokens": 6109370.0,
      "step": 1700
    },
    {
      "entropy": 2.1932441532611846,
      "epoch": 1.3196294151708163,
      "grad_norm": 1.0600680112838745,
      "learning_rate": 1.7270495955474804e-05,
      "loss": 0.06867148280143738,
      "mean_token_accuracy": 0.9842336073517799,
      "num_tokens": 6145950.0,
      "step": 1710
    },
    {
      "entropy": 2.228517660498619,
      "epoch": 1.3273499324454738,
      "grad_norm": 5.043092727661133,
      "learning_rate": 1.7236085934660767e-05,
      "loss": 0.06131213903427124,
      "mean_token_accuracy": 0.9860150918364525,
      "num_tokens": 6179698.0,
      "step": 1720
    },
    {
      "entropy": 2.147132176160812,
      "epoch": 1.3350704497201313,
      "grad_norm": 5.398331642150879,
      "learning_rate": 1.7201495116524904e-05,
      "loss": 0.06988582611083985,
      "mean_token_accuracy": 0.981314155459404,
      "num_tokens": 6217407.0,
      "step": 1730
    },
    {
      "entropy": 2.1961726933717727,
      "epoch": 1.3427909669947886,
      "grad_norm": 0.382192462682724,
      "learning_rate": 1.7166724365336567e-05,
      "loss": 0.04062625765800476,
      "mean_token_accuracy": 0.990032197535038,
      "num_tokens": 6254526.0,
      "step": 1740
    },
    {
      "entropy": 2.1768340557813644,
      "epoch": 1.3505114842694461,
      "grad_norm": 0.16774038970470428,
      "learning_rate": 1.7131774549860826e-05,
      "loss": 0.04462065696716309,
      "mean_token_accuracy": 0.9879740357398987,
      "num_tokens": 6291562.0,
      "step": 1750
    },
    {
      "entropy": 2.2020132303237916,
      "epoch": 1.3582320015441034,
      "grad_norm": 0.23544009029865265,
      "learning_rate": 1.7096646543336762e-05,
      "loss": 0.06595144867897033,
      "mean_token_accuracy": 0.9833995521068573,
      "num_tokens": 6325477.0,
      "step": 1760
    },
    {
      "entropy": 2.2006499111652373,
      "epoch": 1.365952518818761,
      "grad_norm": 4.985975742340088,
      "learning_rate": 1.7061341223455644e-05,
      "loss": 0.05818561315536499,
      "mean_token_accuracy": 0.9816210448741913,
      "num_tokens": 6362473.0,
      "step": 1770
    },
    {
      "entropy": 2.243510177731514,
      "epoch": 1.3736730360934182,
      "grad_norm": 1.5532886981964111,
      "learning_rate": 1.7025859472339026e-05,
      "loss": 0.04559597373008728,
      "mean_token_accuracy": 0.9878525719046592,
      "num_tokens": 6397995.0,
      "step": 1780
    },
    {
      "entropy": 2.2276618093252183,
      "epoch": 1.3813935533680757,
      "grad_norm": 5.254858493804932,
      "learning_rate": 1.699020217651667e-05,
      "loss": 0.07693036198616028,
      "mean_token_accuracy": 0.9713233038783073,
      "num_tokens": 6435700.0,
      "step": 1790
    },
    {
      "entropy": 2.214624211192131,
      "epoch": 1.3891140706427332,
      "grad_norm": 1.3537747859954834,
      "learning_rate": 1.695437022690441e-05,
      "loss": 0.0476935863494873,
      "mean_token_accuracy": 0.9899505183100701,
      "num_tokens": 6470710.0,
      "step": 1800
    },
    {
      "entropy": 2.2236395120620727,
      "epoch": 1.3968345879173905,
      "grad_norm": 2.256129741668701,
      "learning_rate": 1.691836451878191e-05,
      "loss": 0.057245922088623044,
      "mean_token_accuracy": 0.9842681303620339,
      "num_tokens": 6504059.0,
      "step": 1810
    },
    {
      "entropy": 2.194809466600418,
      "epoch": 1.4045551051920477,
      "grad_norm": 2.1219875812530518,
      "learning_rate": 1.688218595177027e-05,
      "loss": 0.095030015707016,
      "mean_token_accuracy": 0.977275763452053,
      "num_tokens": 6544244.0,
      "step": 1820
    },
    {
      "entropy": 2.189770597219467,
      "epoch": 1.4122756224667052,
      "grad_norm": 2.9641566276550293,
      "learning_rate": 1.6845835429809555e-05,
      "loss": 0.03089151084423065,
      "mean_token_accuracy": 0.9881447196006775,
      "num_tokens": 6575887.0,
      "step": 1830
    },
    {
      "entropy": 2.275960248708725,
      "epoch": 1.4199961397413627,
      "grad_norm": 3.321671724319458,
      "learning_rate": 1.6809313861136226e-05,
      "loss": 0.07831167578697204,
      "mean_token_accuracy": 0.9799184516072273,
      "num_tokens": 6604723.0,
      "step": 1840
    },
    {
      "entropy": 2.126830631494522,
      "epoch": 1.42771665701602,
      "grad_norm": 2.0569241046905518,
      "learning_rate": 1.6772622158260418e-05,
      "loss": 0.06082758903503418,
      "mean_token_accuracy": 0.9850939080119133,
      "num_tokens": 6642777.0,
      "step": 1850
    },
    {
      "entropy": 2.176947274804115,
      "epoch": 1.4354371742906775,
      "grad_norm": 0.6401554346084595,
      "learning_rate": 1.6735761237943166e-05,
      "loss": 0.06684135794639587,
      "mean_token_accuracy": 0.987571682035923,
      "num_tokens": 6682558.0,
      "step": 1860
    },
    {
      "entropy": 2.2031300246715544,
      "epoch": 1.4431576915653348,
      "grad_norm": 2.8854594230651855,
      "learning_rate": 1.6698732021173487e-05,
      "loss": 0.09527165293693543,
      "mean_token_accuracy": 0.9732189521193504,
      "num_tokens": 6717269.0,
      "step": 1870
    },
    {
      "entropy": 2.1681358337402346,
      "epoch": 1.4508782088399923,
      "grad_norm": 2.4058427810668945,
      "learning_rate": 1.666153543314537e-05,
      "loss": 0.048515334725379944,
      "mean_token_accuracy": 0.9848218783736229,
      "num_tokens": 6750783.0,
      "step": 1880
    },
    {
      "entropy": 2.209766998887062,
      "epoch": 1.4585987261146496,
      "grad_norm": 2.2834815979003906,
      "learning_rate": 1.6624172403234665e-05,
      "loss": 0.10613754987716675,
      "mean_token_accuracy": 0.9813075616955758,
      "num_tokens": 6784045.0,
      "step": 1890
    },
    {
      "entropy": 2.160476252436638,
      "epoch": 1.466319243389307,
      "grad_norm": 3.1155428886413574,
      "learning_rate": 1.6586643864975855e-05,
      "loss": 0.07376494407653808,
      "mean_token_accuracy": 0.9834692224860191,
      "num_tokens": 6824477.0,
      "step": 1900
    },
    {
      "entropy": 2.1634534388780593,
      "epoch": 1.4740397606639646,
      "grad_norm": 0.3001392185688019,
      "learning_rate": 1.6548950756038732e-05,
      "loss": 0.02613011598587036,
      "mean_token_accuracy": 0.9920083448290825,
      "num_tokens": 6864508.0,
      "step": 1910
    },
    {
      "entropy": 2.2088223576545714,
      "epoch": 1.4817602779386219,
      "grad_norm": 0.03745020180940628,
      "learning_rate": 1.6511094018204973e-05,
      "loss": 0.04845433235168457,
      "mean_token_accuracy": 0.9827080830931664,
      "num_tokens": 6902920.0,
      "step": 1920
    },
    {
      "entropy": 2.2410172671079636,
      "epoch": 1.4894807952132794,
      "grad_norm": 4.202921390533447,
      "learning_rate": 1.6473074597344608e-05,
      "loss": 0.07904167175292968,
      "mean_token_accuracy": 0.9802842602133751,
      "num_tokens": 6934196.0,
      "step": 1930
    },
    {
      "entropy": 2.2144098430871964,
      "epoch": 1.4972013124879366,
      "grad_norm": 3.0965535640716553,
      "learning_rate": 1.6434893443392388e-05,
      "loss": 0.06525328755378723,
      "mean_token_accuracy": 0.9790870904922485,
      "num_tokens": 6975432.0,
      "step": 1940
    },
    {
      "entropy": 2.210488173365593,
      "epoch": 1.5049218297625941,
      "grad_norm": 1.5190048217773438,
      "learning_rate": 1.6396551510324043e-05,
      "loss": 0.048443767428398135,
      "mean_token_accuracy": 0.9841597318649292,
      "num_tokens": 7014183.0,
      "step": 1950
    },
    {
      "entropy": 2.2083949983119964,
      "epoch": 1.5126423470372514,
      "grad_norm": 3.4382431507110596,
      "learning_rate": 1.6358049756132447e-05,
      "loss": 0.050505822896957396,
      "mean_token_accuracy": 0.98332499563694,
      "num_tokens": 7048113.0,
      "step": 1960
    },
    {
      "entropy": 2.1718403518199922,
      "epoch": 1.520362864311909,
      "grad_norm": 4.148841381072998,
      "learning_rate": 1.6319389142803705e-05,
      "loss": 0.08080021142959595,
      "mean_token_accuracy": 0.9764597907662391,
      "num_tokens": 7082017.0,
      "step": 1970
    },
    {
      "entropy": 2.202728974819183,
      "epoch": 1.5280833815865664,
      "grad_norm": 1.0588070154190063,
      "learning_rate": 1.6280570636293084e-05,
      "loss": 0.06684384942054748,
      "mean_token_accuracy": 0.9894752979278565,
      "num_tokens": 7117299.0,
      "step": 1980
    },
    {
      "entropy": 2.1958726704120637,
      "epoch": 1.5358038988612237,
      "grad_norm": 1.6634799242019653,
      "learning_rate": 1.6241595206500897e-05,
      "loss": 0.06764208674430847,
      "mean_token_accuracy": 0.984419621527195,
      "num_tokens": 7153932.0,
      "step": 1990
    },
    {
      "entropy": 2.207414472103119,
      "epoch": 1.543524416135881,
      "grad_norm": 3.5081145763397217,
      "learning_rate": 1.6202463827248265e-05,
      "loss": 0.06235606074333191,
      "mean_token_accuracy": 0.9850886449217796,
      "num_tokens": 7187474.0,
      "step": 2000
    },
    {
      "epoch": 1.543524416135881,
      "eval_entropy": 2.09292504755058,
      "eval_loss": 0.44610196352005005,
      "eval_mean_token_accuracy": 0.9292889271036584,
      "eval_num_tokens": 7187474.0,
      "eval_runtime": 987.9783,
      "eval_samples_per_second": 7.744,
      "eval_steps_per_second": 0.969,
      "step": 2000
    },
    {
      "entropy": 2.1876722425222397,
      "epoch": 1.5512449334105385,
      "grad_norm": 2.9447176456451416,
      "learning_rate": 1.6163177476252787e-05,
      "loss": 0.1231013298034668,
      "mean_token_accuracy": 0.9771831795573235,
      "num_tokens": 7224686.0,
      "step": 2010
    },
    {
      "entropy": 2.2381023466587067,
      "epoch": 1.558965450685196,
      "grad_norm": 4.822196006774902,
      "learning_rate": 1.612373713510411e-05,
      "loss": 0.03697749674320221,
      "mean_token_accuracy": 0.9885532259941101,
      "num_tokens": 7259356.0,
      "step": 2020
    },
    {
      "entropy": 2.2796195298433304,
      "epoch": 1.5666859679598533,
      "grad_norm": 1.0182671546936035,
      "learning_rate": 1.608414378923941e-05,
      "loss": 0.05077391862869263,
      "mean_token_accuracy": 0.988129609823227,
      "num_tokens": 7289152.0,
      "step": 2030
    },
    {
      "entropy": 2.2358617722988128,
      "epoch": 1.5744064852345108,
      "grad_norm": 2.3968863487243652,
      "learning_rate": 1.6044398427918754e-05,
      "loss": 0.06050935983657837,
      "mean_token_accuracy": 0.9812897890806198,
      "num_tokens": 7325050.0,
      "step": 2040
    },
    {
      "entropy": 2.174675610661507,
      "epoch": 1.5821270025091683,
      "grad_norm": 0.130416139960289,
      "learning_rate": 1.60045020442004e-05,
      "loss": 0.057877928018569946,
      "mean_token_accuracy": 0.9885829269886017,
      "num_tokens": 7361592.0,
      "step": 2050
    },
    {
      "entropy": 2.2293392330408097,
      "epoch": 1.5898475197838255,
      "grad_norm": 1.2082692384719849,
      "learning_rate": 1.5964455634915975e-05,
      "loss": 0.07877171635627747,
      "mean_token_accuracy": 0.9803994312882424,
      "num_tokens": 7394992.0,
      "step": 2060
    },
    {
      "entropy": 2.1638178408145903,
      "epoch": 1.5975680370584828,
      "grad_norm": 0.13079890608787537,
      "learning_rate": 1.5924260200645574e-05,
      "loss": 0.03450520932674408,
      "mean_token_accuracy": 0.993475878238678,
      "num_tokens": 7426832.0,
      "step": 2070
    },
    {
      "entropy": 2.31286364197731,
      "epoch": 1.6052885543331403,
      "grad_norm": 1.7269318103790283,
      "learning_rate": 1.5883916745692766e-05,
      "loss": 0.04786880314350128,
      "mean_token_accuracy": 0.9854305148124695,
      "num_tokens": 7459101.0,
      "step": 2080
    },
    {
      "entropy": 2.181331729888916,
      "epoch": 1.6130090716077978,
      "grad_norm": 2.854979991912842,
      "learning_rate": 1.584342627805947e-05,
      "loss": 0.04439874291419983,
      "mean_token_accuracy": 0.9889310583472252,
      "num_tokens": 7498208.0,
      "step": 2090
    },
    {
      "entropy": 2.201298424601555,
      "epoch": 1.620729588882455,
      "grad_norm": 3.3388164043426514,
      "learning_rate": 1.5802789809420812e-05,
      "loss": 0.057651227712631224,
      "mean_token_accuracy": 0.9866861075162887,
      "num_tokens": 7531350.0,
      "step": 2100
    },
    {
      "entropy": 2.1975525587797167,
      "epoch": 1.6284501061571124,
      "grad_norm": 1.4861758947372437,
      "learning_rate": 1.5762008355099824e-05,
      "loss": 0.04551325440406799,
      "mean_token_accuracy": 0.9858617588877678,
      "num_tokens": 7571182.0,
      "step": 2110
    },
    {
      "entropy": 2.2160713732242585,
      "epoch": 1.6361706234317699,
      "grad_norm": 2.163928985595703,
      "learning_rate": 1.5721082934042077e-05,
      "loss": 0.055477970838546754,
      "mean_token_accuracy": 0.9858786046504975,
      "num_tokens": 7602483.0,
      "step": 2120
    },
    {
      "entropy": 2.2408920913934707,
      "epoch": 1.6438911407064274,
      "grad_norm": 0.31985145807266235,
      "learning_rate": 1.5680014568790224e-05,
      "loss": 0.04159039855003357,
      "mean_token_accuracy": 0.9852373600006104,
      "num_tokens": 7637929.0,
      "step": 2130
    },
    {
      "entropy": 2.2200376480817794,
      "epoch": 1.6516116579810847,
      "grad_norm": 0.9541903734207153,
      "learning_rate": 1.5638804285458453e-05,
      "loss": 0.08421515226364136,
      "mean_token_accuracy": 0.980931767821312,
      "num_tokens": 7672999.0,
      "step": 2140
    },
    {
      "entropy": 2.203143599629402,
      "epoch": 1.6593321752557422,
      "grad_norm": 1.1459593772888184,
      "learning_rate": 1.5597453113706854e-05,
      "loss": 0.06661672592163086,
      "mean_token_accuracy": 0.9839463055133819,
      "num_tokens": 7709356.0,
      "step": 2150
    },
    {
      "entropy": 2.1805380463600157,
      "epoch": 1.6670526925303997,
      "grad_norm": 1.9122540950775146,
      "learning_rate": 1.555596208671568e-05,
      "loss": 0.07211873531341553,
      "mean_token_accuracy": 0.9839503139257431,
      "num_tokens": 7743958.0,
      "step": 2160
    },
    {
      "entropy": 2.2309546947479246,
      "epoch": 1.674773209805057,
      "grad_norm": 0.19467999041080475,
      "learning_rate": 1.5514332241159535e-05,
      "loss": 0.02178901433944702,
      "mean_token_accuracy": 0.9955803573131561,
      "num_tokens": 7781564.0,
      "step": 2170
    },
    {
      "entropy": 2.212314310669899,
      "epoch": 1.6824937270797142,
      "grad_norm": 7.478359222412109,
      "learning_rate": 1.5472564617181487e-05,
      "loss": 0.0869008719921112,
      "mean_token_accuracy": 0.9849289789795875,
      "num_tokens": 7819885.0,
      "step": 2180
    },
    {
      "entropy": 2.2491566449403764,
      "epoch": 1.6902142443543717,
      "grad_norm": 6.772735118865967,
      "learning_rate": 1.543066025836706e-05,
      "loss": 0.07113047838211059,
      "mean_token_accuracy": 0.982619047164917,
      "num_tokens": 7851412.0,
      "step": 2190
    },
    {
      "entropy": 2.213401734828949,
      "epoch": 1.6979347616290292,
      "grad_norm": 4.968042373657227,
      "learning_rate": 1.5388620211718174e-05,
      "loss": 0.04826553165912628,
      "mean_token_accuracy": 0.9854778051376343,
      "num_tokens": 7887905.0,
      "step": 2200
    },
    {
      "entropy": 2.296232557296753,
      "epoch": 1.7056552789036865,
      "grad_norm": 1.3846877813339233,
      "learning_rate": 1.5346445527626973e-05,
      "loss": 0.07143614888191223,
      "mean_token_accuracy": 0.9899365901947021,
      "num_tokens": 7917823.0,
      "step": 2210
    },
    {
      "entropy": 2.150168251991272,
      "epoch": 1.7133757961783438,
      "grad_norm": 0.08897116035223007,
      "learning_rate": 1.5304137259849596e-05,
      "loss": 0.03860937356948853,
      "mean_token_accuracy": 0.98958979845047,
      "num_tokens": 7957105.0,
      "step": 2220
    },
    {
      "entropy": 2.1939960032701493,
      "epoch": 1.7210963134530015,
      "grad_norm": 1.8421629667282104,
      "learning_rate": 1.5261696465479823e-05,
      "loss": 0.034920766949653625,
      "mean_token_accuracy": 0.9868922352790832,
      "num_tokens": 7990604.0,
      "step": 2230
    },
    {
      "entropy": 2.248343914747238,
      "epoch": 1.7288168307276588,
      "grad_norm": 0.1955411583185196,
      "learning_rate": 1.52191242049227e-05,
      "loss": 0.026687222719192504,
      "mean_token_accuracy": 0.9909184098243713,
      "num_tokens": 8025473.0,
      "step": 2240
    },
    {
      "entropy": 2.176126691699028,
      "epoch": 1.736537348002316,
      "grad_norm": 3.4646553993225098,
      "learning_rate": 1.5176421541868003e-05,
      "loss": 0.041455024480819704,
      "mean_token_accuracy": 0.987688897550106,
      "num_tokens": 8063276.0,
      "step": 2250
    },
    {
      "entropy": 2.2357524156570436,
      "epoch": 1.7442578652769736,
      "grad_norm": 2.85046124458313,
      "learning_rate": 1.5133589543263693e-05,
      "loss": 0.029646474123001098,
      "mean_token_accuracy": 0.9920761302113533,
      "num_tokens": 8095731.0,
      "step": 2260
    },
    {
      "entropy": 2.230524700880051,
      "epoch": 1.751978382551631,
      "grad_norm": 0.2627425789833069,
      "learning_rate": 1.5090629279289247e-05,
      "loss": 0.05477217435836792,
      "mean_token_accuracy": 0.9860235080122948,
      "num_tokens": 8134788.0,
      "step": 2270
    },
    {
      "entropy": 2.232674279808998,
      "epoch": 1.7596988998262884,
      "grad_norm": 5.077950477600098,
      "learning_rate": 1.5047541823328913e-05,
      "loss": 0.039683285355567935,
      "mean_token_accuracy": 0.9911200389266014,
      "num_tokens": 8168847.0,
      "step": 2280
    },
    {
      "entropy": 2.1635669827461244,
      "epoch": 1.7674194171009456,
      "grad_norm": 5.3769731521606445,
      "learning_rate": 1.5004328251944898e-05,
      "loss": 0.06116831302642822,
      "mean_token_accuracy": 0.9856468245387078,
      "num_tokens": 8203225.0,
      "step": 2290
    },
    {
      "entropy": 2.2570940554142,
      "epoch": 1.7751399343756031,
      "grad_norm": 1.6257141828536987,
      "learning_rate": 1.4960989644850461e-05,
      "loss": 0.06121355891227722,
      "mean_token_accuracy": 0.9865431234240531,
      "num_tokens": 8235676.0,
      "step": 2300
    },
    {
      "entropy": 2.148748704791069,
      "epoch": 1.7828604516502606,
      "grad_norm": 1.410240650177002,
      "learning_rate": 1.4917527084882962e-05,
      "loss": 0.060073697566986085,
      "mean_token_accuracy": 0.9826793894171715,
      "num_tokens": 8271787.0,
      "step": 2310
    },
    {
      "entropy": 2.259695219993591,
      "epoch": 1.790580968924918,
      "grad_norm": 0.7643431425094604,
      "learning_rate": 1.4873941657976758e-05,
      "loss": 0.045375460386276247,
      "mean_token_accuracy": 0.991543410718441,
      "num_tokens": 8305741.0,
      "step": 2320
    },
    {
      "entropy": 2.2634156942367554,
      "epoch": 1.7983014861995754,
      "grad_norm": 1.8315523862838745,
      "learning_rate": 1.483023445313613e-05,
      "loss": 0.06367403268814087,
      "mean_token_accuracy": 0.9873386204242707,
      "num_tokens": 8339552.0,
      "step": 2330
    },
    {
      "entropy": 2.2267695903778075,
      "epoch": 1.806022003474233,
      "grad_norm": 0.7167631387710571,
      "learning_rate": 1.4786406562408018e-05,
      "loss": 0.04319778382778168,
      "mean_token_accuracy": 0.9907146960496902,
      "num_tokens": 8373748.0,
      "step": 2340
    },
    {
      "entropy": 2.2331454545259475,
      "epoch": 1.8137425207488902,
      "grad_norm": 0.07811649143695831,
      "learning_rate": 1.4742459080854776e-05,
      "loss": 0.07732252478599548,
      "mean_token_accuracy": 0.9835749432444573,
      "num_tokens": 8405602.0,
      "step": 2350
    },
    {
      "entropy": 2.206145229935646,
      "epoch": 1.8214630380235475,
      "grad_norm": 2.4456584453582764,
      "learning_rate": 1.4698393106526794e-05,
      "loss": 0.04599563479423523,
      "mean_token_accuracy": 0.9887336835265159,
      "num_tokens": 8440623.0,
      "step": 2360
    },
    {
      "entropy": 2.185474079847336,
      "epoch": 1.829183555298205,
      "grad_norm": 2.7905468940734863,
      "learning_rate": 1.4654209740435058e-05,
      "loss": 0.10312718152999878,
      "mean_token_accuracy": 0.9773416921496392,
      "num_tokens": 8478808.0,
      "step": 2370
    },
    {
      "entropy": 2.1824015408754347,
      "epoch": 1.8369040725728625,
      "grad_norm": 3.6272330284118652,
      "learning_rate": 1.4609910086523656e-05,
      "loss": 0.03959389925003052,
      "mean_token_accuracy": 0.988817447423935,
      "num_tokens": 8518513.0,
      "step": 2380
    },
    {
      "entropy": 2.2420704424381257,
      "epoch": 1.8446245898475198,
      "grad_norm": 2.7875003814697266,
      "learning_rate": 1.4565495251642177e-05,
      "loss": 0.03577028214931488,
      "mean_token_accuracy": 0.9905984789133072,
      "num_tokens": 8551680.0,
      "step": 2390
    },
    {
      "entropy": 2.2019168078899383,
      "epoch": 1.852345107122177,
      "grad_norm": 1.5845329761505127,
      "learning_rate": 1.4520966345518076e-05,
      "loss": 0.032679545879364016,
      "mean_token_accuracy": 0.9905828937888146,
      "num_tokens": 8586760.0,
      "step": 2400
    },
    {
      "entropy": 2.1899552077054976,
      "epoch": 1.8600656243968345,
      "grad_norm": 2.2608611583709717,
      "learning_rate": 1.447632448072893e-05,
      "loss": 0.06101080775260925,
      "mean_token_accuracy": 0.9878913164138794,
      "num_tokens": 8623094.0,
      "step": 2410
    },
    {
      "entropy": 2.147029626369476,
      "epoch": 1.867786141671492,
      "grad_norm": 5.116804122924805,
      "learning_rate": 1.443157077267465e-05,
      "loss": 0.06001234650611877,
      "mean_token_accuracy": 0.9865107357501983,
      "num_tokens": 8661163.0,
      "step": 2420
    },
    {
      "entropy": 2.226870185136795,
      "epoch": 1.8755066589461493,
      "grad_norm": 0.31566572189331055,
      "learning_rate": 1.4386706339549608e-05,
      "loss": 0.04123932719230652,
      "mean_token_accuracy": 0.9899785459041596,
      "num_tokens": 8692034.0,
      "step": 2430
    },
    {
      "entropy": 2.1588668793439867,
      "epoch": 1.8832271762208068,
      "grad_norm": 0.9417765140533447,
      "learning_rate": 1.4341732302314695e-05,
      "loss": 0.06338693499565125,
      "mean_token_accuracy": 0.9791759848594666,
      "num_tokens": 8727679.0,
      "step": 2440
    },
    {
      "entropy": 2.2481164067983626,
      "epoch": 1.8909476934954643,
      "grad_norm": 2.853949785232544,
      "learning_rate": 1.4296649784669317e-05,
      "loss": 0.040822142362594606,
      "mean_token_accuracy": 0.9924886748194695,
      "num_tokens": 8765739.0,
      "step": 2450
    },
    {
      "entropy": 2.16390860080719,
      "epoch": 1.8986682107701216,
      "grad_norm": 0.7499203681945801,
      "learning_rate": 1.4251459913023326e-05,
      "loss": 0.03524640500545502,
      "mean_token_accuracy": 0.9882998391985893,
      "num_tokens": 8800881.0,
      "step": 2460
    },
    {
      "entropy": 2.2302677065134047,
      "epoch": 1.9063887280447789,
      "grad_norm": 2.2339160442352295,
      "learning_rate": 1.4206163816468861e-05,
      "loss": 0.05475660562515259,
      "mean_token_accuracy": 0.9859271243214607,
      "num_tokens": 8831211.0,
      "step": 2470
    },
    {
      "entropy": 2.176157131791115,
      "epoch": 1.9141092453194364,
      "grad_norm": 2.105004072189331,
      "learning_rate": 1.416076262675215e-05,
      "loss": 0.054158592224121095,
      "mean_token_accuracy": 0.9859476014971733,
      "num_tokens": 8870055.0,
      "step": 2480
    },
    {
      "entropy": 2.1534419268369676,
      "epoch": 1.9218297625940939,
      "grad_norm": 1.5954450368881226,
      "learning_rate": 1.4115257478245223e-05,
      "loss": 0.03219989836215973,
      "mean_token_accuracy": 0.9903153285384179,
      "num_tokens": 8906260.0,
      "step": 2490
    },
    {
      "entropy": 2.219716268777847,
      "epoch": 1.9295502798687512,
      "grad_norm": 6.508073329925537,
      "learning_rate": 1.4069649507917578e-05,
      "loss": 0.08166940808296204,
      "mean_token_accuracy": 0.9832880064845085,
      "num_tokens": 8938492.0,
      "step": 2500
    },
    {
      "epoch": 1.9295502798687512,
      "eval_entropy": 2.09807589759149,
      "eval_loss": 0.4539739787578583,
      "eval_mean_token_accuracy": 0.9317483626687539,
      "eval_num_tokens": 8938492.0,
      "eval_runtime": 988.2689,
      "eval_samples_per_second": 7.742,
      "eval_steps_per_second": 0.968,
      "step": 2500
    },
    {
      "entropy": 2.234724909067154,
      "epoch": 1.9372707971434087,
      "grad_norm": 1.7193647623062134,
      "learning_rate": 1.4023939855307764e-05,
      "loss": 0.04806544184684754,
      "mean_token_accuracy": 0.9910022959113121,
      "num_tokens": 8972850.0,
      "step": 2510
    },
    {
      "entropy": 2.198562887310982,
      "epoch": 1.9449913144180662,
      "grad_norm": 2.7927255630493164,
      "learning_rate": 1.397812966249492e-05,
      "loss": 0.06402395963668824,
      "mean_token_accuracy": 0.9840932011604309,
      "num_tokens": 9012403.0,
      "step": 2520
    },
    {
      "entropy": 2.2336697190999986,
      "epoch": 1.9527118316927234,
      "grad_norm": 5.629129886627197,
      "learning_rate": 1.3932220074070236e-05,
      "loss": 0.052779823541641235,
      "mean_token_accuracy": 0.9856812432408333,
      "num_tokens": 9049223.0,
      "step": 2530
    },
    {
      "entropy": 2.1691229969263075,
      "epoch": 1.9604323489673807,
      "grad_norm": 0.12791520357131958,
      "learning_rate": 1.3886212237108334e-05,
      "loss": 0.040309852361679076,
      "mean_token_accuracy": 0.9888273656368256,
      "num_tokens": 9086187.0,
      "step": 2540
    },
    {
      "entropy": 2.2491596221923826,
      "epoch": 1.9681528662420382,
      "grad_norm": 0.4085298478603363,
      "learning_rate": 1.384010730113865e-05,
      "loss": 0.02160833030939102,
      "mean_token_accuracy": 0.9909946650266648,
      "num_tokens": 9121957.0,
      "step": 2550
    },
    {
      "entropy": 2.1770587384700777,
      "epoch": 1.9758733835166957,
      "grad_norm": 0.27961045503616333,
      "learning_rate": 1.3793906418116674e-05,
      "loss": 0.07575550675392151,
      "mean_token_accuracy": 0.9829989477992058,
      "num_tokens": 9154073.0,
      "step": 2560
    },
    {
      "entropy": 2.262489286065102,
      "epoch": 1.983593900791353,
      "grad_norm": 0.30133992433547974,
      "learning_rate": 1.3747610742395185e-05,
      "loss": 0.049575185775756835,
      "mean_token_accuracy": 0.9877225264906884,
      "num_tokens": 9184603.0,
      "step": 2570
    },
    {
      "entropy": 2.157144469022751,
      "epoch": 1.9913144180660103,
      "grad_norm": 1.7762395143508911,
      "learning_rate": 1.3701221430695411e-05,
      "loss": 0.07316330075263977,
      "mean_token_accuracy": 0.9852767020463944,
      "num_tokens": 9220584.0,
      "step": 2580
    },
    {
      "entropy": 2.183892551064491,
      "epoch": 1.9990349353406678,
      "grad_norm": 0.4974443316459656,
      "learning_rate": 1.3654739642078115e-05,
      "loss": 0.030621829628944396,
      "mean_token_accuracy": 0.9909188866615295,
      "num_tokens": 9257900.0,
      "step": 2590
    },
    {
      "entropy": 2.1523505900357223,
      "epoch": 2.006176413819726,
      "grad_norm": 1.002280354499817,
      "learning_rate": 1.3608166537914653e-05,
      "loss": 0.03308936357498169,
      "mean_token_accuracy": 0.9905776703679884,
      "num_tokens": 9287566.0,
      "step": 2600
    },
    {
      "entropy": 2.1786624640226364,
      "epoch": 2.0138969310943833,
      "grad_norm": 1.7228094339370728,
      "learning_rate": 1.3561503281857935e-05,
      "loss": 0.024886465072631835,
      "mean_token_accuracy": 0.9927895620465279,
      "num_tokens": 9322676.0,
      "step": 2610
    },
    {
      "entropy": 2.1957814037799834,
      "epoch": 2.0216174483690406,
      "grad_norm": 1.4911882877349854,
      "learning_rate": 1.3514751039813373e-05,
      "loss": 0.02856048047542572,
      "mean_token_accuracy": 0.9912378385663032,
      "num_tokens": 9355944.0,
      "step": 2620
    },
    {
      "entropy": 2.2170669615268705,
      "epoch": 2.0293379656436983,
      "grad_norm": 2.101999282836914,
      "learning_rate": 1.3467910979909737e-05,
      "loss": 0.02958173453807831,
      "mean_token_accuracy": 0.9899841129779816,
      "num_tokens": 9388310.0,
      "step": 2630
    },
    {
      "entropy": 2.2836335778236387,
      "epoch": 2.0370584829183556,
      "grad_norm": 1.5629916191101074,
      "learning_rate": 1.3420984272469966e-05,
      "loss": 0.024695934355258943,
      "mean_token_accuracy": 0.9935079053044319,
      "num_tokens": 9421384.0,
      "step": 2640
    },
    {
      "entropy": 2.1128551036119463,
      "epoch": 2.044779000193013,
      "grad_norm": 0.04739844426512718,
      "learning_rate": 1.337397208998194e-05,
      "loss": 0.008467760682106019,
      "mean_token_accuracy": 0.9983387798070907,
      "num_tokens": 9458111.0,
      "step": 2650
    },
    {
      "entropy": 2.1693550407886506,
      "epoch": 2.05249951746767,
      "grad_norm": 0.9237629771232605,
      "learning_rate": 1.3326875607069167e-05,
      "loss": 0.030665600299835206,
      "mean_token_accuracy": 0.9918258935213089,
      "num_tokens": 9496587.0,
      "step": 2660
    },
    {
      "entropy": 2.15002136528492,
      "epoch": 2.060220034742328,
      "grad_norm": 4.3934807777404785,
      "learning_rate": 1.3279696000461453e-05,
      "loss": 0.06771356463432313,
      "mean_token_accuracy": 0.9814038887619972,
      "num_tokens": 9533430.0,
      "step": 2670
    },
    {
      "entropy": 2.2125822573900225,
      "epoch": 2.067940552016985,
      "grad_norm": 0.43496638536453247,
      "learning_rate": 1.3232434448965492e-05,
      "loss": 0.020783789455890656,
      "mean_token_accuracy": 0.9934862002730369,
      "num_tokens": 9569158.0,
      "step": 2680
    },
    {
      "entropy": 2.1923176765441896,
      "epoch": 2.0756610692916424,
      "grad_norm": 0.4592268466949463,
      "learning_rate": 1.318509213343541e-05,
      "loss": 0.02032013237476349,
      "mean_token_accuracy": 0.996187499165535,
      "num_tokens": 9605532.0,
      "step": 2690
    },
    {
      "entropy": 2.2375913679599764,
      "epoch": 2.0833815865663,
      "grad_norm": 3.984872579574585,
      "learning_rate": 1.3137670236743258e-05,
      "loss": 0.05714105367660523,
      "mean_token_accuracy": 0.984250420331955,
      "num_tokens": 9637666.0,
      "step": 2700
    },
    {
      "entropy": 2.2166593253612517,
      "epoch": 2.0911021038409574,
      "grad_norm": 1.2222201824188232,
      "learning_rate": 1.3090169943749475e-05,
      "loss": 0.01319773644208908,
      "mean_token_accuracy": 0.9959829047322273,
      "num_tokens": 9671041.0,
      "step": 2710
    },
    {
      "entropy": 2.207410234212875,
      "epoch": 2.0988226211156147,
      "grad_norm": 2.778895139694214,
      "learning_rate": 1.3042592441273265e-05,
      "loss": 0.0495807558298111,
      "mean_token_accuracy": 0.9841275036334991,
      "num_tokens": 9709462.0,
      "step": 2720
    },
    {
      "entropy": 2.240642195940018,
      "epoch": 2.106543138390272,
      "grad_norm": 0.054825589060783386,
      "learning_rate": 1.2994938918062956e-05,
      "loss": 0.02018197178840637,
      "mean_token_accuracy": 0.9940782889723778,
      "num_tokens": 9743175.0,
      "step": 2730
    },
    {
      "entropy": 2.1023289799690246,
      "epoch": 2.1142636556649297,
      "grad_norm": 1.3066190481185913,
      "learning_rate": 1.2947210564766285e-05,
      "loss": 0.029668596386909486,
      "mean_token_accuracy": 0.9919564932584762,
      "num_tokens": 9780470.0,
      "step": 2740
    },
    {
      "entropy": 2.194093734025955,
      "epoch": 2.121984172939587,
      "grad_norm": 1.0239670276641846,
      "learning_rate": 1.289940857390066e-05,
      "loss": 0.031669008731842044,
      "mean_token_accuracy": 0.9926641970872879,
      "num_tokens": 9817526.0,
      "step": 2750
    },
    {
      "entropy": 2.2564922988414766,
      "epoch": 2.1297046902142442,
      "grad_norm": 1.27794349193573,
      "learning_rate": 1.2851534139823363e-05,
      "loss": 0.019582782685756684,
      "mean_token_accuracy": 0.995488366484642,
      "num_tokens": 9852050.0,
      "step": 2760
    },
    {
      "entropy": 2.185793325304985,
      "epoch": 2.137425207488902,
      "grad_norm": 0.7653716206550598,
      "learning_rate": 1.280358845870171e-05,
      "loss": 0.02054794430732727,
      "mean_token_accuracy": 0.9923271596431732,
      "num_tokens": 9892115.0,
      "step": 2770
    },
    {
      "entropy": 2.2001701056957246,
      "epoch": 2.1451457247635592,
      "grad_norm": 0.39544498920440674,
      "learning_rate": 1.2755572728483155e-05,
      "loss": 0.023257075250148772,
      "mean_token_accuracy": 0.9951051697134972,
      "num_tokens": 9931964.0,
      "step": 2780
    },
    {
      "entropy": 2.240946352481842,
      "epoch": 2.1528662420382165,
      "grad_norm": 0.26407626271247864,
      "learning_rate": 1.2707488148865363e-05,
      "loss": 0.013235661387443542,
      "mean_token_accuracy": 0.9952550038695336,
      "num_tokens": 9966565.0,
      "step": 2790
    },
    {
      "entropy": 2.2113855719566344,
      "epoch": 2.160586759312874,
      "grad_norm": 3.220576047897339,
      "learning_rate": 1.265933592126625e-05,
      "loss": 0.036707454919815065,
      "mean_token_accuracy": 0.991096867620945,
      "num_tokens": 9998961.0,
      "step": 2800
    },
    {
      "entropy": 2.190717473626137,
      "epoch": 2.1683072765875315,
      "grad_norm": 2.822577476501465,
      "learning_rate": 1.2611117248793936e-05,
      "loss": 0.03832524418830872,
      "mean_token_accuracy": 0.9897032305598259,
      "num_tokens": 10033957.0,
      "step": 2810
    },
    {
      "entropy": 2.14252789914608,
      "epoch": 2.176027793862189,
      "grad_norm": 1.2037594318389893,
      "learning_rate": 1.2562833336216706e-05,
      "loss": 0.030560973286628722,
      "mean_token_accuracy": 0.9927359268069267,
      "num_tokens": 10067990.0,
      "step": 2820
    },
    {
      "entropy": 2.169364756345749,
      "epoch": 2.183748311136846,
      "grad_norm": 0.36512044072151184,
      "learning_rate": 1.2514485389932904e-05,
      "loss": 0.024552282691001893,
      "mean_token_accuracy": 0.9914642333984375,
      "num_tokens": 10107375.0,
      "step": 2830
    },
    {
      "entropy": 2.215904402732849,
      "epoch": 2.1914688284115034,
      "grad_norm": 1.6125507354736328,
      "learning_rate": 1.2466074617940789e-05,
      "loss": 0.04732250571250916,
      "mean_token_accuracy": 0.9921780049800872,
      "num_tokens": 10141732.0,
      "step": 2840
    },
    {
      "entropy": 2.2249450385570526,
      "epoch": 2.199189345686161,
      "grad_norm": 1.7302135229110718,
      "learning_rate": 1.2417602229808351e-05,
      "loss": 0.017911496758460998,
      "mean_token_accuracy": 0.995235213637352,
      "num_tokens": 10178996.0,
      "step": 2850
    },
    {
      "entropy": 2.2548143804073333,
      "epoch": 2.2069098629608184,
      "grad_norm": 0.04718457907438278,
      "learning_rate": 1.236906943664309e-05,
      "loss": 0.047148984670639035,
      "mean_token_accuracy": 0.992178562283516,
      "num_tokens": 10214640.0,
      "step": 2860
    },
    {
      "entropy": 2.2417348712682723,
      "epoch": 2.2146303802354756,
      "grad_norm": 0.028584061190485954,
      "learning_rate": 1.2320477451061764e-05,
      "loss": 0.019759944081306456,
      "mean_token_accuracy": 0.9941247820854187,
      "num_tokens": 10251016.0,
      "step": 2870
    },
    {
      "entropy": 2.2414058953523637,
      "epoch": 2.2223508975101334,
      "grad_norm": 0.14690132439136505,
      "learning_rate": 1.227182748716007e-05,
      "loss": 0.02443784922361374,
      "mean_token_accuracy": 0.995506352186203,
      "num_tokens": 10292140.0,
      "step": 2880
    },
    {
      "entropy": 2.161432334780693,
      "epoch": 2.2300714147847907,
      "grad_norm": 1.1667348146438599,
      "learning_rate": 1.2223120760482334e-05,
      "loss": 0.059571588039398195,
      "mean_token_accuracy": 0.9856904774904252,
      "num_tokens": 10332661.0,
      "step": 2890
    },
    {
      "entropy": 2.201232245564461,
      "epoch": 2.237791932059448,
      "grad_norm": 0.03672393038868904,
      "learning_rate": 1.217435848799113e-05,
      "loss": 0.014521844685077667,
      "mean_token_accuracy": 0.9963999673724174,
      "num_tokens": 10369258.0,
      "step": 2900
    },
    {
      "entropy": 2.2226219087839127,
      "epoch": 2.245512449334105,
      "grad_norm": 0.8930593729019165,
      "learning_rate": 1.2125541888036865e-05,
      "loss": 0.022989581525325774,
      "mean_token_accuracy": 0.9940146535634995,
      "num_tokens": 10401439.0,
      "step": 2910
    },
    {
      "entropy": 2.2541381388902666,
      "epoch": 2.253232966608763,
      "grad_norm": 1.267367959022522,
      "learning_rate": 1.2076672180327354e-05,
      "loss": 0.04482119679450989,
      "mean_token_accuracy": 0.9880507871508598,
      "num_tokens": 10436132.0,
      "step": 2920
    },
    {
      "entropy": 2.254253861308098,
      "epoch": 2.26095348388342,
      "grad_norm": 2.6567845344543457,
      "learning_rate": 1.2027750585897325e-05,
      "loss": 0.03449807465076447,
      "mean_token_accuracy": 0.9864874675869941,
      "num_tokens": 10470241.0,
      "step": 2930
    },
    {
      "entropy": 2.1793852925300596,
      "epoch": 2.2686740011580775,
      "grad_norm": 0.10617174953222275,
      "learning_rate": 1.1978778327077932e-05,
      "loss": 0.01429850161075592,
      "mean_token_accuracy": 0.9945758923888206,
      "num_tokens": 10501127.0,
      "step": 2940
    },
    {
      "entropy": 2.1325310707092284,
      "epoch": 2.2763945184327348,
      "grad_norm": 1.1946356296539307,
      "learning_rate": 1.1929756627466207e-05,
      "loss": 0.05270189046859741,
      "mean_token_accuracy": 0.9898444160819053,
      "num_tokens": 10541230.0,
      "step": 2950
    },
    {
      "entropy": 2.1456130146980286,
      "epoch": 2.2841150357073925,
      "grad_norm": 0.13995744287967682,
      "learning_rate": 1.1880686711894476e-05,
      "loss": 0.02509075403213501,
      "mean_token_accuracy": 0.9976190477609634,
      "num_tokens": 10573896.0,
      "step": 2960
    },
    {
      "entropy": 2.2185772597789764,
      "epoch": 2.2918355529820498,
      "grad_norm": 0.05365545675158501,
      "learning_rate": 1.1831569806399773e-05,
      "loss": 0.06382806301116943,
      "mean_token_accuracy": 0.9850220456719398,
      "num_tokens": 10608069.0,
      "step": 2970
    },
    {
      "entropy": 2.2425498723983766,
      "epoch": 2.299556070256707,
      "grad_norm": 0.649382472038269,
      "learning_rate": 1.178240713819319e-05,
      "loss": 0.01412668377161026,
      "mean_token_accuracy": 0.994785113632679,
      "num_tokens": 10646749.0,
      "step": 2980
    },
    {
      "entropy": 2.191881862282753,
      "epoch": 2.3072765875313648,
      "grad_norm": 5.6314496994018555,
      "learning_rate": 1.173319993562924e-05,
      "loss": 0.04553419649600983,
      "mean_token_accuracy": 0.9872561976313591,
      "num_tokens": 10688044.0,
      "step": 2990
    },
    {
      "entropy": 2.15625539124012,
      "epoch": 2.314997104806022,
      "grad_norm": 1.1487791538238525,
      "learning_rate": 1.1683949428175137e-05,
      "loss": 0.03611789345741272,
      "mean_token_accuracy": 0.9944777116179466,
      "num_tokens": 10723778.0,
      "step": 3000
    },
    {
      "epoch": 2.314997104806022,
      "eval_entropy": 2.0888531302708193,
      "eval_loss": 0.46160778403282166,
      "eval_mean_token_accuracy": 0.9338101559671862,
      "eval_num_tokens": 10723778.0,
      "eval_runtime": 988.0337,
      "eval_samples_per_second": 7.744,
      "eval_steps_per_second": 0.969,
      "step": 3000
    },
    {
      "entropy": 2.217298111319542,
      "epoch": 2.3227176220806793,
      "grad_norm": 0.7061363458633423,
      "learning_rate": 1.1634656846380088e-05,
      "loss": 0.02380364537239075,
      "mean_token_accuracy": 0.9964717343449593,
      "num_tokens": 10761784.0,
      "step": 3010
    },
    {
      "entropy": 2.2075847864151,
      "epoch": 2.330438139355337,
      "grad_norm": 5.256769180297852,
      "learning_rate": 1.1585323421844563e-05,
      "loss": 0.020036756992340088,
      "mean_token_accuracy": 0.9942559063434601,
      "num_tokens": 10799548.0,
      "step": 3020
    },
    {
      "entropy": 2.2639325708150864,
      "epoch": 2.3381586566299943,
      "grad_norm": 1.2079206705093384,
      "learning_rate": 1.1535950387189491e-05,
      "loss": 0.026534247398376464,
      "mean_token_accuracy": 0.9935628846287727,
      "num_tokens": 10834454.0,
      "step": 3030
    },
    {
      "entropy": 2.2246354073286057,
      "epoch": 2.3458791739046516,
      "grad_norm": 4.4090070724487305,
      "learning_rate": 1.1486538976025494e-05,
      "loss": 0.03687326312065124,
      "mean_token_accuracy": 0.9897530093789101,
      "num_tokens": 10870495.0,
      "step": 3040
    },
    {
      "entropy": 2.264278048276901,
      "epoch": 2.353599691179309,
      "grad_norm": 1.5192763805389404,
      "learning_rate": 1.1437090422922048e-05,
      "loss": 0.03048618733882904,
      "mean_token_accuracy": 0.9923379093408584,
      "num_tokens": 10905172.0,
      "step": 3050
    },
    {
      "entropy": 2.2067640751600264,
      "epoch": 2.361320208453966,
      "grad_norm": 3.1806492805480957,
      "learning_rate": 1.1387605963376634e-05,
      "loss": 0.026115420460700988,
      "mean_token_accuracy": 0.993264289200306,
      "num_tokens": 10944797.0,
      "step": 3060
    },
    {
      "entropy": 2.191044235229492,
      "epoch": 2.369040725728624,
      "grad_norm": 0.313163161277771,
      "learning_rate": 1.1338086833783879e-05,
      "loss": 0.022575947642326354,
      "mean_token_accuracy": 0.9924851134419441,
      "num_tokens": 10981103.0,
      "step": 3070
    },
    {
      "entropy": 2.2448153853416444,
      "epoch": 2.376761243003281,
      "grad_norm": 0.07588159292936325,
      "learning_rate": 1.1288534271404654e-05,
      "loss": 0.02497773468494415,
      "mean_token_accuracy": 0.9956817016005516,
      "num_tokens": 11012602.0,
      "step": 3080
    },
    {
      "entropy": 2.1922047078609466,
      "epoch": 2.3844817602779385,
      "grad_norm": 4.014681339263916,
      "learning_rate": 1.1238949514335171e-05,
      "loss": 0.04230898320674896,
      "mean_token_accuracy": 0.9890714272856712,
      "num_tokens": 11049617.0,
      "step": 3090
    },
    {
      "entropy": 2.2160174280405043,
      "epoch": 2.392202277552596,
      "grad_norm": 0.2612691819667816,
      "learning_rate": 1.1189333801476043e-05,
      "loss": 0.027991071343421936,
      "mean_token_accuracy": 0.9910157874226571,
      "num_tokens": 11083167.0,
      "step": 3100
    },
    {
      "entropy": 2.1471309036016466,
      "epoch": 2.3999227948272535,
      "grad_norm": 0.20606683194637299,
      "learning_rate": 1.1139688372501322e-05,
      "loss": 0.008393615484237671,
      "mean_token_accuracy": 0.9982551485300064,
      "num_tokens": 11120707.0,
      "step": 3110
    },
    {
      "entropy": 2.2330644577741623,
      "epoch": 2.4076433121019107,
      "grad_norm": 3.892418622970581,
      "learning_rate": 1.109001446782754e-05,
      "loss": 0.0618414044380188,
      "mean_token_accuracy": 0.9849733024835586,
      "num_tokens": 11152504.0,
      "step": 3120
    },
    {
      "entropy": 2.1754608184099196,
      "epoch": 2.4153638293765685,
      "grad_norm": 2.0742340087890625,
      "learning_rate": 1.1040313328582704e-05,
      "loss": 0.04327288269996643,
      "mean_token_accuracy": 0.9893003240227699,
      "num_tokens": 11186267.0,
      "step": 3130
    },
    {
      "entropy": 2.1819115340709687,
      "epoch": 2.4230843466512257,
      "grad_norm": 0.9018492102622986,
      "learning_rate": 1.0990586196575294e-05,
      "loss": 0.02030205577611923,
      "mean_token_accuracy": 0.9951718047261238,
      "num_tokens": 11221627.0,
      "step": 3140
    },
    {
      "entropy": 2.2373082458972933,
      "epoch": 2.430804863925883,
      "grad_norm": 0.7242811918258667,
      "learning_rate": 1.0940834314263233e-05,
      "loss": 0.030833426117897033,
      "mean_token_accuracy": 0.9951138079166413,
      "num_tokens": 11257677.0,
      "step": 3150
    },
    {
      "entropy": 2.2836509525775908,
      "epoch": 2.4385253812005403,
      "grad_norm": 0.30014318227767944,
      "learning_rate": 1.089105892472284e-05,
      "loss": 0.02001524865627289,
      "mean_token_accuracy": 0.9914347141981125,
      "num_tokens": 11292106.0,
      "step": 3160
    },
    {
      "entropy": 2.139942041039467,
      "epoch": 2.446245898475198,
      "grad_norm": 0.8197279572486877,
      "learning_rate": 1.0841261271617773e-05,
      "loss": 0.06748307943344116,
      "mean_token_accuracy": 0.988334609568119,
      "num_tokens": 11327418.0,
      "step": 3170
    },
    {
      "entropy": 2.213589632511139,
      "epoch": 2.4539664157498553,
      "grad_norm": 0.3989240229129791,
      "learning_rate": 1.0791442599167962e-05,
      "loss": 0.023147647082805634,
      "mean_token_accuracy": 0.9967875018715858,
      "num_tokens": 11361726.0,
      "step": 3180
    },
    {
      "entropy": 2.256254014372826,
      "epoch": 2.4616869330245126,
      "grad_norm": 1.1078554391860962,
      "learning_rate": 1.0741604152118515e-05,
      "loss": 0.029124677181243896,
      "mean_token_accuracy": 0.9931992501020431,
      "num_tokens": 11398586.0,
      "step": 3190
    },
    {
      "entropy": 2.165381821990013,
      "epoch": 2.46940745029917,
      "grad_norm": 1.1475622653961182,
      "learning_rate": 1.0691747175708611e-05,
      "loss": 0.028723520040512086,
      "mean_token_accuracy": 0.990369887650013,
      "num_tokens": 11432096.0,
      "step": 3200
    },
    {
      "entropy": 2.222157561779022,
      "epoch": 2.4771279675738276,
      "grad_norm": 0.06832505017518997,
      "learning_rate": 1.0641872915640406e-05,
      "loss": 0.04921720027923584,
      "mean_token_accuracy": 0.9877250343561172,
      "num_tokens": 11470784.0,
      "step": 3210
    },
    {
      "entropy": 2.2426993161439897,
      "epoch": 2.484848484848485,
      "grad_norm": 4.916550159454346,
      "learning_rate": 1.0591982618047889e-05,
      "loss": 0.06242421865463257,
      "mean_token_accuracy": 0.9849206641316414,
      "num_tokens": 11502948.0,
      "step": 3220
    },
    {
      "entropy": 2.197218120098114,
      "epoch": 2.492569002123142,
      "grad_norm": 2.124483346939087,
      "learning_rate": 1.0542077529465758e-05,
      "loss": 0.016325852274894713,
      "mean_token_accuracy": 0.9951209232211113,
      "num_tokens": 11539468.0,
      "step": 3230
    },
    {
      "entropy": 2.2902637124061584,
      "epoch": 2.5002895193978,
      "grad_norm": 0.08580861240625381,
      "learning_rate": 1.049215889679827e-05,
      "loss": 0.015471380949020386,
      "mean_token_accuracy": 0.9910388752818108,
      "num_tokens": 11569402.0,
      "step": 3240
    },
    {
      "entropy": 2.218478021025658,
      "epoch": 2.508010036672457,
      "grad_norm": 3.2291533946990967,
      "learning_rate": 1.0442227967288087e-05,
      "loss": 0.023704257607460023,
      "mean_token_accuracy": 0.9939404681324959,
      "num_tokens": 11605925.0,
      "step": 3250
    },
    {
      "entropy": 2.1201782315969466,
      "epoch": 2.5157305539471144,
      "grad_norm": 2.169576406478882,
      "learning_rate": 1.0392285988485118e-05,
      "loss": 0.032733604311943054,
      "mean_token_accuracy": 0.9920709192752838,
      "num_tokens": 11643802.0,
      "step": 3260
    },
    {
      "entropy": 2.2365442246198652,
      "epoch": 2.5234510712217717,
      "grad_norm": 0.022659538313746452,
      "learning_rate": 1.0342334208215343e-05,
      "loss": 0.019719691574573518,
      "mean_token_accuracy": 0.9956497728824616,
      "num_tokens": 11678574.0,
      "step": 3270
    },
    {
      "entropy": 2.258416032791138,
      "epoch": 2.5311715884964294,
      "grad_norm": 1.1894478797912598,
      "learning_rate": 1.0292373874549632e-05,
      "loss": 0.027323293685913085,
      "mean_token_accuracy": 0.9939205542206764,
      "num_tokens": 11709479.0,
      "step": 3280
    },
    {
      "entropy": 2.2412182718515394,
      "epoch": 2.5388921057710867,
      "grad_norm": 0.7658277153968811,
      "learning_rate": 1.0242406235772573e-05,
      "loss": 0.036611801385879515,
      "mean_token_accuracy": 0.9948377966880798,
      "num_tokens": 11747308.0,
      "step": 3290
    },
    {
      "entropy": 2.206353172659874,
      "epoch": 2.546612623045744,
      "grad_norm": 4.169549942016602,
      "learning_rate": 1.0192432540351269e-05,
      "loss": 0.023987260460853577,
      "mean_token_accuracy": 0.992442001402378,
      "num_tokens": 11779775.0,
      "step": 3300
    },
    {
      "entropy": 2.174012392759323,
      "epoch": 2.5543331403204013,
      "grad_norm": 1.4817591905593872,
      "learning_rate": 1.0142454036904154e-05,
      "loss": 0.014469130337238312,
      "mean_token_accuracy": 0.997749999165535,
      "num_tokens": 11813976.0,
      "step": 3310
    },
    {
      "entropy": 2.1796777188777923,
      "epoch": 2.562053657595059,
      "grad_norm": 0.08421381562948227,
      "learning_rate": 1.0092471974169802e-05,
      "loss": 0.02185429334640503,
      "mean_token_accuracy": 0.9927170544862747,
      "num_tokens": 11849438.0,
      "step": 3320
    },
    {
      "entropy": 2.2282682776451113,
      "epoch": 2.5697741748697163,
      "grad_norm": 0.30909499526023865,
      "learning_rate": 1.0042487600975699e-05,
      "loss": 0.03119277358055115,
      "mean_token_accuracy": 0.9929424881935119,
      "num_tokens": 11883486.0,
      "step": 3330
    },
    {
      "entropy": 2.2394611954689028,
      "epoch": 2.5774946921443735,
      "grad_norm": 0.10575172305107117,
      "learning_rate": 9.992502166207077e-06,
      "loss": 0.009619663655757903,
      "mean_token_accuracy": 0.9981060609221458,
      "num_tokens": 11920020.0,
      "step": 3340
    },
    {
      "entropy": 2.201454371213913,
      "epoch": 2.5852152094190313,
      "grad_norm": 5.886312484741211,
      "learning_rate": 9.94251691877568e-06,
      "loss": 0.025818082690238952,
      "mean_token_accuracy": 0.9924077332019806,
      "num_tokens": 11955168.0,
      "step": 3350
    },
    {
      "entropy": 2.1938006192445756,
      "epoch": 2.5929357266936885,
      "grad_norm": 0.12290452420711517,
      "learning_rate": 9.892533107588577e-06,
      "loss": 0.01246689334511757,
      "mean_token_accuracy": 0.9964157700538635,
      "num_tokens": 11994023.0,
      "step": 3360
    },
    {
      "entropy": 2.192418694496155,
      "epoch": 2.600656243968346,
      "grad_norm": 0.488265722990036,
      "learning_rate": 9.842551981516952e-06,
      "loss": 0.018000373244285585,
      "mean_token_accuracy": 0.9964497670531273,
      "num_tokens": 12028712.0,
      "step": 3370
    },
    {
      "entropy": 2.268085116147995,
      "epoch": 2.6083767612430035,
      "grad_norm": 0.14518892765045166,
      "learning_rate": 9.792574789364893e-06,
      "loss": 0.020170669257640838,
      "mean_token_accuracy": 0.9963870957493782,
      "num_tokens": 12059526.0,
      "step": 3380
    },
    {
      "entropy": 2.2167040795087813,
      "epoch": 2.616097278517661,
      "grad_norm": 1.522740125656128,
      "learning_rate": 9.742602779838209e-06,
      "loss": 0.016399385035037996,
      "mean_token_accuracy": 0.9955128505825996,
      "num_tokens": 12097085.0,
      "step": 3390
    },
    {
      "entropy": 2.295898801088333,
      "epoch": 2.623817795792318,
      "grad_norm": 4.225417613983154,
      "learning_rate": 9.69263720151321e-06,
      "loss": 0.029479607939720154,
      "mean_token_accuracy": 0.9938596487045288,
      "num_tokens": 12129924.0,
      "step": 3400
    },
    {
      "entropy": 2.280335706472397,
      "epoch": 2.6315383130669754,
      "grad_norm": 1.20139741897583,
      "learning_rate": 9.64267930280552e-06,
      "loss": 0.050666403770446775,
      "mean_token_accuracy": 0.9906869053840637,
      "num_tokens": 12161521.0,
      "step": 3410
    },
    {
      "entropy": 2.224092537164688,
      "epoch": 2.6392588303416327,
      "grad_norm": 1.7901499271392822,
      "learning_rate": 9.592730331938882e-06,
      "loss": 0.03357548415660858,
      "mean_token_accuracy": 0.9925102308392525,
      "num_tokens": 12197782.0,
      "step": 3420
    },
    {
      "entropy": 2.2111442536115646,
      "epoch": 2.6469793476162904,
      "grad_norm": 1.5995413064956665,
      "learning_rate": 9.542791536913985e-06,
      "loss": 0.027673259377479553,
      "mean_token_accuracy": 0.9946150794625283,
      "num_tokens": 12233888.0,
      "step": 3430
    },
    {
      "entropy": 2.2556965053081512,
      "epoch": 2.6546998648909477,
      "grad_norm": 3.171156644821167,
      "learning_rate": 9.492864165477246e-06,
      "loss": 0.019532108306884767,
      "mean_token_accuracy": 0.9935697302222252,
      "num_tokens": 12267840.0,
      "step": 3440
    },
    {
      "entropy": 2.1733025908470154,
      "epoch": 2.662420382165605,
      "grad_norm": 0.10952026396989822,
      "learning_rate": 9.442949465089679e-06,
      "loss": 0.04825982451438904,
      "mean_token_accuracy": 0.9895769953727722,
      "num_tokens": 12303601.0,
      "step": 3450
    },
    {
      "entropy": 2.239233008027077,
      "epoch": 2.6701408994402627,
      "grad_norm": 0.07487324625253677,
      "learning_rate": 9.393048682895698e-06,
      "loss": 0.051196861267089847,
      "mean_token_accuracy": 0.983622993528843,
      "num_tokens": 12340150.0,
      "step": 3460
    },
    {
      "entropy": 2.2360308557748794,
      "epoch": 2.67786141671492,
      "grad_norm": 2.012169122695923,
      "learning_rate": 9.34316306569196e-06,
      "loss": 0.01992444396018982,
      "mean_token_accuracy": 0.9927385210990906,
      "num_tokens": 12380777.0,
      "step": 3470
    },
    {
      "entropy": 2.235995998978615,
      "epoch": 2.6855819339895772,
      "grad_norm": 0.08367231488227844,
      "learning_rate": 9.293293859896224e-06,
      "loss": 0.018683603405952452,
      "mean_token_accuracy": 0.9944416150450707,
      "num_tokens": 12424408.0,
      "step": 3480
    },
    {
      "entropy": 2.202517831325531,
      "epoch": 2.693302451264235,
      "grad_norm": 0.12537157535552979,
      "learning_rate": 9.243442311516194e-06,
      "loss": 0.035655727982521056,
      "mean_token_accuracy": 0.9909148395061493,
      "num_tokens": 12459199.0,
      "step": 3490
    },
    {
      "entropy": 2.216241243481636,
      "epoch": 2.7010229685388922,
      "grad_norm": 2.079157590866089,
      "learning_rate": 9.193609666118403e-06,
      "loss": 0.02550240755081177,
      "mean_token_accuracy": 0.9924755334854126,
      "num_tokens": 12492382.0,
      "step": 3500
    },
    {
      "epoch": 2.7010229685388922,
      "eval_entropy": 2.08757205492775,
      "eval_loss": 0.46482330560684204,
      "eval_mean_token_accuracy": 0.9347644107593389,
      "eval_num_tokens": 12492382.0,
      "eval_runtime": 987.4485,
      "eval_samples_per_second": 7.748,
      "eval_steps_per_second": 0.969,
      "step": 3500
    },
    {
      "entropy": 2.1927063912153244,
      "epoch": 2.7087434858135495,
      "grad_norm": 4.612570285797119,
      "learning_rate": 9.143797168797074e-06,
      "loss": 0.020228767395019533,
      "mean_token_accuracy": 0.9914082810282707,
      "num_tokens": 12530445.0,
      "step": 3510
    },
    {
      "entropy": 2.2404502600431444,
      "epoch": 2.716464003088207,
      "grad_norm": 2.8573319911956787,
      "learning_rate": 9.094006064143023e-06,
      "loss": 0.031410235166549685,
      "mean_token_accuracy": 0.9914770245552063,
      "num_tokens": 12564152.0,
      "step": 3520
    },
    {
      "entropy": 2.1382099032402038,
      "epoch": 2.724184520362864,
      "grad_norm": 3.369036912918091,
      "learning_rate": 9.04423759621257e-06,
      "loss": 0.048682579398155214,
      "mean_token_accuracy": 0.9871285542845726,
      "num_tokens": 12601866.0,
      "step": 3530
    },
    {
      "entropy": 2.2182486951351166,
      "epoch": 2.731905037637522,
      "grad_norm": 0.04350515827536583,
      "learning_rate": 8.994493008496426e-06,
      "loss": 0.037138843536376955,
      "mean_token_accuracy": 0.9921187847852707,
      "num_tokens": 12636504.0,
      "step": 3540
    },
    {
      "entropy": 2.238505133986473,
      "epoch": 2.739625554912179,
      "grad_norm": 0.463403582572937,
      "learning_rate": 8.944773543888657e-06,
      "loss": 0.017538127303123475,
      "mean_token_accuracy": 0.9944233506917953,
      "num_tokens": 12671889.0,
      "step": 3550
    },
    {
      "entropy": 2.2769761711359022,
      "epoch": 2.7473460721868364,
      "grad_norm": 1.676416277885437,
      "learning_rate": 8.895080444655606e-06,
      "loss": 0.028448593616485596,
      "mean_token_accuracy": 0.9939671754837036,
      "num_tokens": 12707437.0,
      "step": 3560
    },
    {
      "entropy": 2.2299616605043413,
      "epoch": 2.755066589461494,
      "grad_norm": 0.42011213302612305,
      "learning_rate": 8.845414952404878e-06,
      "loss": 0.022680895030498506,
      "mean_token_accuracy": 0.9940501064062118,
      "num_tokens": 12741280.0,
      "step": 3570
    },
    {
      "entropy": 2.1800275564193727,
      "epoch": 2.7627871067361514,
      "grad_norm": 3.0085341930389404,
      "learning_rate": 8.79577830805428e-06,
      "loss": 0.04312008917331696,
      "mean_token_accuracy": 0.9904634907841683,
      "num_tokens": 12777520.0,
      "step": 3580
    },
    {
      "entropy": 2.121428969502449,
      "epoch": 2.7705076240108086,
      "grad_norm": 0.18500223755836487,
      "learning_rate": 8.746171751800862e-06,
      "loss": 0.010257864743471146,
      "mean_token_accuracy": 0.9983589738607407,
      "num_tokens": 12811089.0,
      "step": 3590
    },
    {
      "entropy": 2.1989814430475234,
      "epoch": 2.7782281412854664,
      "grad_norm": 0.542945146560669,
      "learning_rate": 8.696596523089897e-06,
      "loss": 0.00803714692592621,
      "mean_token_accuracy": 0.998145604133606,
      "num_tokens": 12846203.0,
      "step": 3600
    },
    {
      "entropy": 2.231951543688774,
      "epoch": 2.7859486585601236,
      "grad_norm": 0.2957513928413391,
      "learning_rate": 8.647053860583921e-06,
      "loss": 0.020568850636482238,
      "mean_token_accuracy": 0.9942832916975022,
      "num_tokens": 12881146.0,
      "step": 3610
    },
    {
      "entropy": 2.2153997004032133,
      "epoch": 2.793669175834781,
      "grad_norm": 4.239973545074463,
      "learning_rate": 8.597545002131792e-06,
      "loss": 0.01559099406003952,
      "mean_token_accuracy": 0.99518121778965,
      "num_tokens": 12920233.0,
      "step": 3620
    },
    {
      "entropy": 2.257417145371437,
      "epoch": 2.801389693109438,
      "grad_norm": 3.4071872234344482,
      "learning_rate": 8.548071184737756e-06,
      "loss": 0.020855729281902314,
      "mean_token_accuracy": 0.9936988726258278,
      "num_tokens": 12952742.0,
      "step": 3630
    },
    {
      "entropy": 2.1990304440259933,
      "epoch": 2.8091102103840955,
      "grad_norm": 6.179783344268799,
      "learning_rate": 8.498633644530535e-06,
      "loss": 0.03551109731197357,
      "mean_token_accuracy": 0.9895797207951545,
      "num_tokens": 12987526.0,
      "step": 3640
    },
    {
      "entropy": 2.2407626271247865,
      "epoch": 2.816830727658753,
      "grad_norm": 2.062673568725586,
      "learning_rate": 8.449233616732453e-06,
      "loss": 0.026538869738578795,
      "mean_token_accuracy": 0.9921577915549278,
      "num_tokens": 13025654.0,
      "step": 3650
    },
    {
      "entropy": 2.234692195057869,
      "epoch": 2.8245512449334105,
      "grad_norm": 2.2553117275238037,
      "learning_rate": 8.399872335628564e-06,
      "loss": 0.017977765202522276,
      "mean_token_accuracy": 0.9920303016901016,
      "num_tokens": 13062506.0,
      "step": 3660
    },
    {
      "entropy": 2.2229584991931914,
      "epoch": 2.8322717622080678,
      "grad_norm": 2.068716287612915,
      "learning_rate": 8.35055103453582e-06,
      "loss": 0.026064303517341614,
      "mean_token_accuracy": 0.9947251975536346,
      "num_tokens": 13097457.0,
      "step": 3670
    },
    {
      "entropy": 2.1777986973524093,
      "epoch": 2.8399922794827255,
      "grad_norm": 0.49508559703826904,
      "learning_rate": 8.301270945772245e-06,
      "loss": 0.01860121488571167,
      "mean_token_accuracy": 0.9952414557337761,
      "num_tokens": 13132851.0,
      "step": 3680
    },
    {
      "entropy": 2.233269226551056,
      "epoch": 2.8477127967573828,
      "grad_norm": 0.03142388164997101,
      "learning_rate": 8.252033300626154e-06,
      "loss": 0.03146423697471619,
      "mean_token_accuracy": 0.9943786114454269,
      "num_tokens": 13165995.0,
      "step": 3690
    },
    {
      "entropy": 2.2296706050634385,
      "epoch": 2.85543331403204,
      "grad_norm": 0.21639715135097504,
      "learning_rate": 8.202839329325396e-06,
      "loss": 0.019184540212154388,
      "mean_token_accuracy": 0.9943373009562493,
      "num_tokens": 13203076.0,
      "step": 3700
    },
    {
      "entropy": 2.196987953782082,
      "epoch": 2.8631538313066978,
      "grad_norm": 0.4990762770175934,
      "learning_rate": 8.153690261006586e-06,
      "loss": 0.025900739431381225,
      "mean_token_accuracy": 0.9961180582642555,
      "num_tokens": 13242240.0,
      "step": 3710
    },
    {
      "entropy": 2.2462848991155626,
      "epoch": 2.870874348581355,
      "grad_norm": 2.7010302543640137,
      "learning_rate": 8.104587323684435e-06,
      "loss": 0.03213064074516296,
      "mean_token_accuracy": 0.9937118411064148,
      "num_tokens": 13278586.0,
      "step": 3720
    },
    {
      "entropy": 2.2904939532279966,
      "epoch": 2.8785948658560123,
      "grad_norm": 0.5824764370918274,
      "learning_rate": 8.05553174422104e-06,
      "loss": 0.04053815901279449,
      "mean_token_accuracy": 0.9919193282723426,
      "num_tokens": 13313746.0,
      "step": 3730
    },
    {
      "entropy": 2.2101560860872267,
      "epoch": 2.8863153831306696,
      "grad_norm": 0.2210099995136261,
      "learning_rate": 8.006524748295234e-06,
      "loss": 0.01130208745598793,
      "mean_token_accuracy": 0.998863635957241,
      "num_tokens": 13354478.0,
      "step": 3740
    },
    {
      "entropy": 2.2668700993061064,
      "epoch": 2.8940359004053273,
      "grad_norm": 0.10139039158821106,
      "learning_rate": 7.95756756037197e-06,
      "loss": 0.030500581860542296,
      "mean_token_accuracy": 0.992310605943203,
      "num_tokens": 13386730.0,
      "step": 3750
    },
    {
      "entropy": 2.3063734799623488,
      "epoch": 2.9017564176799846,
      "grad_norm": 1.5737172365188599,
      "learning_rate": 7.908661403671721e-06,
      "loss": 0.04344950020313263,
      "mean_token_accuracy": 0.9915585055947304,
      "num_tokens": 13418333.0,
      "step": 3760
    },
    {
      "entropy": 2.1441037923097612,
      "epoch": 2.909476934954642,
      "grad_norm": 0.03719672933220863,
      "learning_rate": 7.859807500139922e-06,
      "loss": 0.02446196973323822,
      "mean_token_accuracy": 0.9906012058258057,
      "num_tokens": 13457112.0,
      "step": 3770
    },
    {
      "entropy": 2.2902497231960295,
      "epoch": 2.917197452229299,
      "grad_norm": 4.060549736022949,
      "learning_rate": 7.811007070416425e-06,
      "loss": 0.03251109719276428,
      "mean_token_accuracy": 0.987868283689022,
      "num_tokens": 13492524.0,
      "step": 3780
    },
    {
      "entropy": 2.2127520233392715,
      "epoch": 2.924917969503957,
      "grad_norm": 0.7151080369949341,
      "learning_rate": 7.762261333805026e-06,
      "loss": 0.029196158051490784,
      "mean_token_accuracy": 0.9946857452392578,
      "num_tokens": 13528544.0,
      "step": 3790
    },
    {
      "entropy": 2.162434884905815,
      "epoch": 2.932638486778614,
      "grad_norm": 0.06961432099342346,
      "learning_rate": 7.713571508242981e-06,
      "loss": 0.026906999945640563,
      "mean_token_accuracy": 0.9946787610650063,
      "num_tokens": 13567498.0,
      "step": 3800
    },
    {
      "entropy": 2.246299198269844,
      "epoch": 2.9403590040532714,
      "grad_norm": 3.7265074253082275,
      "learning_rate": 7.664938810270566e-06,
      "loss": 0.029045966267585755,
      "mean_token_accuracy": 0.9936078891158104,
      "num_tokens": 13602625.0,
      "step": 3810
    },
    {
      "entropy": 2.2388866275548933,
      "epoch": 2.948079521327929,
      "grad_norm": 0.24144315719604492,
      "learning_rate": 7.616364455000714e-06,
      "loss": 0.014337979257106781,
      "mean_token_accuracy": 0.9960026681423187,
      "num_tokens": 13641679.0,
      "step": 3820
    },
    {
      "entropy": 2.188111361861229,
      "epoch": 2.9558000386025864,
      "grad_norm": 6.046823501586914,
      "learning_rate": 7.567849656088628e-06,
      "loss": 0.03005032241344452,
      "mean_token_accuracy": 0.9918044999241828,
      "num_tokens": 13683014.0,
      "step": 3830
    },
    {
      "entropy": 2.204224619269371,
      "epoch": 2.9635205558772437,
      "grad_norm": 5.8294453620910645,
      "learning_rate": 7.519395625701462e-06,
      "loss": 0.038737067580223085,
      "mean_token_accuracy": 0.9885923460125923,
      "num_tokens": 13718071.0,
      "step": 3840
    },
    {
      "entropy": 2.2484711676836016,
      "epoch": 2.9712410731519014,
      "grad_norm": 9.640524864196777,
      "learning_rate": 7.471003574488041e-06,
      "loss": 0.028097692131996154,
      "mean_token_accuracy": 0.9957665964961052,
      "num_tokens": 13751151.0,
      "step": 3850
    },
    {
      "entropy": 2.2508881419897078,
      "epoch": 2.9789615904265587,
      "grad_norm": 0.10017109662294388,
      "learning_rate": 7.422674711548605e-06,
      "loss": 0.01202818676829338,
      "mean_token_accuracy": 0.9953813180327415,
      "num_tokens": 13791068.0,
      "step": 3860
    },
    {
      "entropy": 2.2337160974740984,
      "epoch": 2.986682107701216,
      "grad_norm": 0.043931517750024796,
      "learning_rate": 7.374410244404613e-06,
      "loss": 0.009058024734258652,
      "mean_token_accuracy": 0.9986373543739319,
      "num_tokens": 13829239.0,
      "step": 3870
    },
    {
      "entropy": 2.1862533658742906,
      "epoch": 2.9944026249758733,
      "grad_norm": 0.072396419942379,
      "learning_rate": 7.326211378968542e-06,
      "loss": 0.012882831692695617,
      "mean_token_accuracy": 0.9978046581149101,
      "num_tokens": 13868022.0,
      "step": 3880
    },
    {
      "entropy": 2.1555983053671346,
      "epoch": 3.0015441034549313,
      "grad_norm": 0.04370846599340439,
      "learning_rate": 7.278079319513793e-06,
      "loss": 0.03105989098548889,
      "mean_token_accuracy": 0.9914975681820432,
      "num_tokens": 13899213.0,
      "step": 3890
    },
    {
      "entropy": 2.2709316343069075,
      "epoch": 3.009264620729589,
      "grad_norm": 0.431661456823349,
      "learning_rate": 7.230015268644588e-06,
      "loss": 0.015339791774749756,
      "mean_token_accuracy": 0.9981125354766845,
      "num_tokens": 13935501.0,
      "step": 3900
    },
    {
      "entropy": 2.1740885227918625,
      "epoch": 3.0169851380042463,
      "grad_norm": 0.2746732234954834,
      "learning_rate": 7.182020427265902e-06,
      "loss": 0.0041766297072172165,
      "mean_token_accuracy": 1.0,
      "num_tokens": 13974859.0,
      "step": 3910
    },
    {
      "entropy": 2.2257029473781587,
      "epoch": 3.0247056552789036,
      "grad_norm": 2.3247573375701904,
      "learning_rate": 7.134095994553489e-06,
      "loss": 0.01182505115866661,
      "mean_token_accuracy": 0.9958305478096008,
      "num_tokens": 14011902.0,
      "step": 3920
    },
    {
      "entropy": 2.198028963804245,
      "epoch": 3.0324261725535613,
      "grad_norm": 0.07981577515602112,
      "learning_rate": 7.086243167923905e-06,
      "loss": 0.022176089882850646,
      "mean_token_accuracy": 0.9963392838835716,
      "num_tokens": 14046566.0,
      "step": 3930
    },
    {
      "entropy": 2.242706462740898,
      "epoch": 3.0401466898282186,
      "grad_norm": 0.02186143584549427,
      "learning_rate": 7.038463143004591e-06,
      "loss": 0.012292072176933289,
      "mean_token_accuracy": 0.9961706340312958,
      "num_tokens": 14079313.0,
      "step": 3940
    },
    {
      "entropy": 2.241307318210602,
      "epoch": 3.047867207102876,
      "grad_norm": 0.0767030119895935,
      "learning_rate": 6.990757113603992e-06,
      "loss": 0.007230133563280105,
      "mean_token_accuracy": 0.9980000004172325,
      "num_tokens": 14117695.0,
      "step": 3950
    },
    {
      "entropy": 2.210197499394417,
      "epoch": 3.055587724377533,
      "grad_norm": 0.13448522984981537,
      "learning_rate": 6.943126271681747e-06,
      "loss": 0.018068452179431916,
      "mean_token_accuracy": 0.9969620048999787,
      "num_tokens": 14153987.0,
      "step": 3960
    },
    {
      "entropy": 2.253513675928116,
      "epoch": 3.063308241652191,
      "grad_norm": 3.672414541244507,
      "learning_rate": 6.895571807318897e-06,
      "loss": 0.023470789194107056,
      "mean_token_accuracy": 0.9966314926743507,
      "num_tokens": 14185621.0,
      "step": 3970
    },
    {
      "entropy": 2.2444525837898253,
      "epoch": 3.071028758926848,
      "grad_norm": 0.12743113934993744,
      "learning_rate": 6.848094908688139e-06,
      "loss": 0.012810063362121583,
      "mean_token_accuracy": 0.9977913543581962,
      "num_tokens": 14219216.0,
      "step": 3980
    },
    {
      "entropy": 2.2262933045625686,
      "epoch": 3.0787492762015054,
      "grad_norm": 3.4074060916900635,
      "learning_rate": 6.8006967620241615e-06,
      "loss": 0.025683704018592834,
      "mean_token_accuracy": 0.9924832716584205,
      "num_tokens": 14257742.0,
      "step": 3990
    },
    {
      "entropy": 2.261911395192146,
      "epoch": 3.086469793476163,
      "grad_norm": 0.056527040898799896,
      "learning_rate": 6.753378551593992e-06,
      "loss": 0.003825421258807182,
      "mean_token_accuracy": 0.9986842110753059,
      "num_tokens": 14291203.0,
      "step": 4000
    },
    {
      "epoch": 3.086469793476163,
      "eval_entropy": 2.1145772489260732,
      "eval_loss": 0.4855208396911621,
      "eval_mean_token_accuracy": 0.9354811434088082,
      "eval_num_tokens": 14291203.0,
      "eval_runtime": 989.2063,
      "eval_samples_per_second": 7.734,
      "eval_steps_per_second": 0.967,
      "step": 4000
    },
    {
      "entropy": 2.2831456065177917,
      "epoch": 3.0941903107508204,
      "grad_norm": 0.058800771832466125,
      "learning_rate": 6.70614145966741e-06,
      "loss": 0.013551701605319978,
      "mean_token_accuracy": 0.9963570073246956,
      "num_tokens": 14322403.0,
      "step": 4010
    },
    {
      "entropy": 2.2009116262197495,
      "epoch": 3.1019108280254777,
      "grad_norm": 0.09004119783639908,
      "learning_rate": 6.658986666487406e-06,
      "loss": 0.012571512162685395,
      "mean_token_accuracy": 0.9967614203691483,
      "num_tokens": 14363415.0,
      "step": 4020
    },
    {
      "entropy": 2.183881738781929,
      "epoch": 3.109631345300135,
      "grad_norm": 0.030173663049936295,
      "learning_rate": 6.611915350240694e-06,
      "loss": 0.006261659413576126,
      "mean_token_accuracy": 0.9986278191208839,
      "num_tokens": 14402620.0,
      "step": 4030
    },
    {
      "entropy": 2.2086421698331833,
      "epoch": 3.1173518625747927,
      "grad_norm": 0.12183760106563568,
      "learning_rate": 6.56492868702828e-06,
      "loss": 0.013231892883777619,
      "mean_token_accuracy": 0.995246222615242,
      "num_tokens": 14436242.0,
      "step": 4040
    },
    {
      "entropy": 2.1381252884864805,
      "epoch": 3.12507237984945,
      "grad_norm": 2.105003833770752,
      "learning_rate": 6.51802785083606e-06,
      "loss": 0.025765135884284973,
      "mean_token_accuracy": 0.9932129815220833,
      "num_tokens": 14476123.0,
      "step": 4050
    },
    {
      "entropy": 2.22107275724411,
      "epoch": 3.1327928971241072,
      "grad_norm": 0.8158308863639832,
      "learning_rate": 6.471214013505507e-06,
      "loss": 0.01576409637928009,
      "mean_token_accuracy": 0.9953061327338218,
      "num_tokens": 14507443.0,
      "step": 4060
    },
    {
      "entropy": 2.2192520588636397,
      "epoch": 3.1405134143987645,
      "grad_norm": 4.943661212921143,
      "learning_rate": 6.42448834470438e-06,
      "loss": 0.013656246662139892,
      "mean_token_accuracy": 0.9962772980332375,
      "num_tokens": 14540908.0,
      "step": 4070
    },
    {
      "entropy": 2.2019663602113724,
      "epoch": 3.1482339316734222,
      "grad_norm": 0.011873590759932995,
      "learning_rate": 6.3778520118974975e-06,
      "loss": 0.004965296760201454,
      "mean_token_accuracy": 0.99921875,
      "num_tokens": 14577340.0,
      "step": 4080
    },
    {
      "entropy": 2.2791995286941527,
      "epoch": 3.1559544489480795,
      "grad_norm": 0.24833594262599945,
      "learning_rate": 6.331306180317577e-06,
      "loss": 0.016860713064670563,
      "mean_token_accuracy": 0.9945571333169937,
      "num_tokens": 14612020.0,
      "step": 4090
    },
    {
      "entropy": 2.2642083793878554,
      "epoch": 3.163674966222737,
      "grad_norm": 4.699156284332275,
      "learning_rate": 6.284852012936118e-06,
      "loss": 0.01841319501399994,
      "mean_token_accuracy": 0.9952612414956092,
      "num_tokens": 14641945.0,
      "step": 4100
    },
    {
      "entropy": 2.2173303991556166,
      "epoch": 3.1713954834973945,
      "grad_norm": 3.3025972843170166,
      "learning_rate": 6.238490670434338e-06,
      "loss": 0.019408072531223296,
      "mean_token_accuracy": 0.9971791654825211,
      "num_tokens": 14675381.0,
      "step": 4110
    },
    {
      "entropy": 2.1818405985832214,
      "epoch": 3.179116000772052,
      "grad_norm": 0.2685317099094391,
      "learning_rate": 6.192223311174179e-06,
      "loss": 0.02000413089990616,
      "mean_token_accuracy": 0.9951157197356224,
      "num_tokens": 14713258.0,
      "step": 4120
    },
    {
      "entropy": 2.2165157198905945,
      "epoch": 3.186836518046709,
      "grad_norm": 3.8232998847961426,
      "learning_rate": 6.146051091169367e-06,
      "loss": 0.01722181886434555,
      "mean_token_accuracy": 0.9972248584032058,
      "num_tokens": 14749403.0,
      "step": 4130
    },
    {
      "entropy": 2.2096123516559603,
      "epoch": 3.1945570353213664,
      "grad_norm": 0.10278035700321198,
      "learning_rate": 6.099975164056524e-06,
      "loss": 0.006322793662548065,
      "mean_token_accuracy": 0.9983333334326744,
      "num_tokens": 14785151.0,
      "step": 4140
    },
    {
      "entropy": 2.1967154651880265,
      "epoch": 3.202277552596024,
      "grad_norm": 1.6875110864639282,
      "learning_rate": 6.053996681066334e-06,
      "loss": 0.02624501585960388,
      "mean_token_accuracy": 0.9960680216550827,
      "num_tokens": 14819187.0,
      "step": 4150
    },
    {
      "entropy": 2.2883317291736605,
      "epoch": 3.2099980698706814,
      "grad_norm": 0.17266136407852173,
      "learning_rate": 6.008116790994805e-06,
      "loss": 0.01184847354888916,
      "mean_token_accuracy": 0.9966648116707801,
      "num_tokens": 14851008.0,
      "step": 4160
    },
    {
      "entropy": 2.185150256752968,
      "epoch": 3.2177185871453386,
      "grad_norm": 1.444366693496704,
      "learning_rate": 5.962336640174545e-06,
      "loss": 0.015763652324676514,
      "mean_token_accuracy": 0.9958777576684952,
      "num_tokens": 14887540.0,
      "step": 4170
    },
    {
      "entropy": 2.2751846730709078,
      "epoch": 3.225439104419996,
      "grad_norm": 0.21020469069480896,
      "learning_rate": 5.916657372446117e-06,
      "loss": 0.006749919056892395,
      "mean_token_accuracy": 0.9985294118523598,
      "num_tokens": 14922146.0,
      "step": 4180
    },
    {
      "entropy": 2.1631966710090635,
      "epoch": 3.2331596216946537,
      "grad_norm": 0.33755046129226685,
      "learning_rate": 5.8710801291294815e-06,
      "loss": 0.03264306783676148,
      "mean_token_accuracy": 0.9941071435809136,
      "num_tokens": 14958922.0,
      "step": 4190
    },
    {
      "entropy": 2.2516215592622757,
      "epoch": 3.240880138969311,
      "grad_norm": 0.3011329174041748,
      "learning_rate": 5.825606048995459e-06,
      "loss": 0.008013246208429336,
      "mean_token_accuracy": 0.9982007578015327,
      "num_tokens": 14995286.0,
      "step": 4200
    },
    {
      "entropy": 2.231572899222374,
      "epoch": 3.248600656243968,
      "grad_norm": 2.2834465503692627,
      "learning_rate": 5.780236268237286e-06,
      "loss": 0.015553775429725646,
      "mean_token_accuracy": 0.9944224849343299,
      "num_tokens": 15031929.0,
      "step": 4210
    },
    {
      "entropy": 2.2093691408634184,
      "epoch": 3.256321173518626,
      "grad_norm": 0.22805458307266235,
      "learning_rate": 5.7349719204422214e-06,
      "loss": 0.011248494684696197,
      "mean_token_accuracy": 0.9973071217536926,
      "num_tokens": 15072819.0,
      "step": 4220
    },
    {
      "entropy": 2.22184799015522,
      "epoch": 3.264041690793283,
      "grad_norm": 0.4345404803752899,
      "learning_rate": 5.689814136563241e-06,
      "loss": 0.006262413412332535,
      "mean_token_accuracy": 0.998863635957241,
      "num_tokens": 15108990.0,
      "step": 4230
    },
    {
      "entropy": 2.2079828560352324,
      "epoch": 3.2717622080679405,
      "grad_norm": 0.5222209095954895,
      "learning_rate": 5.644764044890752e-06,
      "loss": 0.013622181117534637,
      "mean_token_accuracy": 0.9969899669289589,
      "num_tokens": 15141615.0,
      "step": 4240
    },
    {
      "entropy": 2.149238818883896,
      "epoch": 3.2794827253425978,
      "grad_norm": 2.1469385623931885,
      "learning_rate": 5.599822771024419e-06,
      "loss": 0.015719014406204223,
      "mean_token_accuracy": 0.9969118013978004,
      "num_tokens": 15180770.0,
      "step": 4250
    },
    {
      "entropy": 2.241379123926163,
      "epoch": 3.2872032426172555,
      "grad_norm": 0.027104590088129044,
      "learning_rate": 5.554991437845048e-06,
      "loss": 0.0024414468556642533,
      "mean_token_accuracy": 1.0,
      "num_tokens": 15211585.0,
      "step": 4260
    },
    {
      "entropy": 2.1398084282875063,
      "epoch": 3.2949237598919128,
      "grad_norm": 0.11109408736228943,
      "learning_rate": 5.510271165486509e-06,
      "loss": 0.006358934193849563,
      "mean_token_accuracy": 0.9989583328366279,
      "num_tokens": 15248698.0,
      "step": 4270
    },
    {
      "entropy": 2.2906727343797684,
      "epoch": 3.30264427716657,
      "grad_norm": 0.28413844108581543,
      "learning_rate": 5.465663071307762e-06,
      "loss": 0.012974724173545837,
      "mean_token_accuracy": 0.9986111104488373,
      "num_tokens": 15280972.0,
      "step": 4280
    },
    {
      "entropy": 2.236059454083443,
      "epoch": 3.3103647944412273,
      "grad_norm": 0.6353345513343811,
      "learning_rate": 5.421168269864947e-06,
      "loss": 0.010030826926231385,
      "mean_token_accuracy": 0.9976310476660728,
      "num_tokens": 15318573.0,
      "step": 4290
    },
    {
      "entropy": 2.277936393022537,
      "epoch": 3.318085311715885,
      "grad_norm": 0.15945784747600555,
      "learning_rate": 5.376787872883518e-06,
      "loss": 0.02424613982439041,
      "mean_token_accuracy": 0.9960007324814797,
      "num_tokens": 15349593.0,
      "step": 4300
    },
    {
      "entropy": 2.1999513030052187,
      "epoch": 3.3258058289905423,
      "grad_norm": 0.8073568344116211,
      "learning_rate": 5.332522989230474e-06,
      "loss": 0.014913980662822724,
      "mean_token_accuracy": 0.9958407923579216,
      "num_tokens": 15384204.0,
      "step": 4310
    },
    {
      "entropy": 2.273113438487053,
      "epoch": 3.3335263462651996,
      "grad_norm": 0.010039719752967358,
      "learning_rate": 5.288374724886656e-06,
      "loss": 0.008377297222614289,
      "mean_token_accuracy": 0.9974475517868996,
      "num_tokens": 15419086.0,
      "step": 4320
    },
    {
      "entropy": 2.123158890008926,
      "epoch": 3.3412468635398573,
      "grad_norm": 0.423715204000473,
      "learning_rate": 5.2443441829191175e-06,
      "loss": 0.018509870767593382,
      "mean_token_accuracy": 0.9957080200314522,
      "num_tokens": 15456863.0,
      "step": 4330
    },
    {
      "entropy": 2.177235481142998,
      "epoch": 3.3489673808145146,
      "grad_norm": 0.035253945738077164,
      "learning_rate": 5.200432463453552e-06,
      "loss": 0.012387002259492874,
      "mean_token_accuracy": 0.9969292849302291,
      "num_tokens": 15494675.0,
      "step": 4340
    },
    {
      "entropy": 2.1819762349128724,
      "epoch": 3.356687898089172,
      "grad_norm": 0.7576441764831543,
      "learning_rate": 5.1566406636468085e-06,
      "loss": 0.028994646668434144,
      "mean_token_accuracy": 0.9916816547513008,
      "num_tokens": 15529963.0,
      "step": 4350
    },
    {
      "entropy": 2.1758675992488863,
      "epoch": 3.3644084153638296,
      "grad_norm": 0.2250443994998932,
      "learning_rate": 5.112969877659493e-06,
      "loss": 0.004997969418764114,
      "mean_token_accuracy": 0.9987468659877777,
      "num_tokens": 15564461.0,
      "step": 4360
    },
    {
      "entropy": 2.239803743362427,
      "epoch": 3.372128932638487,
      "grad_norm": 0.1805865317583084,
      "learning_rate": 5.06942119662861e-06,
      "loss": 0.005810170248150825,
      "mean_token_accuracy": 0.9986111104488373,
      "num_tokens": 15598051.0,
      "step": 4370
    },
    {
      "entropy": 2.125052031874657,
      "epoch": 3.379849449913144,
      "grad_norm": 4.371738433837891,
      "learning_rate": 5.02599570864031e-06,
      "loss": 0.015517064929008484,
      "mean_token_accuracy": 0.9934965163469315,
      "num_tokens": 15637669.0,
      "step": 4380
    },
    {
      "entropy": 2.2460622638463974,
      "epoch": 3.3875699671878015,
      "grad_norm": 0.18027924001216888,
      "learning_rate": 4.982694498702702e-06,
      "loss": 0.005739504098892212,
      "mean_token_accuracy": 0.9993750005960464,
      "num_tokens": 15670862.0,
      "step": 4390
    },
    {
      "entropy": 2.213028836250305,
      "epoch": 3.3952904844624587,
      "grad_norm": 1.9674371480941772,
      "learning_rate": 4.939518648718746e-06,
      "loss": 0.012049780786037445,
      "mean_token_accuracy": 0.9981812179088593,
      "num_tokens": 15706922.0,
      "step": 4400
    },
    {
      "entropy": 2.2505041897296905,
      "epoch": 3.4030110017371165,
      "grad_norm": 1.6243531703948975,
      "learning_rate": 4.896469237459228e-06,
      "loss": 0.003813191503286362,
      "mean_token_accuracy": 0.999074074625969,
      "num_tokens": 15745049.0,
      "step": 4410
    },
    {
      "entropy": 2.102014201879501,
      "epoch": 3.4107315190117737,
      "grad_norm": 2.138728618621826,
      "learning_rate": 4.853547340535775e-06,
      "loss": 0.014606839418411255,
      "mean_token_accuracy": 0.998145604133606,
      "num_tokens": 15782169.0,
      "step": 4420
    },
    {
      "entropy": 2.166767033934593,
      "epoch": 3.418452036286431,
      "grad_norm": 0.09994740039110184,
      "learning_rate": 4.810754030374024e-06,
      "loss": 0.022449350357055663,
      "mean_token_accuracy": 0.9930004492402077,
      "num_tokens": 15819096.0,
      "step": 4430
    },
    {
      "entropy": 2.1858236521482466,
      "epoch": 3.4261725535610887,
      "grad_norm": 0.3162212073802948,
      "learning_rate": 4.768090376186797e-06,
      "loss": 0.023777911067008974,
      "mean_token_accuracy": 0.9946548566222191,
      "num_tokens": 15862846.0,
      "step": 4440
    },
    {
      "entropy": 2.189643532037735,
      "epoch": 3.433893070835746,
      "grad_norm": 0.39837151765823364,
      "learning_rate": 4.72555744394739e-06,
      "loss": 0.018805976212024688,
      "mean_token_accuracy": 0.9932140588760376,
      "num_tokens": 15904709.0,
      "step": 4450
    },
    {
      "entropy": 2.2112290412187576,
      "epoch": 3.4416135881104033,
      "grad_norm": 0.025752931833267212,
      "learning_rate": 4.68315629636296e-06,
      "loss": 0.0037055592983961105,
      "mean_token_accuracy": 0.9986111104488373,
      "num_tokens": 15940870.0,
      "step": 4460
    },
    {
      "entropy": 2.195486468076706,
      "epoch": 3.449334105385061,
      "grad_norm": 3.4531970024108887,
      "learning_rate": 4.64088799284794e-06,
      "loss": 0.009186282008886337,
      "mean_token_accuracy": 0.996559077501297,
      "num_tokens": 15977007.0,
      "step": 4470
    },
    {
      "entropy": 2.2477256029844286,
      "epoch": 3.4570546226597183,
      "grad_norm": 0.017128420993685722,
      "learning_rate": 4.598753589497595e-06,
      "loss": 0.034723791480064395,
      "mean_token_accuracy": 0.990346123278141,
      "num_tokens": 16010302.0,
      "step": 4480
    },
    {
      "entropy": 2.266435816884041,
      "epoch": 3.4647751399343756,
      "grad_norm": 0.011056693270802498,
      "learning_rate": 4.5567541390616165e-06,
      "loss": 0.020518609881401063,
      "mean_token_accuracy": 0.994847884774208,
      "num_tokens": 16045264.0,
      "step": 4490
    },
    {
      "entropy": 2.2363200038671494,
      "epoch": 3.472495657209033,
      "grad_norm": 0.28764382004737854,
      "learning_rate": 4.514890690917844e-06,
      "loss": 0.030601662397384644,
      "mean_token_accuracy": 0.9958360910415649,
      "num_tokens": 16084354.0,
      "step": 4500
    },
    {
      "epoch": 3.472495657209033,
      "eval_entropy": 2.094737304290883,
      "eval_loss": 0.49774986505508423,
      "eval_mean_token_accuracy": 0.9358505434127935,
      "eval_num_tokens": 16084354.0,
      "eval_runtime": 987.909,
      "eval_samples_per_second": 7.745,
      "eval_steps_per_second": 0.969,
      "step": 4500
    },
    {
      "entropy": 2.167369270324707,
      "epoch": 3.4802161744836906,
      "grad_norm": 0.8214945793151855,
      "learning_rate": 4.473164291046016e-06,
      "loss": 0.0059088427573442456,
      "mean_token_accuracy": 0.9981317937374115,
      "num_tokens": 16120147.0,
      "step": 4510
    },
    {
      "entropy": 2.227228084206581,
      "epoch": 3.487936691758348,
      "grad_norm": 0.32604676485061646,
      "learning_rate": 4.431575982001651e-06,
      "loss": 0.011142008751630784,
      "mean_token_accuracy": 0.9969166651368141,
      "num_tokens": 16153451.0,
      "step": 4520
    },
    {
      "entropy": 2.2228987962007523,
      "epoch": 3.495657209033005,
      "grad_norm": 4.1868085861206055,
      "learning_rate": 4.390126802890008e-06,
      "loss": 0.021856479346752167,
      "mean_token_accuracy": 0.9982320085167885,
      "num_tokens": 16189202.0,
      "step": 4530
    },
    {
      "entropy": 2.173022910952568,
      "epoch": 3.5033777263076624,
      "grad_norm": 1.626468539237976,
      "learning_rate": 4.348817789340102e-06,
      "loss": 0.006640783697366715,
      "mean_token_accuracy": 0.9993055552244187,
      "num_tokens": 16229153.0,
      "step": 4540
    },
    {
      "entropy": 2.2181044608354568,
      "epoch": 3.51109824358232,
      "grad_norm": 0.957091748714447,
      "learning_rate": 4.307649973478838e-06,
      "loss": 0.021261365711688997,
      "mean_token_accuracy": 0.9973930478096008,
      "num_tokens": 16261419.0,
      "step": 4550
    },
    {
      "entropy": 2.163629150390625,
      "epoch": 3.5188187608569774,
      "grad_norm": 3.54240083694458,
      "learning_rate": 4.266624383905238e-06,
      "loss": 0.014290031790733338,
      "mean_token_accuracy": 0.9942340418696404,
      "num_tokens": 16297418.0,
      "step": 4560
    },
    {
      "entropy": 2.165030711889267,
      "epoch": 3.5265392781316347,
      "grad_norm": 0.8268080353736877,
      "learning_rate": 4.225742045664713e-06,
      "loss": 0.013842463493347168,
      "mean_token_accuracy": 0.9971296295523644,
      "num_tokens": 16333646.0,
      "step": 4570
    },
    {
      "entropy": 2.2287199676036833,
      "epoch": 3.5342597954062924,
      "grad_norm": 0.16437366604804993,
      "learning_rate": 4.1850039802234754e-06,
      "loss": 0.0037029199302196503,
      "mean_token_accuracy": 0.9989583328366279,
      "num_tokens": 16367085.0,
      "step": 4580
    },
    {
      "entropy": 2.2051072388887407,
      "epoch": 3.5419803126809497,
      "grad_norm": 0.2944578528404236,
      "learning_rate": 4.144411205442996e-06,
      "loss": 0.022799597680568696,
      "mean_token_accuracy": 0.995043994486332,
      "num_tokens": 16403393.0,
      "step": 4590
    },
    {
      "entropy": 2.192304465174675,
      "epoch": 3.549700829955607,
      "grad_norm": 3.0054590702056885,
      "learning_rate": 4.103964735554602e-06,
      "loss": 0.031678736209869385,
      "mean_token_accuracy": 0.9951188251376152,
      "num_tokens": 16438181.0,
      "step": 4600
    },
    {
      "entropy": 2.24237465262413,
      "epoch": 3.5574213472302643,
      "grad_norm": 3.488093376159668,
      "learning_rate": 4.063665581134101e-06,
      "loss": 0.016337577998638154,
      "mean_token_accuracy": 0.9961950778961182,
      "num_tokens": 16474359.0,
      "step": 4610
    },
    {
      "entropy": 2.1587339490652084,
      "epoch": 3.565141864504922,
      "grad_norm": 0.4733858108520508,
      "learning_rate": 4.023514749076553e-06,
      "loss": 0.022992467880249022,
      "mean_token_accuracy": 0.9947966426610947,
      "num_tokens": 16508074.0,
      "step": 4620
    },
    {
      "entropy": 2.2088340640068056,
      "epoch": 3.5728623817795793,
      "grad_norm": 0.09064114093780518,
      "learning_rate": 3.983513242571113e-06,
      "loss": 0.024549849331378937,
      "mean_token_accuracy": 0.9951631486415863,
      "num_tokens": 16545666.0,
      "step": 4630
    },
    {
      "entropy": 2.1512410193681717,
      "epoch": 3.5805828990542365,
      "grad_norm": 0.4027511179447174,
      "learning_rate": 3.943662061075957e-06,
      "loss": 0.008296293765306472,
      "mean_token_accuracy": 0.9972971126437187,
      "num_tokens": 16582524.0,
      "step": 4640
    },
    {
      "entropy": 2.2184607803821565,
      "epoch": 3.588303416328894,
      "grad_norm": 0.21228182315826416,
      "learning_rate": 3.903962200293305e-06,
      "loss": 0.006698296964168548,
      "mean_token_accuracy": 0.9988095238804817,
      "num_tokens": 16619971.0,
      "step": 4650
    },
    {
      "entropy": 2.1930524677038195,
      "epoch": 3.5960239336035515,
      "grad_norm": 0.12852530181407928,
      "learning_rate": 3.86441465214457e-06,
      "loss": 0.019022318720817565,
      "mean_token_accuracy": 0.9959166675806046,
      "num_tokens": 16655670.0,
      "step": 4660
    },
    {
      "entropy": 2.2068610817193983,
      "epoch": 3.603744450878209,
      "grad_norm": 0.11005378514528275,
      "learning_rate": 3.825020404745539e-06,
      "loss": 0.02366378754377365,
      "mean_token_accuracy": 0.9947434559464454,
      "num_tokens": 16692946.0,
      "step": 4670
    },
    {
      "entropy": 2.158995199203491,
      "epoch": 3.611464968152866,
      "grad_norm": 0.06957825273275375,
      "learning_rate": 3.785780442381708e-06,
      "loss": 0.006554947793483734,
      "mean_token_accuracy": 0.9985294118523598,
      "num_tokens": 16728342.0,
      "step": 4680
    },
    {
      "entropy": 2.201757091283798,
      "epoch": 3.619185485427524,
      "grad_norm": 0.013743036426603794,
      "learning_rate": 3.746695745483676e-06,
      "loss": 0.022013702988624574,
      "mean_token_accuracy": 0.9964899107813835,
      "num_tokens": 16765997.0,
      "step": 4690
    },
    {
      "entropy": 2.189955911040306,
      "epoch": 3.626906002702181,
      "grad_norm": 1.3227832317352295,
      "learning_rate": 3.7077672906026674e-06,
      "loss": 0.009854133427143096,
      "mean_token_accuracy": 0.9972507312893868,
      "num_tokens": 16806476.0,
      "step": 4700
    },
    {
      "entropy": 2.2155461460351944,
      "epoch": 3.6346265199768384,
      "grad_norm": 0.07080277055501938,
      "learning_rate": 3.66899605038611e-06,
      "loss": 0.018130703270435332,
      "mean_token_accuracy": 0.9930429771542549,
      "num_tokens": 16839996.0,
      "step": 4710
    },
    {
      "entropy": 2.2547535479068754,
      "epoch": 3.642347037251496,
      "grad_norm": 0.03352458029985428,
      "learning_rate": 3.6303829935533387e-06,
      "loss": 0.023316796123981475,
      "mean_token_accuracy": 0.9934576332569123,
      "num_tokens": 16870710.0,
      "step": 4720
    },
    {
      "entropy": 2.2180339872837065,
      "epoch": 3.6500675545261534,
      "grad_norm": 0.08932402729988098,
      "learning_rate": 3.5919290848714083e-06,
      "loss": 0.019556549191474915,
      "mean_token_accuracy": 0.9943825855851174,
      "num_tokens": 16905867.0,
      "step": 4730
    },
    {
      "entropy": 2.25754671394825,
      "epoch": 3.6577880718008107,
      "grad_norm": 0.36979636549949646,
      "learning_rate": 3.5536352851309654e-06,
      "loss": 0.010746852308511735,
      "mean_token_accuracy": 0.9972794115543365,
      "num_tokens": 16942946.0,
      "step": 4740
    },
    {
      "entropy": 2.173688989877701,
      "epoch": 3.665508589075468,
      "grad_norm": 0.9019256830215454,
      "learning_rate": 3.5155025511222553e-06,
      "loss": 0.013900287449359894,
      "mean_token_accuracy": 0.9956018269062042,
      "num_tokens": 16982691.0,
      "step": 4750
    },
    {
      "entropy": 2.186745339632034,
      "epoch": 3.6732291063501252,
      "grad_norm": 1.9558967351913452,
      "learning_rate": 3.4775318356112085e-06,
      "loss": 0.0026500292122364046,
      "mean_token_accuracy": 0.9984375,
      "num_tokens": 17014619.0,
      "step": 4760
    },
    {
      "entropy": 2.190760698914528,
      "epoch": 3.680949623624783,
      "grad_norm": 0.6850957274436951,
      "learning_rate": 3.439724087315649e-06,
      "loss": 0.008563292026519776,
      "mean_token_accuracy": 0.997857142984867,
      "num_tokens": 17049479.0,
      "step": 4770
    },
    {
      "entropy": 2.2078921467065813,
      "epoch": 3.6886701408994402,
      "grad_norm": 0.48178425431251526,
      "learning_rate": 3.4020802508815844e-06,
      "loss": 0.00671912282705307,
      "mean_token_accuracy": 0.9985702604055404,
      "num_tokens": 17084774.0,
      "step": 4780
    },
    {
      "entropy": 2.199252927303314,
      "epoch": 3.6963906581740975,
      "grad_norm": 3.1879372596740723,
      "learning_rate": 3.3646012668595797e-06,
      "loss": 0.012920960783958435,
      "mean_token_accuracy": 0.9954613089561463,
      "num_tokens": 17118756.0,
      "step": 4790
    },
    {
      "entropy": 2.192939519882202,
      "epoch": 3.7041111754487552,
      "grad_norm": 3.9493212699890137,
      "learning_rate": 3.3272880716812994e-06,
      "loss": 0.010051518678665161,
      "mean_token_accuracy": 0.996506179869175,
      "num_tokens": 17151694.0,
      "step": 4800
    },
    {
      "entropy": 2.2035789877176284,
      "epoch": 3.7118316927234125,
      "grad_norm": 0.0355636328458786,
      "learning_rate": 3.2901415976360773e-06,
      "loss": 0.012046564370393753,
      "mean_token_accuracy": 0.9979532152414322,
      "num_tokens": 17188300.0,
      "step": 4810
    },
    {
      "entropy": 2.265298753976822,
      "epoch": 3.71955220999807,
      "grad_norm": 0.5305169224739075,
      "learning_rate": 3.2531627728476357e-06,
      "loss": 0.015401627123355865,
      "mean_token_accuracy": 0.9960317447781563,
      "num_tokens": 17226628.0,
      "step": 4820
    },
    {
      "entropy": 2.235144093632698,
      "epoch": 3.7272727272727275,
      "grad_norm": 1.9314122200012207,
      "learning_rate": 3.2163525212509008e-06,
      "loss": 0.012413406372070312,
      "mean_token_accuracy": 0.9970962628722191,
      "num_tokens": 17261288.0,
      "step": 4830
    },
    {
      "entropy": 2.1957934617996218,
      "epoch": 3.734993244547385,
      "grad_norm": 0.05546269565820694,
      "learning_rate": 3.1797117625689024e-06,
      "loss": 0.00331185944378376,
      "mean_token_accuracy": 1.0,
      "num_tokens": 17296586.0,
      "step": 4840
    },
    {
      "entropy": 2.2121754467487333,
      "epoch": 3.742713761822042,
      "grad_norm": 0.16634555160999298,
      "learning_rate": 3.1432414122898057e-06,
      "loss": 0.01630573719739914,
      "mean_token_accuracy": 0.99408238530159,
      "num_tokens": 17330163.0,
      "step": 4850
    },
    {
      "entropy": 2.2177652209997176,
      "epoch": 3.7504342790966994,
      "grad_norm": 0.054232124239206314,
      "learning_rate": 3.1069423816440324e-06,
      "loss": 0.005872461199760437,
      "mean_token_accuracy": 0.9970823854207993,
      "num_tokens": 17364167.0,
      "step": 4860
    },
    {
      "entropy": 2.2218244522809982,
      "epoch": 3.7581547963713566,
      "grad_norm": 0.05728960037231445,
      "learning_rate": 3.070815577581505e-06,
      "loss": 0.023728246986865997,
      "mean_token_accuracy": 0.994246032834053,
      "num_tokens": 17396732.0,
      "step": 4870
    },
    {
      "entropy": 2.270163568854332,
      "epoch": 3.7658753136460144,
      "grad_norm": 2.1181461811065674,
      "learning_rate": 3.0348619027489656e-06,
      "loss": 0.013901802897453307,
      "mean_token_accuracy": 0.9959643810987473,
      "num_tokens": 17433024.0,
      "step": 4880
    },
    {
      "entropy": 2.194043418765068,
      "epoch": 3.7735958309206716,
      "grad_norm": 0.1262357532978058,
      "learning_rate": 2.9990822554674325e-06,
      "loss": 0.009613892436027527,
      "mean_token_accuracy": 0.9968301028013229,
      "num_tokens": 17472584.0,
      "step": 4890
    },
    {
      "entropy": 2.2226660668849947,
      "epoch": 3.781316348195329,
      "grad_norm": 0.12729032337665558,
      "learning_rate": 2.9634775297097693e-06,
      "loss": 0.0013892247341573238,
      "mean_token_accuracy": 1.0,
      "num_tokens": 17509334.0,
      "step": 4900
    },
    {
      "entropy": 2.2277002543210984,
      "epoch": 3.7890368654699866,
      "grad_norm": 0.04032934457063675,
      "learning_rate": 2.928048615078324e-06,
      "loss": 0.008014556765556336,
      "mean_token_accuracy": 0.9958013474941254,
      "num_tokens": 17551004.0,
      "step": 4910
    },
    {
      "entropy": 2.23101692199707,
      "epoch": 3.796757382744644,
      "grad_norm": 0.05918664112687111,
      "learning_rate": 2.8927963967827124e-06,
      "loss": 0.011012130230665208,
      "mean_token_accuracy": 0.9966880336403847,
      "num_tokens": 17590813.0,
      "step": 4920
    },
    {
      "entropy": 2.309200382232666,
      "epoch": 3.804477900019301,
      "grad_norm": 0.039273522794246674,
      "learning_rate": 2.8577217556177115e-06,
      "loss": 0.010339566320180894,
      "mean_token_accuracy": 0.9992424249649048,
      "num_tokens": 17622252.0,
      "step": 4930
    },
    {
      "entropy": 2.2433013200759886,
      "epoch": 3.812198417293959,
      "grad_norm": 0.3272627294063568,
      "learning_rate": 2.8228255679412353e-06,
      "loss": 0.009843137860298157,
      "mean_token_accuracy": 0.99921875,
      "num_tokens": 17662561.0,
      "step": 4940
    },
    {
      "entropy": 2.2392791122198106,
      "epoch": 3.819918934568616,
      "grad_norm": 0.038887813687324524,
      "learning_rate": 2.7881087056524468e-06,
      "loss": 0.008080794662237167,
      "mean_token_accuracy": 0.9986842110753059,
      "num_tokens": 17694695.0,
      "step": 4950
    },
    {
      "entropy": 2.2464512795209886,
      "epoch": 3.8276394518432735,
      "grad_norm": 0.4352750778198242,
      "learning_rate": 2.7535720361699703e-06,
      "loss": 0.006981483101844788,
      "mean_token_accuracy": 0.9978636354207993,
      "num_tokens": 17726808.0,
      "step": 4960
    },
    {
      "entropy": 2.2429711014032363,
      "epoch": 3.8353599691179308,
      "grad_norm": 3.1303551197052,
      "learning_rate": 2.7192164224102267e-06,
      "loss": 0.009275702387094497,
      "mean_token_accuracy": 0.9982230380177498,
      "num_tokens": 17763685.0,
      "step": 4970
    },
    {
      "entropy": 2.167110252380371,
      "epoch": 3.843080486392588,
      "grad_norm": 3.1086466312408447,
      "learning_rate": 2.6850427227658625e-06,
      "loss": 0.01786840558052063,
      "mean_token_accuracy": 0.9957186207175255,
      "num_tokens": 17800532.0,
      "step": 4980
    },
    {
      "entropy": 2.202439108490944,
      "epoch": 3.8508010036672458,
      "grad_norm": 0.03255437687039375,
      "learning_rate": 2.6510517910843014e-06,
      "loss": 0.014580619335174561,
      "mean_token_accuracy": 0.9978412076830864,
      "num_tokens": 17836043.0,
      "step": 4990
    },
    {
      "entropy": 2.2670375287532805,
      "epoch": 3.858521520941903,
      "grad_norm": 0.09342753142118454,
      "learning_rate": 2.617244476646429e-06,
      "loss": 0.01826583743095398,
      "mean_token_accuracy": 0.9961647734045982,
      "num_tokens": 17872352.0,
      "step": 5000
    },
    {
      "epoch": 3.858521520941903,
      "eval_entropy": 2.0908612560951845,
      "eval_loss": 0.507649302482605,
      "eval_mean_token_accuracy": 0.9354228066551025,
      "eval_num_tokens": 17872352.0,
      "eval_runtime": 987.5554,
      "eval_samples_per_second": 7.747,
      "eval_steps_per_second": 0.969,
      "step": 5000
    },
    {
      "entropy": 2.1657870292663572,
      "epoch": 3.8662420382165603,
      "grad_norm": 0.19421711564064026,
      "learning_rate": 2.5836216241453473e-06,
      "loss": 0.014168056845664977,
      "mean_token_accuracy": 0.9971320345997811,
      "num_tokens": 17912375.0,
      "step": 5010
    },
    {
      "entropy": 2.1820801228284834,
      "epoch": 3.873962555491218,
      "grad_norm": 1.4638241529464722,
      "learning_rate": 2.5501840736652884e-06,
      "loss": 0.008419310301542282,
      "mean_token_accuracy": 0.9989999994635582,
      "num_tokens": 17950474.0,
      "step": 5020
    },
    {
      "entropy": 2.2266716092824934,
      "epoch": 3.8816830727658753,
      "grad_norm": 0.12025623768568039,
      "learning_rate": 2.5169326606606127e-06,
      "loss": 0.004219979792833328,
      "mean_token_accuracy": 0.9989130437374115,
      "num_tokens": 17983123.0,
      "step": 5030
    },
    {
      "entropy": 2.23225157558918,
      "epoch": 3.8894035900405326,
      "grad_norm": 2.4465901851654053,
      "learning_rate": 2.4838682159349504e-06,
      "loss": 0.01536642462015152,
      "mean_token_accuracy": 0.9948759660124779,
      "num_tokens": 18016866.0,
      "step": 5040
    },
    {
      "entropy": 2.2503413647413253,
      "epoch": 3.8971241073151903,
      "grad_norm": 0.26093390583992004,
      "learning_rate": 2.4509915656204242e-06,
      "loss": 0.018379612267017363,
      "mean_token_accuracy": 0.9953591957688331,
      "num_tokens": 18050369.0,
      "step": 5050
    },
    {
      "entropy": 2.346876162290573,
      "epoch": 3.9048446245898476,
      "grad_norm": 0.04715586453676224,
      "learning_rate": 2.418303531157017e-06,
      "loss": 0.006166347116231918,
      "mean_token_accuracy": 0.9976365551352501,
      "num_tokens": 18085766.0,
      "step": 5060
    },
    {
      "entropy": 2.2075785368680956,
      "epoch": 3.912565141864505,
      "grad_norm": 0.06558074057102203,
      "learning_rate": 2.3858049292720565e-06,
      "loss": 0.010641612112522125,
      "mean_token_accuracy": 0.9973457798361778,
      "num_tokens": 18118325.0,
      "step": 5070
    },
    {
      "entropy": 2.1510800033807755,
      "epoch": 3.920285659139162,
      "grad_norm": 2.4715569019317627,
      "learning_rate": 2.353496571959791e-06,
      "loss": 0.008443721383810044,
      "mean_token_accuracy": 0.9969614043831825,
      "num_tokens": 18155458.0,
      "step": 5080
    },
    {
      "entropy": 2.219439336657524,
      "epoch": 3.92800617641382,
      "grad_norm": 0.02134764939546585,
      "learning_rate": 2.3213792664611123e-06,
      "loss": 0.020188172161579133,
      "mean_token_accuracy": 0.9954755708575249,
      "num_tokens": 18188388.0,
      "step": 5090
    },
    {
      "entropy": 2.1883194625377653,
      "epoch": 3.935726693688477,
      "grad_norm": 0.009646475315093994,
      "learning_rate": 2.289453815243391e-06,
      "loss": 0.009002915769815444,
      "mean_token_accuracy": 0.9987499997019768,
      "num_tokens": 18228648.0,
      "step": 5100
    },
    {
      "entropy": 2.1794584810733797,
      "epoch": 3.9434472109631344,
      "grad_norm": 0.06897470355033875,
      "learning_rate": 2.2577210159804107e-06,
      "loss": 0.01597502678632736,
      "mean_token_accuracy": 0.9932748913764954,
      "num_tokens": 18267139.0,
      "step": 5110
    },
    {
      "entropy": 2.217776373028755,
      "epoch": 3.9511677282377917,
      "grad_norm": 3.1233739852905273,
      "learning_rate": 2.226181661532455e-06,
      "loss": 0.02588752806186676,
      "mean_token_accuracy": 0.993584081530571,
      "num_tokens": 18305735.0,
      "step": 5120
    },
    {
      "entropy": 2.188844358921051,
      "epoch": 3.9588882455124494,
      "grad_norm": 0.6708783507347107,
      "learning_rate": 2.1948365399264802e-06,
      "loss": 0.014759351313114167,
      "mean_token_accuracy": 0.9980769217014313,
      "num_tokens": 18339445.0,
      "step": 5130
    },
    {
      "entropy": 2.1940950095653533,
      "epoch": 3.9666087627871067,
      "grad_norm": 0.0733635202050209,
      "learning_rate": 2.163686434336448e-06,
      "loss": 0.015484817326068878,
      "mean_token_accuracy": 0.9958730831742286,
      "num_tokens": 18372350.0,
      "step": 5140
    },
    {
      "entropy": 2.249452704191208,
      "epoch": 3.974329280061764,
      "grad_norm": 0.0156388096511364,
      "learning_rate": 2.1327321230637355e-06,
      "loss": 0.02897282540798187,
      "mean_token_accuracy": 0.994414983689785,
      "num_tokens": 18408315.0,
      "step": 5150
    },
    {
      "entropy": 2.2180632591247558,
      "epoch": 3.9820497973364217,
      "grad_norm": 0.5558152794837952,
      "learning_rate": 2.101974379517697e-06,
      "loss": 0.0015709832310676574,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18441147.0,
      "step": 5160
    },
    {
      "entropy": 2.251400685310364,
      "epoch": 3.989770314611079,
      "grad_norm": 0.07538606226444244,
      "learning_rate": 2.071413972196351e-06,
      "loss": 0.010088697820901871,
      "mean_token_accuracy": 0.9971969693899154,
      "num_tokens": 18477218.0,
      "step": 5170
    },
    {
      "entropy": 2.265915700793266,
      "epoch": 3.9974908318857363,
      "grad_norm": 0.07861519604921341,
      "learning_rate": 2.04105166466716e-06,
      "loss": 0.013128484785556793,
      "mean_token_accuracy": 0.9976960763335228,
      "num_tokens": 18513822.0,
      "step": 5180
    },
    {
      "entropy": 2.2155757626971684,
      "epoch": 4.004632310364794,
      "grad_norm": 0.15793584287166595,
      "learning_rate": 2.0108882155479602e-06,
      "loss": 0.014671583473682404,
      "mean_token_accuracy": 0.9959569025684047,
      "num_tokens": 18544084.0,
      "step": 5190
    },
    {
      "entropy": 2.216670718789101,
      "epoch": 4.012352827639452,
      "grad_norm": 0.10165757685899734,
      "learning_rate": 1.9809243784880215e-06,
      "loss": 0.001284267660230398,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18579882.0,
      "step": 5200
    },
    {
      "entropy": 2.1769883900880815,
      "epoch": 4.020073344914109,
      "grad_norm": 0.02957017533481121,
      "learning_rate": 1.9511609021491895e-06,
      "loss": 0.003650863468647003,
      "mean_token_accuracy": 0.9989999994635582,
      "num_tokens": 18615819.0,
      "step": 5210
    },
    {
      "entropy": 2.237304484844208,
      "epoch": 4.027793862188767,
      "grad_norm": 0.22377407550811768,
      "learning_rate": 1.921598530187202e-06,
      "loss": 0.0013460190035402774,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18649294.0,
      "step": 5220
    },
    {
      "entropy": 2.2296341180801393,
      "epoch": 4.035514379463424,
      "grad_norm": 0.028993308544158936,
      "learning_rate": 1.8922380012330943e-06,
      "loss": 0.001667243242263794,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18686007.0,
      "step": 5230
    },
    {
      "entropy": 2.258844423294067,
      "epoch": 4.043234896738081,
      "grad_norm": 0.047745853662490845,
      "learning_rate": 1.8630800488747625e-06,
      "loss": 0.005907272547483444,
      "mean_token_accuracy": 0.9991935476660728,
      "num_tokens": 18718290.0,
      "step": 5240
    },
    {
      "entropy": 2.2396200627088545,
      "epoch": 4.050955414012739,
      "grad_norm": 0.11687777936458588,
      "learning_rate": 1.8341254016386102e-06,
      "loss": 0.011040620505809784,
      "mean_token_accuracy": 0.9929608583450318,
      "num_tokens": 18758891.0,
      "step": 5250
    },
    {
      "entropy": 2.2722030729055405,
      "epoch": 4.058675931287397,
      "grad_norm": 1.0893008708953857,
      "learning_rate": 1.8053747829713618e-06,
      "loss": 0.0026631668210029603,
      "mean_token_accuracy": 0.998863635957241,
      "num_tokens": 18791486.0,
      "step": 5260
    },
    {
      "entropy": 2.2243568778038023,
      "epoch": 4.066396448562053,
      "grad_norm": 0.028796734288334846,
      "learning_rate": 1.7768289112219883e-06,
      "loss": 0.005374731123447418,
      "mean_token_accuracy": 0.9990384608507157,
      "num_tokens": 18825388.0,
      "step": 5270
    },
    {
      "entropy": 2.172247883677483,
      "epoch": 4.074116965836711,
      "grad_norm": 0.03655334934592247,
      "learning_rate": 1.7484884996237472e-06,
      "loss": 0.0015311230905354022,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18863479.0,
      "step": 5280
    },
    {
      "entropy": 2.1923954993486405,
      "epoch": 4.081837483111369,
      "grad_norm": 0.11251391470432281,
      "learning_rate": 1.7203542562763699e-06,
      "loss": 0.0013637151569128036,
      "mean_token_accuracy": 1.0,
      "num_tokens": 18899886.0,
      "step": 5290
    },
    {
      "entropy": 2.1101179629564286,
      "epoch": 4.089558000386026,
      "grad_norm": 0.019721994176506996,
      "learning_rate": 1.6924268841283731e-06,
      "loss": 0.005000091716647148,
      "mean_token_accuracy": 0.9991935476660728,
      "num_tokens": 18943812.0,
      "step": 5300
    },
    {
      "entropy": 2.2096335530281066,
      "epoch": 4.097278517660683,
      "grad_norm": 0.3568638563156128,
      "learning_rate": 1.6647070809594846e-06,
      "loss": 0.010236522555351258,
      "mean_token_accuracy": 0.99627845287323,
      "num_tokens": 18976301.0,
      "step": 5310
    },
    {
      "entropy": 2.2267425060272217,
      "epoch": 4.10499903493534,
      "grad_norm": 0.1415398120880127,
      "learning_rate": 1.6371955393632177e-06,
      "loss": 0.012716498970985413,
      "mean_token_accuracy": 0.997547847032547,
      "num_tokens": 19015300.0,
      "step": 5320
    },
    {
      "entropy": 2.2259861052036287,
      "epoch": 4.112719552209998,
      "grad_norm": 0.0552356131374836,
      "learning_rate": 1.60989294672956e-06,
      "loss": 0.00908339098095894,
      "mean_token_accuracy": 0.9977884605526924,
      "num_tokens": 19051745.0,
      "step": 5330
    },
    {
      "entropy": 2.2949888169765473,
      "epoch": 4.120440069484656,
      "grad_norm": 0.13113631308078766,
      "learning_rate": 1.5827999852278097e-06,
      "loss": 0.0018477724865078927,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19083371.0,
      "step": 5340
    },
    {
      "entropy": 2.2030975073575974,
      "epoch": 4.1281605867593125,
      "grad_norm": 2.422152280807495,
      "learning_rate": 1.5559173317895148e-06,
      "loss": 0.005163485929369926,
      "mean_token_accuracy": 0.9991379305720329,
      "num_tokens": 19119667.0,
      "step": 5350
    },
    {
      "entropy": 2.203478366136551,
      "epoch": 4.13588110403397,
      "grad_norm": 0.4935871362686157,
      "learning_rate": 1.5292456580915703e-06,
      "loss": 0.0074889764189720156,
      "mean_token_accuracy": 0.9986068099737168,
      "num_tokens": 19158058.0,
      "step": 5360
    },
    {
      "entropy": 2.176115852594376,
      "epoch": 4.143601621308628,
      "grad_norm": 0.047461602836847305,
      "learning_rate": 1.5027856305394416e-06,
      "loss": 0.01174907609820366,
      "mean_token_accuracy": 0.9964136898517608,
      "num_tokens": 19191330.0,
      "step": 5370
    },
    {
      "entropy": 2.3252247244119646,
      "epoch": 4.151322138583285,
      "grad_norm": 0.07587450742721558,
      "learning_rate": 1.4765379102504973e-06,
      "loss": 0.004352330043911934,
      "mean_token_accuracy": 0.9991935476660728,
      "num_tokens": 19223065.0,
      "step": 5380
    },
    {
      "entropy": 2.2030561834573748,
      "epoch": 4.1590426558579425,
      "grad_norm": 0.0413011871278286,
      "learning_rate": 1.4505031530375024e-06,
      "loss": 0.0010141530074179173,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19258355.0,
      "step": 5390
    },
    {
      "entropy": 2.172553300857544,
      "epoch": 4.1667631731326,
      "grad_norm": 0.016729436814785004,
      "learning_rate": 1.424682009392232e-06,
      "loss": 0.008941520750522614,
      "mean_token_accuracy": 0.997578963637352,
      "num_tokens": 19293261.0,
      "step": 5400
    },
    {
      "entropy": 2.225874611735344,
      "epoch": 4.174483690407257,
      "grad_norm": 0.11103388667106628,
      "learning_rate": 1.399075124469217e-06,
      "loss": 0.004546413570642472,
      "mean_token_accuracy": 0.99921875,
      "num_tokens": 19327624.0,
      "step": 5410
    },
    {
      "entropy": 2.167690780758858,
      "epoch": 4.182204207681915,
      "grad_norm": 2.885631799697876,
      "learning_rate": 1.3736831380696224e-06,
      "loss": 0.01201588660478592,
      "mean_token_accuracy": 0.9971944436430931,
      "num_tokens": 19362087.0,
      "step": 5420
    },
    {
      "entropy": 2.269624611735344,
      "epoch": 4.1899247249565725,
      "grad_norm": 0.44468164443969727,
      "learning_rate": 1.3485066846252615e-06,
      "loss": 0.009978090226650239,
      "mean_token_accuracy": 0.9983333334326744,
      "num_tokens": 19396605.0,
      "step": 5430
    },
    {
      "entropy": 2.2737204790115357,
      "epoch": 4.197645242231229,
      "grad_norm": 0.1409270316362381,
      "learning_rate": 1.323546393182752e-06,
      "loss": 0.008260860294103622,
      "mean_token_accuracy": 0.9979480162262917,
      "num_tokens": 19434246.0,
      "step": 5440
    },
    {
      "entropy": 2.2025551408529283,
      "epoch": 4.205365759505887,
      "grad_norm": 2.0743470191955566,
      "learning_rate": 1.2988028873877867e-06,
      "loss": 0.007827727496623993,
      "mean_token_accuracy": 0.9981812179088593,
      "num_tokens": 19472258.0,
      "step": 5450
    },
    {
      "entropy": 2.0870246827602386,
      "epoch": 4.213086276780544,
      "grad_norm": 0.0854920968413353,
      "learning_rate": 1.2742767854695603e-06,
      "loss": 0.001750888116657734,
      "mean_token_accuracy": 0.9986111104488373,
      "num_tokens": 19515091.0,
      "step": 5460
    },
    {
      "entropy": 2.2541315644979476,
      "epoch": 4.220806794055202,
      "grad_norm": 0.050843846052885056,
      "learning_rate": 1.2499687002253214e-06,
      "loss": 0.012310877442359924,
      "mean_token_accuracy": 0.9966666668653488,
      "num_tokens": 19550438.0,
      "step": 5470
    },
    {
      "entropy": 2.281228691339493,
      "epoch": 4.228527311329859,
      "grad_norm": 0.10463957488536835,
      "learning_rate": 1.2258792390050589e-06,
      "loss": 0.009670499712228775,
      "mean_token_accuracy": 0.9984359726309776,
      "num_tokens": 19587460.0,
      "step": 5480
    },
    {
      "entropy": 2.19995211660862,
      "epoch": 4.236247828604516,
      "grad_norm": 0.029270397499203682,
      "learning_rate": 1.2020090036963272e-06,
      "loss": 0.0012936697341501713,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19622266.0,
      "step": 5490
    },
    {
      "entropy": 2.2033345282077788,
      "epoch": 4.243968345879174,
      "grad_norm": 0.2233721911907196,
      "learning_rate": 1.1783585907092077e-06,
      "loss": 0.011895059794187545,
      "mean_token_accuracy": 0.9973854154348374,
      "num_tokens": 19661206.0,
      "step": 5500
    },
    {
      "epoch": 4.243968345879174,
      "eval_entropy": 2.1031606457699303,
      "eval_loss": 0.5105812549591064,
      "eval_mean_token_accuracy": 0.93590848367416,
      "eval_num_tokens": 19661206.0,
      "eval_runtime": 987.4206,
      "eval_samples_per_second": 7.748,
      "eval_steps_per_second": 0.969,
      "step": 5500
    },
    {
      "entropy": 2.166568320989609,
      "epoch": 4.251688863153832,
      "grad_norm": 0.06564086675643921,
      "learning_rate": 1.1549285909614138e-06,
      "loss": 0.0009026847779750824,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19698305.0,
      "step": 5510
    },
    {
      "entropy": 2.2002768576145173,
      "epoch": 4.2594093804284885,
      "grad_norm": 0.12140869349241257,
      "learning_rate": 1.131719589863516e-06,
      "loss": 0.0019102182239294052,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19734728.0,
      "step": 5520
    },
    {
      "entropy": 2.2461003422737122,
      "epoch": 4.267129897703146,
      "grad_norm": 0.06065747141838074,
      "learning_rate": 1.1087321673043184e-06,
      "loss": 0.010919121652841568,
      "mean_token_accuracy": 0.9966435179114341,
      "num_tokens": 19766601.0,
      "step": 5530
    },
    {
      "entropy": 2.21742702126503,
      "epoch": 4.274850414977804,
      "grad_norm": 0.31923985481262207,
      "learning_rate": 1.0859668976363802e-06,
      "loss": 0.004073232784867287,
      "mean_token_accuracy": 0.9993055552244187,
      "num_tokens": 19799159.0,
      "step": 5540
    },
    {
      "entropy": 2.195229309797287,
      "epoch": 4.282570932252461,
      "grad_norm": 0.03684071823954582,
      "learning_rate": 1.063424349661647e-06,
      "loss": 0.0008944393135607243,
      "mean_token_accuracy": 1.0,
      "num_tokens": 19833600.0,
      "step": 5550
    },
    {
      "entropy": 2.204106843471527,
      "epoch": 4.2902914495271185,
      "grad_norm": 0.054600995033979416,
      "learning_rate": 1.0411050866172512e-06,
      "loss": 0.013252517580986023,
      "mean_token_accuracy": 0.9959087640047073,
      "num_tokens": 19878221.0,
      "step": 5560
    },
    {
      "entropy": 2.2110326796770097,
      "epoch": 4.298011966801775,
      "grad_norm": 1.9496335983276367,
      "learning_rate": 1.0190096661614424e-06,
      "loss": 0.019130887091159822,
      "mean_token_accuracy": 0.9931336134672165,
      "num_tokens": 19907682.0,
      "step": 5570
    },
    {
      "entropy": 2.2047807425260544,
      "epoch": 4.305732484076433,
      "grad_norm": 0.13513988256454468,
      "learning_rate": 9.971386403596417e-07,
      "loss": 0.008160575479269027,
      "mean_token_accuracy": 0.9981125354766845,
      "num_tokens": 19942095.0,
      "step": 5580
    },
    {
      "entropy": 2.188329595327377,
      "epoch": 4.313453001351091,
      "grad_norm": 1.0191106796264648,
      "learning_rate": 9.75492555670655e-07,
      "loss": 0.0053262699395418165,
      "mean_token_accuracy": 0.9979999989271164,
      "num_tokens": 19984249.0,
      "step": 5590
    },
    {
      "entropy": 2.2830605089664457,
      "epoch": 4.321173518625748,
      "grad_norm": 0.13544659316539764,
      "learning_rate": 9.540719529330212e-07,
      "loss": 0.007930701971054077,
      "mean_token_accuracy": 0.9986089944839478,
      "num_tokens": 20016851.0,
      "step": 5600
    },
    {
      "entropy": 2.222521474957466,
      "epoch": 4.328894035900405,
      "grad_norm": 0.042033832520246506,
      "learning_rate": 9.328773673514979e-07,
      "loss": 0.009003970026969909,
      "mean_token_accuracy": 0.9984122976660729,
      "num_tokens": 20049220.0,
      "step": 5610
    },
    {
      "entropy": 2.2348962247371675,
      "epoch": 4.336614553175063,
      "grad_norm": 0.030303576961159706,
      "learning_rate": 9.119093284836855e-07,
      "loss": 0.008937902003526687,
      "mean_token_accuracy": 0.9973191484808922,
      "num_tokens": 20082476.0,
      "step": 5620
    },
    {
      "entropy": 2.1816387206315992,
      "epoch": 4.34433507044972,
      "grad_norm": 0.06456495821475983,
      "learning_rate": 8.911683602267973e-07,
      "loss": 0.012813526391983032,
      "mean_token_accuracy": 0.9954789370298386,
      "num_tokens": 20120303.0,
      "step": 5630
    },
    {
      "entropy": 2.1896758466959,
      "epoch": 4.352055587724378,
      "grad_norm": 0.10599888116121292,
      "learning_rate": 8.706549808045761e-07,
      "loss": 0.02139887660741806,
      "mean_token_accuracy": 0.9958462625741958,
      "num_tokens": 20156756.0,
      "step": 5640
    },
    {
      "entropy": 2.2255977392196655,
      "epoch": 4.359776104999035,
      "grad_norm": 0.27150967717170715,
      "learning_rate": 8.503697027543368e-07,
      "loss": 0.004352736845612526,
      "mean_token_accuracy": 0.9980769231915474,
      "num_tokens": 20192296.0,
      "step": 5650
    },
    {
      "entropy": 2.22255657017231,
      "epoch": 4.367496622273692,
      "grad_norm": 0.11026787757873535,
      "learning_rate": 8.303130329141629e-07,
      "loss": 0.0016281832009553908,
      "mean_token_accuracy": 1.0,
      "num_tokens": 20226479.0,
      "step": 5660
    },
    {
      "entropy": 2.246255025267601,
      "epoch": 4.37521713954835,
      "grad_norm": 0.46250292658805847,
      "learning_rate": 8.104854724102495e-07,
      "loss": 0.011007238179445267,
      "mean_token_accuracy": 0.9953571423888207,
      "num_tokens": 20270096.0,
      "step": 5670
    },
    {
      "entropy": 2.1503868639469146,
      "epoch": 4.382937656823007,
      "grad_norm": 0.04919159784913063,
      "learning_rate": 7.908875166443741e-07,
      "loss": 0.004710817709565163,
      "mean_token_accuracy": 0.9985272988677025,
      "num_tokens": 20305299.0,
      "step": 5680
    },
    {
      "entropy": 2.2465969413518905,
      "epoch": 4.3906581740976645,
      "grad_norm": 0.17782022058963776,
      "learning_rate": 7.715196552815219e-07,
      "loss": 0.003526587411761284,
      "mean_token_accuracy": 0.9982142850756646,
      "num_tokens": 20342660.0,
      "step": 5690
    },
    {
      "entropy": 2.197134277224541,
      "epoch": 4.398378691372322,
      "grad_norm": 0.2121172547340393,
      "learning_rate": 7.5238237223765e-07,
      "loss": 0.004980428516864777,
      "mean_token_accuracy": 0.9987499997019768,
      "num_tokens": 20375889.0,
      "step": 5700
    },
    {
      "entropy": 2.220217686891556,
      "epoch": 4.406099208646979,
      "grad_norm": 0.34957343339920044,
      "learning_rate": 7.334761456676021e-07,
      "loss": 0.004869958385825157,
      "mean_token_accuracy": 0.9981249988079071,
      "num_tokens": 20410586.0,
      "step": 5710
    },
    {
      "entropy": 2.244478589296341,
      "epoch": 4.413819725921637,
      "grad_norm": 0.10005008429288864,
      "learning_rate": 7.148014479531551e-07,
      "loss": 0.003159433975815773,
      "mean_token_accuracy": 0.998863635957241,
      "num_tokens": 20445560.0,
      "step": 5720
    },
    {
      "entropy": 2.2202999114990236,
      "epoch": 4.4215402431962945,
      "grad_norm": 0.05696292221546173,
      "learning_rate": 6.963587456912157e-07,
      "loss": 0.003368407115340233,
      "mean_token_accuracy": 0.99921875,
      "num_tokens": 20479977.0,
      "step": 5730
    },
    {
      "entropy": 2.2504466265439986,
      "epoch": 4.429260760470951,
      "grad_norm": 0.15865400433540344,
      "learning_rate": 6.781484996821697e-07,
      "loss": 0.003355879709124565,
      "mean_token_accuracy": 0.9971590921282768,
      "num_tokens": 20515597.0,
      "step": 5740
    },
    {
      "entropy": 2.2659272760152818,
      "epoch": 4.436981277745609,
      "grad_norm": 0.07037786394357681,
      "learning_rate": 6.601711649183628e-07,
      "loss": 0.0016187068074941635,
      "mean_token_accuracy": 1.0,
      "num_tokens": 20552502.0,
      "step": 5750
    },
    {
      "entropy": 2.3120163559913633,
      "epoch": 4.444701795020267,
      "grad_norm": 0.19772270321846008,
      "learning_rate": 6.42427190572732e-07,
      "loss": 0.0062440957874059675,
      "mean_token_accuracy": 0.9991666659712791,
      "num_tokens": 20586728.0,
      "step": 5760
    },
    {
      "entropy": 2.258056941628456,
      "epoch": 4.452422312294924,
      "grad_norm": 0.2031521052122116,
      "learning_rate": 6.249170199875832e-07,
      "loss": 0.006161601468920708,
      "mean_token_accuracy": 0.9992647051811219,
      "num_tokens": 20620711.0,
      "step": 5770
    },
    {
      "entropy": 2.259738165140152,
      "epoch": 4.460142829569581,
      "grad_norm": 0.02412845566868782,
      "learning_rate": 6.07641090663521e-07,
      "loss": 0.0021015578880906106,
      "mean_token_accuracy": 0.9991071432828903,
      "num_tokens": 20651230.0,
      "step": 5780
    },
    {
      "entropy": 2.247869810461998,
      "epoch": 4.467863346844238,
      "grad_norm": 0.11743807792663574,
      "learning_rate": 5.905998342485086e-07,
      "loss": 0.008077496290206909,
      "mean_token_accuracy": 0.998356680572033,
      "num_tokens": 20683130.0,
      "step": 5790
    },
    {
      "entropy": 2.249860090017319,
      "epoch": 4.475583864118896,
      "grad_norm": 4.197149276733398,
      "learning_rate": 5.737936765270813e-07,
      "loss": 0.011671672016382218,
      "mean_token_accuracy": 0.9986111104488373,
      "num_tokens": 20722216.0,
      "step": 5800
    },
    {
      "entropy": 2.2498649269342423,
      "epoch": 4.483304381393554,
      "grad_norm": 0.130318745970726,
      "learning_rate": 5.572230374097231e-07,
      "loss": 0.004165193811058998,
      "mean_token_accuracy": 0.9988095238804817,
      "num_tokens": 20758556.0,
      "step": 5810
    },
    {
      "entropy": 2.172873851656914,
      "epoch": 4.49102489866821,
      "grad_norm": 0.1512354016304016,
      "learning_rate": 5.408883309223556e-07,
      "loss": 0.010192253440618516,
      "mean_token_accuracy": 0.997826087474823,
      "num_tokens": 20797172.0,
      "step": 5820
    },
    {
      "entropy": 2.2314586699008943,
      "epoch": 4.498745415942868,
      "grad_norm": 0.23559877276420593,
      "learning_rate": 5.247899651960064e-07,
      "loss": 0.009908857196569443,
      "mean_token_accuracy": 0.9950735285878182,
      "num_tokens": 20832728.0,
      "step": 5830
    },
    {
      "entropy": 2.219519907236099,
      "epoch": 4.506465933217526,
      "grad_norm": 0.054366808384656906,
      "learning_rate": 5.089283424566094e-07,
      "loss": 0.016809307038784027,
      "mean_token_accuracy": 0.9969444438815117,
      "num_tokens": 20868737.0,
      "step": 5840
    },
    {
      "entropy": 2.258474165201187,
      "epoch": 4.514186450492183,
      "grad_norm": 0.20110376179218292,
      "learning_rate": 4.933038590149508e-07,
      "loss": 0.0023351117968559267,
      "mean_token_accuracy": 1.0,
      "num_tokens": 20904237.0,
      "step": 5850
    },
    {
      "entropy": 2.2490910053253175,
      "epoch": 4.52190696776684,
      "grad_norm": 0.06749659776687622,
      "learning_rate": 4.77916905256769e-07,
      "loss": 0.0015241201967000962,
      "mean_token_accuracy": 1.0,
      "num_tokens": 20938015.0,
      "step": 5860
    },
    {
      "entropy": 2.2541816264390944,
      "epoch": 4.529627485041498,
      "grad_norm": 0.07041777670383453,
      "learning_rate": 4.627678656330026e-07,
      "loss": 0.003881530836224556,
      "mean_token_accuracy": 0.9989999994635582,
      "num_tokens": 20971752.0,
      "step": 5870
    },
    {
      "entropy": 2.225889188051224,
      "epoch": 4.537348002316155,
      "grad_norm": 0.04327480122447014,
      "learning_rate": 4.478571186501857e-07,
      "loss": 0.013946136832237244,
      "mean_token_accuracy": 0.9973797842860221,
      "num_tokens": 21005854.0,
      "step": 5880
    },
    {
      "entropy": 2.2675861060619353,
      "epoch": 4.545068519590813,
      "grad_norm": 1.7059868574142456,
      "learning_rate": 4.3318503686098577e-07,
      "loss": 0.0015708981081843376,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21041256.0,
      "step": 5890
    },
    {
      "entropy": 2.2442140936851502,
      "epoch": 4.5527890368654695,
      "grad_norm": 0.06668271869421005,
      "learning_rate": 4.1875198685489613e-07,
      "loss": 0.0081165112555027,
      "mean_token_accuracy": 0.9988095238804817,
      "num_tokens": 21073826.0,
      "step": 5900
    },
    {
      "entropy": 2.238309305906296,
      "epoch": 4.560509554140127,
      "grad_norm": 0.022307725623250008,
      "learning_rate": 4.04558329249084e-07,
      "loss": 0.006697001308202744,
      "mean_token_accuracy": 0.998863635957241,
      "num_tokens": 21110342.0,
      "step": 5910
    },
    {
      "entropy": 2.200860995054245,
      "epoch": 4.568230071414785,
      "grad_norm": 0.17545939981937408,
      "learning_rate": 3.906044186793689e-07,
      "loss": 0.004396386817097664,
      "mean_token_accuracy": 0.9992647051811219,
      "num_tokens": 21147596.0,
      "step": 5920
    },
    {
      "entropy": 2.24030060172081,
      "epoch": 4.575950588689442,
      "grad_norm": 0.06349517405033112,
      "learning_rate": 3.768906037913689e-07,
      "loss": 0.010619433969259262,
      "mean_token_accuracy": 0.9960175439715385,
      "num_tokens": 21181840.0,
      "step": 5930
    },
    {
      "entropy": 2.1731992214918137,
      "epoch": 4.5836711059640995,
      "grad_norm": 0.07084771990776062,
      "learning_rate": 3.6341722723179264e-07,
      "loss": 0.007943939417600632,
      "mean_token_accuracy": 0.9991379305720329,
      "num_tokens": 21215224.0,
      "step": 5940
    },
    {
      "entropy": 2.2863466501235963,
      "epoch": 4.591391623238757,
      "grad_norm": 0.044842347502708435,
      "learning_rate": 3.501846256398689e-07,
      "loss": 0.013139717280864716,
      "mean_token_accuracy": 0.9968560606241226,
      "num_tokens": 21248186.0,
      "step": 5950
    },
    {
      "entropy": 2.208678925037384,
      "epoch": 4.599112140513414,
      "grad_norm": 0.14837071299552917,
      "learning_rate": 3.371931296389397e-07,
      "loss": 0.0007539391051977872,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21282267.0,
      "step": 5960
    },
    {
      "entropy": 2.2153519988059998,
      "epoch": 4.606832657788072,
      "grad_norm": 2.5055654048919678,
      "learning_rate": 3.2444306382820256e-07,
      "loss": 0.03023122251033783,
      "mean_token_accuracy": 0.9947824060916901,
      "num_tokens": 21315323.0,
      "step": 5970
    },
    {
      "entropy": 2.2592311084270476,
      "epoch": 4.6145531750627296,
      "grad_norm": 0.1265610009431839,
      "learning_rate": 3.119347467745992e-07,
      "loss": 0.001410143543034792,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21353041.0,
      "step": 5980
    },
    {
      "entropy": 2.2527008682489393,
      "epoch": 4.622273692337386,
      "grad_norm": 0.589449405670166,
      "learning_rate": 2.99668491004852e-07,
      "loss": 0.005680891871452332,
      "mean_token_accuracy": 0.9983333334326744,
      "num_tokens": 21386776.0,
      "step": 5990
    },
    {
      "entropy": 2.2103109836578367,
      "epoch": 4.629994209612044,
      "grad_norm": 1.9735628366470337,
      "learning_rate": 2.876446029976554e-07,
      "loss": 0.005637718737125397,
      "mean_token_accuracy": 0.9986278191208839,
      "num_tokens": 21422729.0,
      "step": 6000
    },
    {
      "epoch": 4.629994209612044,
      "eval_entropy": 2.1154677377100897,
      "eval_loss": 0.5169965028762817,
      "eval_mean_token_accuracy": 0.9358510735640332,
      "eval_num_tokens": 21422729.0,
      "eval_runtime": 988.4165,
      "eval_samples_per_second": 7.741,
      "eval_steps_per_second": 0.968,
      "step": 6000
    },
    {
      "entropy": 2.227972522377968,
      "epoch": 4.637714726886701,
      "grad_norm": 1.7792558670043945,
      "learning_rate": 2.7586338317602713e-07,
      "loss": 0.006323382258415222,
      "mean_token_accuracy": 0.99921875,
      "num_tokens": 21457481.0,
      "step": 6010
    },
    {
      "entropy": 2.286092558503151,
      "epoch": 4.645435244161359,
      "grad_norm": 0.7590512037277222,
      "learning_rate": 2.6432512589979144e-07,
      "loss": 0.006863045692443848,
      "mean_token_accuracy": 0.998242424428463,
      "num_tokens": 21491657.0,
      "step": 6020
    },
    {
      "entropy": 2.1695028364658357,
      "epoch": 4.653155761436016,
      "grad_norm": 3.8523592948913574,
      "learning_rate": 2.530301194582296e-07,
      "loss": 0.005904960632324219,
      "mean_token_accuracy": 0.9977777764201164,
      "num_tokens": 21531112.0,
      "step": 6030
    },
    {
      "entropy": 2.2056106328964233,
      "epoch": 4.660876278710674,
      "grad_norm": 0.10862927883863449,
      "learning_rate": 2.41978646062877e-07,
      "loss": 0.0022114580497145653,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21567253.0,
      "step": 6040
    },
    {
      "entropy": 2.2164981573820115,
      "epoch": 4.668596795985331,
      "grad_norm": 0.3193495273590088,
      "learning_rate": 2.3117098184046837e-07,
      "loss": 0.007381512224674225,
      "mean_token_accuracy": 0.9967272713780403,
      "num_tokens": 21608628.0,
      "step": 6050
    },
    {
      "entropy": 2.1946250647306442,
      "epoch": 4.676317313259989,
      "grad_norm": 0.02586934342980385,
      "learning_rate": 2.206073968260436e-07,
      "loss": 0.006831285357475281,
      "mean_token_accuracy": 0.9986111104488373,
      "num_tokens": 21644056.0,
      "step": 6060
    },
    {
      "entropy": 2.2137088149785997,
      "epoch": 4.6840378305346455,
      "grad_norm": 0.05132552608847618,
      "learning_rate": 2.1028815495619526e-07,
      "loss": 0.0044960763305425646,
      "mean_token_accuracy": 0.999074074625969,
      "num_tokens": 21681633.0,
      "step": 6070
    },
    {
      "entropy": 2.2435118734836577,
      "epoch": 4.691758347809303,
      "grad_norm": 0.0880841463804245,
      "learning_rate": 2.002135140624828e-07,
      "loss": 0.010966504365205765,
      "mean_token_accuracy": 0.9978095233440399,
      "num_tokens": 21720250.0,
      "step": 6080
    },
    {
      "entropy": 2.2539133608341215,
      "epoch": 4.699478865083961,
      "grad_norm": 0.03336416557431221,
      "learning_rate": 1.9038372586497766e-07,
      "loss": 0.0027033800259232523,
      "mean_token_accuracy": 0.9980000004172325,
      "num_tokens": 21756172.0,
      "step": 6090
    },
    {
      "entropy": 2.171377721428871,
      "epoch": 4.707199382358618,
      "grad_norm": 0.02603979967534542,
      "learning_rate": 1.8079903596598504e-07,
      "loss": 0.015214832127094268,
      "mean_token_accuracy": 0.9969905868172646,
      "num_tokens": 21797144.0,
      "step": 6100
    },
    {
      "entropy": 2.2282418727874758,
      "epoch": 4.7149198996332755,
      "grad_norm": 0.02482684515416622,
      "learning_rate": 1.714596838439031e-07,
      "loss": 0.0006937965285032987,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21833771.0,
      "step": 6110
    },
    {
      "entropy": 2.271139058470726,
      "epoch": 4.722640416907932,
      "grad_norm": 0.041452571749687195,
      "learning_rate": 1.6236590284723796e-07,
      "loss": 0.0011482076719403268,
      "mean_token_accuracy": 1.0,
      "num_tokens": 21866597.0,
      "step": 6120
    },
    {
      "entropy": 2.179414650797844,
      "epoch": 4.73036093418259,
      "grad_norm": 0.09968980401754379,
      "learning_rate": 1.535179201887782e-07,
      "loss": 0.008539145439863205,
      "mean_token_accuracy": 0.9981481477618217,
      "num_tokens": 21901353.0,
      "step": 6130
    },
    {
      "entropy": 2.2338360488414764,
      "epoch": 4.738081451457248,
      "grad_norm": 3.942858934402466,
      "learning_rate": 1.4491595693991056e-07,
      "loss": 0.008534059673547745,
      "mean_token_accuracy": 0.9989583328366279,
      "num_tokens": 21936884.0,
      "step": 6140
    },
    {
      "entropy": 2.24261232316494,
      "epoch": 4.7458019687319055,
      "grad_norm": 0.16906267404556274,
      "learning_rate": 1.3656022802510326e-07,
      "loss": 0.018931737542152403,
      "mean_token_accuracy": 0.9983006909489631,
      "num_tokens": 21971041.0,
      "step": 6150
    },
    {
      "entropy": 2.305382552742958,
      "epoch": 4.753522486006562,
      "grad_norm": 0.04022558033466339,
      "learning_rate": 1.2845094221653475e-07,
      "loss": 0.006191138923168182,
      "mean_token_accuracy": 0.9983004376292228,
      "num_tokens": 22001583.0,
      "step": 6160
    },
    {
      "entropy": 2.304952707886696,
      "epoch": 4.76124300328122,
      "grad_norm": 0.0716034322977066,
      "learning_rate": 1.2058830212887117e-07,
      "loss": 0.014426815509796142,
      "mean_token_accuracy": 0.9962768808007241,
      "num_tokens": 22036327.0,
      "step": 6170
    },
    {
      "entropy": 2.2286325991153717,
      "epoch": 4.768963520555877,
      "grad_norm": 0.08237482607364655,
      "learning_rate": 1.129725042142138e-07,
      "loss": 0.006697164475917816,
      "mean_token_accuracy": 0.9982928246259689,
      "num_tokens": 22071393.0,
      "step": 6180
    },
    {
      "entropy": 2.2289715498685836,
      "epoch": 4.776684037830535,
      "grad_norm": 0.09096929430961609,
      "learning_rate": 1.0560373875718177e-07,
      "loss": 0.0037596296519041062,
      "mean_token_accuracy": 0.9989999994635582,
      "num_tokens": 22105599.0,
      "step": 6190
    },
    {
      "entropy": 2.226081871986389,
      "epoch": 4.784404555105192,
      "grad_norm": 1.364410400390625,
      "learning_rate": 9.848218987016267e-08,
      "loss": 0.004093809798359871,
      "mean_token_accuracy": 0.9989999994635582,
      "num_tokens": 22144292.0,
      "step": 6200
    },
    {
      "entropy": 2.3211822539567946,
      "epoch": 4.792125072379849,
      "grad_norm": 0.08333883434534073,
      "learning_rate": 9.160803548871277e-08,
      "loss": 0.0013872952200472354,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22173297.0,
      "step": 6210
    },
    {
      "entropy": 2.181807669997215,
      "epoch": 4.799845589654507,
      "grad_norm": 0.08239463716745377,
      "learning_rate": 8.498144736710734e-08,
      "loss": 0.024519316852092743,
      "mean_token_accuracy": 0.99544607847929,
      "num_tokens": 22213011.0,
      "step": 6220
    },
    {
      "entropy": 2.1619095504283905,
      "epoch": 4.807566106929165,
      "grad_norm": 0.39841601252555847,
      "learning_rate": 7.860259107405288e-08,
      "loss": 0.0010081185027956963,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22247236.0,
      "step": 6230
    },
    {
      "entropy": 2.2368269830942156,
      "epoch": 4.8152866242038215,
      "grad_norm": 0.07764973491430283,
      "learning_rate": 7.24716259885483e-08,
      "loss": 0.026798546314239502,
      "mean_token_accuracy": 0.9924574360251427,
      "num_tokens": 22288494.0,
      "step": 6240
    },
    {
      "entropy": 2.245153045654297,
      "epoch": 4.823007141478479,
      "grad_norm": 0.26748162508010864,
      "learning_rate": 6.658870529590355e-08,
      "loss": 0.00493730865418911,
      "mean_token_accuracy": 0.999074074625969,
      "num_tokens": 22326075.0,
      "step": 6250
    },
    {
      "entropy": 2.219322806596756,
      "epoch": 4.830727658753137,
      "grad_norm": 0.04285896196961403,
      "learning_rate": 6.095397598391284e-08,
      "loss": 0.008515120297670365,
      "mean_token_accuracy": 0.9957702025771141,
      "num_tokens": 22368028.0,
      "step": 6260
    },
    {
      "entropy": 2.247615697979927,
      "epoch": 4.838448176027794,
      "grad_norm": 1.9107056856155396,
      "learning_rate": 5.556757883917851e-08,
      "loss": 0.0032505668699741364,
      "mean_token_accuracy": 0.9991379305720329,
      "num_tokens": 22406097.0,
      "step": 6270
    },
    {
      "entropy": 2.205378809571266,
      "epoch": 4.8461686933024515,
      "grad_norm": 2.558126211166382,
      "learning_rate": 5.0429648443601765e-08,
      "loss": 0.01382438838481903,
      "mean_token_accuracy": 0.9963152587413788,
      "num_tokens": 22447091.0,
      "step": 6280
    },
    {
      "entropy": 2.1460685402154924,
      "epoch": 4.853889210577108,
      "grad_norm": 0.6871592998504639,
      "learning_rate": 4.55403131710086e-08,
      "loss": 0.0034082386642694472,
      "mean_token_accuracy": 0.9983333334326744,
      "num_tokens": 22482163.0,
      "step": 6290
    },
    {
      "entropy": 2.1781392335891723,
      "epoch": 4.861609727851766,
      "grad_norm": 0.03307396173477173,
      "learning_rate": 4.089969518395243e-08,
      "loss": 0.0009966903366148473,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22521045.0,
      "step": 6300
    },
    {
      "entropy": 2.276013654470444,
      "epoch": 4.869330245126424,
      "grad_norm": 0.35160934925079346,
      "learning_rate": 3.650791043065427e-08,
      "loss": 0.0020426543429493903,
      "mean_token_accuracy": 0.999074074625969,
      "num_tokens": 22554853.0,
      "step": 6310
    },
    {
      "entropy": 2.253162696957588,
      "epoch": 4.877050762401081,
      "grad_norm": 0.1488099992275238,
      "learning_rate": 3.236506864211286e-08,
      "loss": 0.007266855984926224,
      "mean_token_accuracy": 0.9969222694635391,
      "num_tokens": 22590884.0,
      "step": 6320
    },
    {
      "entropy": 2.27145434319973,
      "epoch": 4.884771279675738,
      "grad_norm": 3.7629544734954834,
      "learning_rate": 2.847127332935573e-08,
      "loss": 0.01586032509803772,
      "mean_token_accuracy": 0.995819628238678,
      "num_tokens": 22623671.0,
      "step": 6330
    },
    {
      "entropy": 2.236271634697914,
      "epoch": 4.892491796950396,
      "grad_norm": 0.6222274303436279,
      "learning_rate": 2.482662178085682e-08,
      "loss": 0.012338303774595261,
      "mean_token_accuracy": 0.9965226069092751,
      "num_tokens": 22658331.0,
      "step": 6340
    },
    {
      "entropy": 2.197663214802742,
      "epoch": 4.900212314225053,
      "grad_norm": 0.11816233396530151,
      "learning_rate": 2.143120506010843e-08,
      "loss": 0.008864825218915939,
      "mean_token_accuracy": 0.9983317211270333,
      "num_tokens": 22697599.0,
      "step": 6350
    },
    {
      "entropy": 2.2309891253709795,
      "epoch": 4.907932831499711,
      "grad_norm": 1.5716471672058105,
      "learning_rate": 1.8285108003338603e-08,
      "loss": 0.007599858939647675,
      "mean_token_accuracy": 0.9981518805027008,
      "num_tokens": 22730261.0,
      "step": 6360
    },
    {
      "entropy": 2.225965863466263,
      "epoch": 4.915653348774368,
      "grad_norm": 0.03380768746137619,
      "learning_rate": 1.538840921739837e-08,
      "loss": 0.0017374027520418167,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22768240.0,
      "step": 6370
    },
    {
      "entropy": 2.2355996876955033,
      "epoch": 4.923373866049025,
      "grad_norm": 0.04010441154241562,
      "learning_rate": 1.2741181077793319e-08,
      "loss": 0.0016989083960652351,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22803213.0,
      "step": 6380
    },
    {
      "entropy": 2.2543944001197813,
      "epoch": 4.931094383323683,
      "grad_norm": 0.22062981128692627,
      "learning_rate": 1.0343489726878375e-08,
      "loss": 0.008954495936632157,
      "mean_token_accuracy": 0.9964285716414452,
      "num_tokens": 22833895.0,
      "step": 6390
    },
    {
      "entropy": 2.2120563268661497,
      "epoch": 4.93881490059834,
      "grad_norm": 0.09145352989435196,
      "learning_rate": 8.195395072201352e-09,
      "loss": 0.0006910903844982386,
      "mean_token_accuracy": 1.0,
      "num_tokens": 22872270.0,
      "step": 6400
    },
    {
      "entropy": 2.263408675789833,
      "epoch": 4.946535417872997,
      "grad_norm": 0.17131362855434418,
      "learning_rate": 6.2969507850108025e-09,
      "loss": 0.007691174000501633,
      "mean_token_accuracy": 0.998504464328289,
      "num_tokens": 22906940.0,
      "step": 6410
    },
    {
      "entropy": 2.201318547129631,
      "epoch": 4.954255935147655,
      "grad_norm": 2.934361696243286,
      "learning_rate": 4.648204298912662e-09,
      "loss": 0.005105702206492424,
      "mean_token_accuracy": 0.9974216759204865,
      "num_tokens": 22943748.0,
      "step": 6420
    },
    {
      "entropy": 2.2332019776105883,
      "epoch": 4.961976452422312,
      "grad_norm": 0.023339014500379562,
      "learning_rate": 3.249196808683408e-09,
      "loss": 0.008095689117908478,
      "mean_token_accuracy": 0.9983243241906166,
      "num_tokens": 22977903.0,
      "step": 6430
    },
    {
      "entropy": 2.242179808020592,
      "epoch": 4.96969696969697,
      "grad_norm": 0.043394219130277634,
      "learning_rate": 2.0999632692431104e-09,
      "loss": 0.0013438466936349868,
      "mean_token_accuracy": 1.0,
      "num_tokens": 23011768.0,
      "step": 6440
    },
    {
      "entropy": 2.2419181168079376,
      "epoch": 4.9774174869716274,
      "grad_norm": 0.046509750187397,
      "learning_rate": 1.2005323947816839e-09,
      "loss": 0.008913324773311615,
      "mean_token_accuracy": 0.9981812179088593,
      "num_tokens": 23047974.0,
      "step": 6450
    },
    {
      "entropy": 2.259080970287323,
      "epoch": 4.985138004246284,
      "grad_norm": 2.01216197013855,
      "learning_rate": 5.509266580416839e-10,
      "loss": 0.0047379877418279644,
      "mean_token_accuracy": 0.9958045974373817,
      "num_tokens": 23081068.0,
      "step": 6460
    },
    {
      "entropy": 2.214487534761429,
      "epoch": 4.992858521520942,
      "grad_norm": 1.7724125385284424,
      "learning_rate": 1.5116228975653457e-10,
      "loss": 0.007440639287233352,
      "mean_token_accuracy": 0.9989130437374115,
      "num_tokens": 23120802.0,
      "step": 6470
    },
    {
      "entropy": 2.282172141848384,
      "epoch": 5.0,
      "grad_norm": 0.03656819462776184,
      "learning_rate": 1.2492782430761908e-12,
      "loss": 0.006634878367185593,
      "mean_token_accuracy": 0.9977477482847266,
      "num_tokens": 23154615.0,
      "step": 6480
    },
    {
      "epoch": 5.0,
      "eval_entropy": 2.1159368567830468,
      "eval_loss": 0.517891526222229,
      "eval_mean_token_accuracy": 0.9358265156407092,
      "eval_num_tokens": 23154615.0,
      "eval_runtime": 988.2948,
      "eval_samples_per_second": 7.742,
      "eval_steps_per_second": 0.968,
      "step": 6480
    },
    {
      "epoch": 5.0,
      "step": 6480,
      "total_flos": 1.3561958615698637e+18,
      "train_loss": 0.07709400485848927,
      "train_runtime": 23143.6512,
      "train_samples_per_second": 2.239,
      "train_steps_per_second": 0.28
    }
  ],
  "logging_steps": 10,
  "max_steps": 6480,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.3561958615698637e+18,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}