{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 13.424242424242424,
  "eval_steps": 100,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.9522146363556385,
      "epoch": 0.16835016835016836,
      "grad_norm": 0.376953125,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.9937530517578125,
      "mean_token_accuracy": 0.7483265110850335,
      "num_tokens": 1591793.0,
      "step": 25
    },
    {
      "entropy": 0.3762529502809048,
      "epoch": 0.3367003367003367,
      "grad_norm": 0.228515625,
      "learning_rate": 1.6333333333333335e-05,
      "loss": 0.3625458145141602,
      "mean_token_accuracy": 0.8780256441235542,
      "num_tokens": 3174240.0,
      "step": 50
    },
    {
      "entropy": 0.3107199192792177,
      "epoch": 0.5050505050505051,
      "grad_norm": 0.154296875,
      "learning_rate": 1.9997430173759876e-05,
      "loss": 0.30434131622314453,
      "mean_token_accuracy": 0.8953498190641404,
      "num_tokens": 4755519.0,
      "step": 75
    },
    {
      "entropy": 0.2847785219550133,
      "epoch": 0.6734006734006734,
      "grad_norm": 0.2001953125,
      "learning_rate": 1.9980063397715685e-05,
      "loss": 0.28054773330688476,
      "mean_token_accuracy": 0.9025071159005165,
      "num_tokens": 6334024.0,
      "step": 100
    },
    {
      "epoch": 0.6734006734006734,
      "eval_entropy": 0.2793949246406555,
      "eval_loss": 0.2698107361793518,
      "eval_mean_token_accuracy": 0.9054524502754211,
      "eval_num_tokens": 6334024.0,
      "eval_runtime": 14.0488,
      "eval_samples_per_second": 17.795,
      "eval_steps_per_second": 8.898,
      "eval_token_accuracy": 0.9051616133787281,
      "step": 100
    },
    {
      "entropy": 0.26899013079702855,
      "epoch": 0.8417508417508418,
      "grad_norm": 0.2001953125,
      "learning_rate": 1.9946341631587086e-05,
      "loss": 0.26340150833129883,
      "mean_token_accuracy": 0.9085881695151329,
      "num_tokens": 7921274.0,
      "step": 125
    },
    {
      "entropy": 0.24236563958075583,
      "epoch": 1.0067340067340067,
      "grad_norm": 0.2109375,
      "learning_rate": 1.9896320137462984e-05,
      "loss": 0.23591676712036133,
      "mean_token_accuracy": 0.9202974551186269,
      "num_tokens": 9458752.0,
      "step": 150
    },
    {
      "entropy": 0.19062039852142335,
      "epoch": 1.1750841750841752,
      "grad_norm": 0.1513671875,
      "learning_rate": 1.983008088887478e-05,
      "loss": 0.18400527954101562,
      "mean_token_accuracy": 0.9372689473628998,
      "num_tokens": 11049146.0,
      "step": 175
    },
    {
      "entropy": 0.17757252234965562,
      "epoch": 1.3434343434343434,
      "grad_norm": 0.1455078125,
      "learning_rate": 1.9747732436460955e-05,
      "loss": 0.17247369766235351,
      "mean_token_accuracy": 0.9412289550900459,
      "num_tokens": 12629069.0,
      "step": 200
    },
    {
      "epoch": 1.3434343434343434,
      "eval_entropy": 0.17348989713191987,
      "eval_loss": 0.1803748607635498,
      "eval_mean_token_accuracy": 0.9387603945732117,
      "eval_num_tokens": 12629069.0,
      "eval_runtime": 13.9222,
      "eval_samples_per_second": 17.957,
      "eval_steps_per_second": 8.978,
      "eval_token_accuracy": 0.9386888733819055,
      "step": 200
    },
    {
      "entropy": 0.17212937109172344,
      "epoch": 1.5117845117845117,
      "grad_norm": 0.16015625,
      "learning_rate": 1.9649409730077934e-05,
      "loss": 0.1680605125427246,
      "mean_token_accuracy": 0.9425190502405166,
      "num_tokens": 14207351.0,
      "step": 225
    },
    {
      "entropy": 0.16998517245054245,
      "epoch": 1.6801346801346801,
      "grad_norm": 0.166015625,
      "learning_rate": 1.9535273897648857e-05,
      "loss": 0.1662980079650879,
      "mean_token_accuracy": 0.943454926609993,
      "num_tokens": 15787859.0,
      "step": 250
    },
    {
      "entropy": 0.16503205563873052,
      "epoch": 1.8484848484848486,
      "grad_norm": 0.1552734375,
      "learning_rate": 1.9405511981112553e-05,
      "loss": 0.16160566329956055,
      "mean_token_accuracy": 0.9454404374957085,
      "num_tokens": 17374190.0,
      "step": 275
    },
    {
      "entropy": 0.1615225655615938,
      "epoch": 2.0134680134680134,
      "grad_norm": 0.169921875,
      "learning_rate": 1.926033662990558e-05,
      "loss": 0.15662331581115724,
      "mean_token_accuracy": 0.9472035674416289,
      "num_tokens": 18919416.0,
      "step": 300
    },
    {
      "epoch": 2.0134680134680134,
      "eval_entropy": 0.1541357452273369,
      "eval_loss": 0.1643277257680893,
      "eval_mean_token_accuracy": 0.9455152935981751,
      "eval_num_tokens": 18919416.0,
      "eval_runtime": 13.8817,
      "eval_samples_per_second": 18.009,
      "eval_steps_per_second": 9.005,
      "eval_token_accuracy": 0.9454900486056786,
      "step": 300
    },
    {
      "entropy": 0.14235603258013727,
      "epoch": 2.1818181818181817,
      "grad_norm": 0.1787109375,
      "learning_rate": 1.9099985752479505e-05,
      "loss": 0.13640413284301758,
      "mean_token_accuracy": 0.9537694716453552,
      "num_tokens": 20493589.0,
      "step": 325
    },
    {
      "entropy": 0.1364088713005185,
      "epoch": 2.3501683501683504,
      "grad_norm": 0.1767578125,
      "learning_rate": 1.892472212642459e-05,
      "loss": 0.13120804786682128,
      "mean_token_accuracy": 0.9558958530426025,
      "num_tokens": 22071046.0,
      "step": 350
    },
    {
      "entropy": 0.1340012726187706,
      "epoch": 2.5185185185185186,
      "grad_norm": 0.1728515625,
      "learning_rate": 1.8734832967838775e-05,
      "loss": 0.12886263847351073,
      "mean_token_accuracy": 0.956660538315773,
      "num_tokens": 23651892.0,
      "step": 375
    },
    {
      "entropy": 0.13179550088942052,
      "epoch": 2.686868686868687,
      "grad_norm": 0.1572265625,
      "learning_rate": 1.8530629460647658e-05,
      "loss": 0.1262987232208252,
      "mean_token_accuracy": 0.9577056321501732,
      "num_tokens": 25237419.0,
      "step": 400
    },
    {
      "epoch": 2.686868686868687,
      "eval_entropy": 0.13389861524105073,
      "eval_loss": 0.1478790044784546,
      "eval_mean_token_accuracy": 0.9525656714439392,
      "eval_num_tokens": 25237419.0,
      "eval_runtime": 14.0025,
      "eval_samples_per_second": 17.854,
      "eval_steps_per_second": 8.927,
      "eval_token_accuracy": 0.9525425978397923,
      "step": 400
    },
    {
      "entropy": 0.12678295791149138,
      "epoch": 2.855218855218855,
      "grad_norm": 0.1796875,
      "learning_rate": 1.831244624664681e-05,
      "loss": 0.12145834922790527,
      "mean_token_accuracy": 0.9593311008810997,
      "num_tokens": 26831783.0,
      "step": 425
    },
    {
      "entropy": 0.12364728927460251,
      "epoch": 3.0202020202020203,
      "grad_norm": 0.1796875,
      "learning_rate": 1.808064087710212e-05,
      "loss": 0.11712233543395996,
      "mean_token_accuracy": 0.9609322833771609,
      "num_tokens": 28376572.0,
      "step": 450
    },
    {
      "entropy": 0.09804413143545389,
      "epoch": 3.1885521885521886,
      "grad_norm": 0.216796875,
      "learning_rate": 1.7835593226806902e-05,
      "loss": 0.0911950397491455,
      "mean_token_accuracy": 0.9690286010503769,
      "num_tokens": 29961022.0,
      "step": 475
    },
    {
      "entropy": 0.09506628409028053,
      "epoch": 3.356902356902357,
      "grad_norm": 0.1650390625,
      "learning_rate": 1.7577704871555924e-05,
      "loss": 0.08855191230773926,
      "mean_token_accuracy": 0.9698488730192184,
      "num_tokens": 31553946.0,
      "step": 500
    },
    {
      "epoch": 3.356902356902357,
      "eval_entropy": 0.10651444137096405,
      "eval_loss": 0.1412735879421234,
      "eval_mean_token_accuracy": 0.9571833200454712,
      "eval_num_tokens": 31553946.0,
      "eval_runtime": 13.9481,
      "eval_samples_per_second": 17.924,
      "eval_steps_per_second": 8.962,
      "eval_token_accuracy": 0.9571799455825543,
      "step": 500
    },
    {
      "entropy": 0.09585228541865945,
      "epoch": 3.525252525252525,
      "grad_norm": 0.1767578125,
      "learning_rate": 1.7307398430056595e-05,
      "loss": 0.08968615531921387,
      "mean_token_accuracy": 0.9697143504023552,
      "num_tokens": 33135630.0,
      "step": 525
    },
    {
      "entropy": 0.10036869466304779,
      "epoch": 3.6936026936026938,
      "grad_norm": 0.1669921875,
      "learning_rate": 1.7025116871355737e-05,
      "loss": 0.09400833129882813,
      "mean_token_accuracy": 0.9683772554993629,
      "num_tokens": 34707432.0,
      "step": 550
    },
    {
      "entropy": 0.0957243438065052,
      "epoch": 3.861952861952862,
      "grad_norm": 0.1865234375,
      "learning_rate": 1.6731322788916892e-05,
      "loss": 0.08972453117370606,
      "mean_token_accuracy": 0.9695299303531647,
      "num_tokens": 36292817.0,
      "step": 575
    },
    {
      "entropy": 0.0924541620841744,
      "epoch": 4.026936026936027,
      "grad_norm": 0.1943359375,
      "learning_rate": 1.6426497642537826e-05,
      "loss": 0.08610689163208007,
      "mean_token_accuracy": 0.9709198608690378,
      "num_tokens": 37832802.0,
      "step": 600
    },
    {
      "epoch": 4.026936026936027,
      "eval_entropy": 0.08664591646194458,
      "eval_loss": 0.14084948599338531,
      "eval_mean_token_accuracy": 0.9594513201713561,
      "eval_num_tokens": 37832802.0,
      "eval_runtime": 14.1668,
      "eval_samples_per_second": 17.647,
      "eval_steps_per_second": 8.823,
      "eval_token_accuracy": 0.9594704655647771,
      "step": 600
    },
    {
      "entropy": 0.07393988940864801,
      "epoch": 4.1952861952861955,
      "grad_norm": 0.1767578125,
      "learning_rate": 1.6111140969350504e-05,
      "loss": 0.06671238899230957,
      "mean_token_accuracy": 0.9773511955142021,
      "num_tokens": 39420574.0,
      "step": 625
    },
    {
      "entropy": 0.07375508725643158,
      "epoch": 4.363636363636363,
      "grad_norm": 0.1767578125,
      "learning_rate": 1.5785769565196543e-05,
      "loss": 0.06662603855133056,
      "mean_token_accuracy": 0.9770343449711799,
      "num_tokens": 40995243.0,
      "step": 650
    },
    {
      "entropy": 0.0700146171823144,
      "epoch": 4.531986531986532,
      "grad_norm": 0.173828125,
      "learning_rate": 1.5450916637719683e-05,
      "loss": 0.06388366222381592,
      "mean_token_accuracy": 0.9781549483537674,
      "num_tokens": 42585578.0,
      "step": 675
    },
    {
      "entropy": 0.07131696030497552,
      "epoch": 4.700336700336701,
      "grad_norm": 0.166015625,
      "learning_rate": 1.5107130932563151e-05,
      "loss": 0.0648357915878296,
      "mean_token_accuracy": 0.9776528170704841,
      "num_tokens": 44164669.0,
      "step": 700
    },
    {
      "epoch": 4.700336700336701,
      "eval_entropy": 0.08206569665670395,
      "eval_loss": 0.13808086514472961,
      "eval_mean_token_accuracy": 0.9618017344474793,
      "eval_num_tokens": 44164669.0,
      "eval_runtime": 13.9704,
      "eval_samples_per_second": 17.895,
      "eval_steps_per_second": 8.947,
      "eval_token_accuracy": 0.9618092493569853,
      "step": 700
    },
    {
      "entropy": 0.06924619141966104,
      "epoch": 4.8686868686868685,
      "grad_norm": 0.1689453125,
      "learning_rate": 1.4754975834103877e-05,
      "loss": 0.06314377784729004,
      "mean_token_accuracy": 0.9783227249979973,
      "num_tokens": 45746386.0,
      "step": 725
    },
    {
      "entropy": 0.06477328384181066,
      "epoch": 5.033670033670034,
      "grad_norm": 0.1552734375,
      "learning_rate": 1.4395028442197231e-05,
      "loss": 0.057736573219299314,
      "mean_token_accuracy": 0.9802252127199756,
      "num_tokens": 47291064.0,
      "step": 750
    },
    {
      "entropy": 0.05126072809100151,
      "epoch": 5.202020202020202,
      "grad_norm": 0.1611328125,
      "learning_rate": 1.4027878626445339e-05,
      "loss": 0.04461108207702637,
      "mean_token_accuracy": 0.984737733900547,
      "num_tokens": 48874157.0,
      "step": 775
    },
    {
      "entropy": 0.05017553111538291,
      "epoch": 5.37037037037037,
      "grad_norm": 0.1728515625,
      "learning_rate": 1.365412805953872e-05,
      "loss": 0.043688578605651854,
      "mean_token_accuracy": 0.9849283066391945,
      "num_tokens": 50457004.0,
      "step": 800
    },
    {
      "epoch": 5.37037037037037,
      "eval_entropy": 0.0671493958979845,
      "eval_loss": 0.15137650072574615,
      "eval_mean_token_accuracy": 0.9625004653930664,
      "eval_num_tokens": 50457004.0,
      "eval_runtime": 13.8354,
      "eval_samples_per_second": 18.07,
      "eval_steps_per_second": 9.035,
      "eval_token_accuracy": 0.9625030416255251,
      "step": 800
    },
    {
      "entropy": 0.049754718681797386,
      "epoch": 5.538720538720539,
      "grad_norm": 0.138671875,
      "learning_rate": 1.3274389231255466e-05,
      "loss": 0.043588805198669436,
      "mean_token_accuracy": 0.9849756741523743,
      "num_tokens": 52047687.0,
      "step": 825
    },
    {
      "entropy": 0.048861624505370856,
      "epoch": 5.707070707070707,
      "grad_norm": 0.1484375,
      "learning_rate": 1.2889284444733722e-05,
      "loss": 0.04272536277770996,
      "mean_token_accuracy": 0.9852952674031258,
      "num_tokens": 53626751.0,
      "step": 850
    },
    {
      "entropy": 0.049314973950386046,
      "epoch": 5.875420875420875,
      "grad_norm": 0.3671875,
      "learning_rate": 1.2499444796662354e-05,
      "loss": 0.04330784320831299,
      "mean_token_accuracy": 0.9850323754549026,
      "num_tokens": 55212090.0,
      "step": 875
    },
    {
      "entropy": 0.04571919083329184,
      "epoch": 6.040404040404041,
      "grad_norm": 0.140625,
      "learning_rate": 1.2105509143061072e-05,
      "loss": 0.03899869203567505,
      "mean_token_accuracy": 0.9866329042278991,
      "num_tokens": 56756053.0,
      "step": 900
    },
    {
      "epoch": 6.040404040404041,
      "eval_entropy": 0.055020518347620964,
      "eval_loss": 0.1728222817182541,
      "eval_mean_token_accuracy": 0.9626847972869873,
      "eval_num_tokens": 56756053.0,
      "eval_runtime": 13.8671,
      "eval_samples_per_second": 18.028,
      "eval_steps_per_second": 9.014,
      "eval_token_accuracy": 0.9626900638892185,
      "step": 900
    },
    {
      "entropy": 0.034643625281751154,
      "epoch": 6.2087542087542085,
      "grad_norm": 0.12109375,
      "learning_rate": 1.1708123052344803e-05,
      "loss": 0.028563168048858643,
      "mean_token_accuracy": 0.9902752894163132,
      "num_tokens": 58343293.0,
      "step": 925
    },
    {
      "entropy": 0.03522825219668448,
      "epoch": 6.377104377104377,
      "grad_norm": 0.1533203125,
      "learning_rate": 1.1307937747388034e-05,
      "loss": 0.028969509601593016,
      "mean_token_accuracy": 0.9901717621088028,
      "num_tokens": 59924721.0,
      "step": 950
    },
    {
      "entropy": 0.034883827781304715,
      "epoch": 6.545454545454545,
      "grad_norm": 0.1669921875,
      "learning_rate": 1.090560903832278e-05,
      "loss": 0.028860065937042236,
      "mean_token_accuracy": 0.9901738902926445,
      "num_tokens": 61500807.0,
      "step": 975
    },
    {
      "entropy": 0.03482985371723771,
      "epoch": 6.713804713804714,
      "grad_norm": 0.166015625,
      "learning_rate": 1.0501796247819176e-05,
      "loss": 0.02919963836669922,
      "mean_token_accuracy": 0.9900345727801323,
      "num_tokens": 63084518.0,
      "step": 1000
    },
    {
      "epoch": 6.713804713804714,
      "eval_entropy": 0.0541005334854126,
      "eval_loss": 0.16875509917736053,
      "eval_mean_token_accuracy": 0.9639611830711364,
      "eval_num_tokens": 63084518.0,
      "eval_runtime": 13.9974,
      "eval_samples_per_second": 17.86,
      "eval_steps_per_second": 8.93,
      "eval_token_accuracy": 0.963954977909252,
      "step": 1000
    },
    {
      "entropy": 0.03346874587237835,
      "epoch": 6.882154882154882,
      "grad_norm": 0.1640625,
      "learning_rate": 1.0097161130609774e-05,
      "loss": 0.027893943786621092,
      "mean_token_accuracy": 0.9905096918344498,
      "num_tokens": 64660786.0,
      "step": 1025
    },
    {
      "entropy": 0.03328982365260623,
      "epoch": 7.047138047138047,
      "grad_norm": 0.1533203125,
      "learning_rate": 9.692366789028308e-06,
      "loss": 0.026968984603881835,
      "mean_token_accuracy": 0.990928383201969,
      "num_tokens": 66215390.0,
      "step": 1050
    },
    {
      "entropy": 0.024703437695279717,
      "epoch": 7.215488215488215,
      "grad_norm": 0.140625,
      "learning_rate": 9.288076586340005e-06,
      "loss": 0.018726770877838136,
      "mean_token_accuracy": 0.9938653546571732,
      "num_tokens": 67789474.0,
      "step": 1075
    },
    {
      "entropy": 0.02491601270623505,
      "epoch": 7.383838383838384,
      "grad_norm": 0.1328125,
      "learning_rate": 8.88495305964436e-06,
      "loss": 0.019079389572143553,
      "mean_token_accuracy": 0.9937083786725998,
      "num_tokens": 69374157.0,
      "step": 1100
    },
    {
      "epoch": 7.383838383838384,
      "eval_entropy": 0.04581953993439675,
      "eval_loss": 0.19364017248153687,
      "eval_mean_token_accuracy": 0.9638748369216918,
      "eval_num_tokens": 69374157.0,
      "eval_runtime": 13.8268,
      "eval_samples_per_second": 18.081,
      "eval_steps_per_second": 9.04,
      "eval_token_accuracy": 0.9638604612813639,
      "step": 1100
    },
    {
      "entropy": 0.024102514907717705,
      "epoch": 7.552188552188552,
      "grad_norm": 0.140625,
      "learning_rate": 8.48365683413172e-06,
      "loss": 0.018542931079864503,
      "mean_token_accuracy": 0.9938124868273736,
      "num_tokens": 70951571.0,
      "step": 1125
    },
    {
      "entropy": 0.023655354911461472,
      "epoch": 7.720538720538721,
      "grad_norm": 0.12451171875,
      "learning_rate": 8.084845540473127e-06,
      "loss": 0.0183677613735199,
      "mean_token_accuracy": 0.9938945046067238,
      "num_tokens": 72531150.0,
      "step": 1150
    },
    {
      "entropy": 0.02381689740344882,
      "epoch": 7.888888888888889,
      "grad_norm": 0.12109375,
      "learning_rate": 7.689172737117389e-06,
      "loss": 0.018559412956237795,
      "mean_token_accuracy": 0.9938493025302887,
      "num_tokens": 74120322.0,
      "step": 1175
    },
    {
      "entropy": 0.021846223494266064,
      "epoch": 8.053872053872054,
      "grad_norm": 0.09765625,
      "learning_rate": 7.297286839261659e-06,
      "loss": 0.016244869232177734,
      "mean_token_accuracy": 0.9946904477416253,
      "num_tokens": 75671469.0,
      "step": 1200
    },
    {
      "epoch": 8.053872053872054,
      "eval_entropy": 0.04140873943269253,
      "eval_loss": 0.21278557181358337,
      "eval_mean_token_accuracy": 0.9638264698982238,
      "eval_num_tokens": 75671469.0,
      "eval_runtime": 13.8279,
      "eval_samples_per_second": 18.079,
      "eval_steps_per_second": 9.04,
      "eval_token_accuracy": 0.9638101864792958,
      "step": 1200
    },
    {
      "entropy": 0.018266186909750105,
      "epoch": 8.222222222222221,
      "grad_norm": 0.0966796875,
      "learning_rate": 6.909830056250527e-06,
      "loss": 0.012853317260742188,
      "mean_token_accuracy": 0.9959348595142364,
      "num_tokens": 77255219.0,
      "step": 1225
    },
    {
      "entropy": 0.017858617678284644,
      "epoch": 8.390572390572391,
      "grad_norm": 0.107421875,
      "learning_rate": 6.527437339145097e-06,
      "loss": 0.012499079704284669,
      "mean_token_accuracy": 0.9960082057118416,
      "num_tokens": 78840351.0,
      "step": 1250
    },
    {
      "entropy": 0.018353314991109072,
      "epoch": 8.558922558922559,
      "grad_norm": 0.11376953125,
      "learning_rate": 6.1507353401866896e-06,
      "loss": 0.01300368070602417,
      "mean_token_accuracy": 0.9958306196331977,
      "num_tokens": 80420094.0,
      "step": 1275
    },
    {
      "entropy": 0.017453019949607552,
      "epoch": 8.727272727272727,
      "grad_norm": 0.103515625,
      "learning_rate": 5.780341385860333e-06,
      "loss": 0.012396881580352783,
      "mean_token_accuracy": 0.9960150149464607,
      "num_tokens": 82008321.0,
      "step": 1300
    },
    {
      "epoch": 8.727272727272727,
      "eval_entropy": 0.03956229141354561,
      "eval_loss": 0.22291041910648346,
      "eval_mean_token_accuracy": 0.9638104586601257,
      "eval_num_tokens": 82008321.0,
      "eval_runtime": 13.8719,
      "eval_samples_per_second": 18.022,
      "eval_steps_per_second": 9.011,
      "eval_token_accuracy": 0.9637880655663859,
      "step": 1300
    },
    {
      "entropy": 0.01774411663878709,
      "epoch": 8.895622895622896,
      "grad_norm": 0.10302734375,
      "learning_rate": 5.416862465241033e-06,
      "loss": 0.012684780359268188,
      "mean_token_accuracy": 0.9958784037828445,
      "num_tokens": 83583971.0,
      "step": 1325
    },
    {
      "entropy": 0.01662842344435654,
      "epoch": 9.06060606060606,
      "grad_norm": 0.0751953125,
      "learning_rate": 5.060894235280637e-06,
      "loss": 0.011442055702209472,
      "mean_token_accuracy": 0.9964186977975222,
      "num_tokens": 85130709.0,
      "step": 1350
    },
    {
      "entropy": 0.01486415593419224,
      "epoch": 9.228956228956228,
      "grad_norm": 0.09521484375,
      "learning_rate": 4.713020044665348e-06,
      "loss": 0.009846681952476502,
      "mean_token_accuracy": 0.9969244155287743,
      "num_tokens": 86706204.0,
      "step": 1375
    },
    {
      "entropy": 0.015145676247775554,
      "epoch": 9.397306397306398,
      "grad_norm": 0.0869140625,
      "learning_rate": 4.373809977843676e-06,
      "loss": 0.010009793043136596,
      "mean_token_accuracy": 0.9968555930256844,
      "num_tokens": 88290356.0,
      "step": 1400
    },
    {
      "epoch": 9.397306397306398,
      "eval_entropy": 0.03756700618565083,
      "eval_loss": 0.23951546847820282,
      "eval_mean_token_accuracy": 0.9634870853424072,
      "eval_num_tokens": 88290356.0,
      "eval_runtime": 13.8742,
      "eval_samples_per_second": 18.019,
      "eval_steps_per_second": 9.009,
      "eval_token_accuracy": 0.9634723398093982,
      "step": 1400
    },
    {
      "entropy": 0.014698740621097385,
      "epoch": 9.565656565656566,
      "grad_norm": 0.08056640625,
      "learning_rate": 4.043819920791322e-06,
      "loss": 0.00973641276359558,
      "mean_token_accuracy": 0.996994196176529,
      "num_tokens": 89873272.0,
      "step": 1425
    },
    {
      "entropy": 0.014629123574122787,
      "epoch": 9.734006734006734,
      "grad_norm": 0.0927734375,
      "learning_rate": 3.7235906500440576e-06,
      "loss": 0.00979375422000885,
      "mean_token_accuracy": 0.9969650763273239,
      "num_tokens": 91455959.0,
      "step": 1450
    },
    {
      "entropy": 0.014975303104147315,
      "epoch": 9.902356902356903,
      "grad_norm": 0.0849609375,
      "learning_rate": 3.413646946491458e-06,
      "loss": 0.009887872934341431,
      "mean_token_accuracy": 0.9969008722901345,
      "num_tokens": 93038545.0,
      "step": 1475
    },
    {
      "entropy": 0.014437279744757985,
      "epoch": 10.067340067340067,
      "grad_norm": 0.0654296875,
      "learning_rate": 3.1144967353837196e-06,
      "loss": 0.009359861612319947,
      "mean_token_accuracy": 0.9971185049840382,
      "num_tokens": 94591222.0,
      "step": 1500
    },
    {
      "epoch": 10.067340067340067,
      "eval_entropy": 0.03697392378747463,
      "eval_loss": 0.24305777251720428,
      "eval_mean_token_accuracy": 0.9633980007171631,
      "eval_num_tokens": 94591222.0,
      "eval_runtime": 13.8912,
      "eval_samples_per_second": 17.997,
      "eval_steps_per_second": 8.999,
      "eval_token_accuracy": 0.9633818451656756,
      "step": 1500
    },
    {
      "entropy": 0.013786709853447973,
      "epoch": 10.235690235690235,
      "grad_norm": 0.068359375,
      "learning_rate": 2.8266302539609747e-06,
      "loss": 0.008779547214508056,
      "mean_token_accuracy": 0.9973234468698502,
      "num_tokens": 96172254.0,
      "step": 1525
    },
    {
      "entropy": 0.013736034976318479,
      "epoch": 10.404040404040405,
      "grad_norm": 0.07861328125,
      "learning_rate": 2.5505192480690865e-06,
      "loss": 0.008723703622817993,
      "mean_token_accuracy": 0.9973191824555397,
      "num_tokens": 97747622.0,
      "step": 1550
    },
    {
      "entropy": 0.01360086859203875,
      "epoch": 10.572390572390573,
      "grad_norm": 0.08154296875,
      "learning_rate": 2.2866161990785228e-06,
      "loss": 0.008712610602378846,
      "mean_token_accuracy": 0.997349898815155,
      "num_tokens": 99335839.0,
      "step": 1575
    },
    {
      "entropy": 0.013570407247170806,
      "epoch": 10.74074074074074,
      "grad_norm": 0.0771484375,
      "learning_rate": 2.0353535823732053e-06,
      "loss": 0.008633202314376831,
      "mean_token_accuracy": 0.9973410308361054,
      "num_tokens": 100917282.0,
      "step": 1600
    },
    {
      "epoch": 10.74074074074074,
      "eval_entropy": 0.03649875710904598,
      "eval_loss": 0.2482212781906128,
      "eval_mean_token_accuracy": 0.9633483939170837,
      "eval_num_tokens": 100917282.0,
      "eval_runtime": 13.819,
      "eval_samples_per_second": 18.091,
      "eval_steps_per_second": 9.045,
      "eval_token_accuracy": 0.9633335813556902,
      "step": 1600
    },
    {
      "entropy": 0.013947828765958547,
      "epoch": 10.909090909090908,
      "grad_norm": 0.07958984375,
      "learning_rate": 1.7971431586244814e-06,
      "loss": 0.008911921381950379,
      "mean_token_accuracy": 0.9972717782855034,
      "num_tokens": 102499386.0,
      "step": 1625
    },
    {
      "entropy": 0.013590329364702409,
      "epoch": 11.074074074074074,
      "grad_norm": 0.0703125,
      "learning_rate": 1.5723752990116948e-06,
      "loss": 0.008526145815849304,
      "mean_token_accuracy": 0.9974123969370005,
      "num_tokens": 104048535.0,
      "step": 1650
    },
    {
      "entropy": 0.013840707880444825,
      "epoch": 11.242424242424242,
      "grad_norm": 0.08154296875,
      "learning_rate": 1.3614183454950824e-06,
      "loss": 0.00863082766532898,
      "mean_token_accuracy": 0.9974041077494621,
      "num_tokens": 105626902.0,
      "step": 1675
    },
    {
      "entropy": 0.01306802561506629,
      "epoch": 11.41077441077441,
      "grad_norm": 0.0791015625,
      "learning_rate": 1.1646180071894608e-06,
      "loss": 0.008170877695083617,
      "mean_token_accuracy": 0.9975323963165283,
      "num_tokens": 107208502.0,
      "step": 1700
    },
    {
      "epoch": 11.41077441077441,
      "eval_entropy": 0.036279944390058515,
      "eval_loss": 0.25036948919296265,
      "eval_mean_token_accuracy": 0.9632471032142639,
      "eval_num_tokens": 107208502.0,
      "eval_runtime": 13.8985,
      "eval_samples_per_second": 17.988,
      "eval_steps_per_second": 8.994,
      "eval_token_accuracy": 0.963233031751554,
      "step": 1700
    },
    {
      "entropy": 0.013187658675014972,
      "epoch": 11.57912457912458,
      "grad_norm": 0.07861328125,
      "learning_rate": 9.822967938278172e-07,
      "loss": 0.008353355526924133,
      "mean_token_accuracy": 0.9974633774161339,
      "num_tokens": 108793895.0,
      "step": 1725
    },
    {
      "entropy": 0.013317351192235947,
      "epoch": 11.747474747474747,
      "grad_norm": 0.07177734375,
      "learning_rate": 8.147534872432761e-07,
      "loss": 0.008341993093490601,
      "mean_token_accuracy": 0.9975066068768501,
      "num_tokens": 110382929.0,
      "step": 1750
    },
    {
      "entropy": 0.013570899264886976,
      "epoch": 11.915824915824915,
      "grad_norm": 0.07763671875,
      "learning_rate": 6.622626517355557e-07,
      "loss": 0.008468236327171325,
      "mean_token_accuracy": 0.9973884886503219,
      "num_tokens": 111971394.0,
      "step": 1775
    },
    {
      "entropy": 0.013678188079359884,
      "epoch": 12.080808080808081,
      "grad_norm": 0.0771484375,
      "learning_rate": 5.250741841242735e-07,
      "loss": 0.008458149433135987,
      "mean_token_accuracy": 0.9974453488782961,
      "num_tokens": 113511798.0,
      "step": 1800
    },
    {
      "epoch": 12.080808080808081,
      "eval_entropy": 0.036235544994473456,
      "eval_loss": 0.2506777346134186,
      "eval_mean_token_accuracy": 0.9632981281280517,
      "eval_num_tokens": 113511798.0,
      "eval_runtime": 13.8333,
      "eval_samples_per_second": 18.072,
      "eval_steps_per_second": 9.036,
      "eval_token_accuracy": 0.9632792845694567,
      "step": 1800
    },
    {
      "entropy": 0.013473317860625684,
      "epoch": 12.24915824915825,
      "grad_norm": 0.06884765625,
      "learning_rate": 4.034129042265067e-07,
      "loss": 0.008400481343269348,
      "mean_token_accuracy": 0.9974551931023597,
      "num_tokens": 115097998.0,
      "step": 1825
    },
    {
      "entropy": 0.01321539796423167,
      "epoch": 12.417508417508417,
      "grad_norm": 0.0712890625,
      "learning_rate": 2.974781864296783e-07,
      "loss": 0.008239768147468567,
      "mean_token_accuracy": 0.9975069260597229,
      "num_tokens": 116678393.0,
      "step": 1850
    },
    {
      "entropy": 0.013363241832703352,
      "epoch": 12.585858585858587,
      "grad_norm": 0.072265625,
      "learning_rate": 2.0744363296356872e-07,
      "loss": 0.008337837457656861,
      "mean_token_accuracy": 0.9975020122528077,
      "num_tokens": 118256344.0,
      "step": 1875
    },
    {
      "entropy": 0.013509186143055558,
      "epoch": 12.754208754208754,
      "grad_norm": 0.0751953125,
      "learning_rate": 1.3345678940684615e-07,
      "loss": 0.008378518223762512,
      "mean_token_accuracy": 0.9974912297725678,
      "num_tokens": 119840932.0,
      "step": 1900
    },
    {
      "epoch": 12.754208754208754,
      "eval_entropy": 0.036245557606220245,
      "eval_loss": 0.2508084177970886,
      "eval_mean_token_accuracy": 0.9633139533996582,
      "eval_num_tokens": 119840932.0,
      "eval_runtime": 13.8832,
      "eval_samples_per_second": 18.007,
      "eval_steps_per_second": 9.004,
      "eval_token_accuracy": 0.9632933615140358,
      "step": 1900
    },
    {
      "entropy": 0.013419822426512837,
      "epoch": 12.922558922558922,
      "grad_norm": 0.07861328125,
      "learning_rate": 7.563890289437825e-08,
      "loss": 0.008330971598625184,
      "mean_token_accuracy": 0.9974869236350059,
      "num_tokens": 121419896.0,
      "step": 1925
    },
    {
      "entropy": 0.013293538674978274,
      "epoch": 13.087542087542088,
      "grad_norm": 0.078125,
      "learning_rate": 3.408472342152136e-08,
      "loss": 0.008297484517097473,
      "mean_token_accuracy": 0.9974846158708844,
      "num_tokens": 122967645.0,
      "step": 1950
    },
    {
      "entropy": 0.013380372445099056,
      "epoch": 13.255892255892256,
      "grad_norm": 0.0712890625,
      "learning_rate": 8.862348571043733e-09,
      "loss": 0.008304932117462159,
      "mean_token_accuracy": 0.9974824267625809,
      "num_tokens": 124542187.0,
      "step": 1975
    },
    {
      "entropy": 0.013427549134939909,
      "epoch": 13.424242424242424,
      "grad_norm": 0.07763671875,
      "learning_rate": 1.311191710651194e-11,
      "loss": 0.008341950178146363,
      "mean_token_accuracy": 0.9974777749180794,
      "num_tokens": 126127250.0,
      "step": 2000
    },
    {
      "epoch": 13.424242424242424,
      "eval_entropy": 0.036229205310344695,
      "eval_loss": 0.2507854700088501,
      "eval_mean_token_accuracy": 0.9632595100402832,
      "eval_num_tokens": 126127250.0,
      "eval_runtime": 13.8215,
      "eval_samples_per_second": 18.088,
      "eval_steps_per_second": 9.044,
      "eval_token_accuracy": 0.9632390647278022,
      "step": 2000
    }
  ],
  "logging_steps": 25,
  "max_steps": 2000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 14,
  "save_steps": 400,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.700374075876745e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}