[
  {
    "loss": 2.0326,
    "grad_norm": 1.0961493253707886,
    "learning_rate": 9e-06,
    "entropy": 1.4562864780426026,
    "num_tokens": 113103.0,
    "mean_token_accuracy": 0.5854019731283188,
    "epoch": 0.0034019391052900155,
    "step": 10
  },
  {
    "loss": 1.9722,
    "grad_norm": 0.5352939963340759,
    "learning_rate": 1.9e-05,
    "entropy": 1.6092135667800904,
    "num_tokens": 221983.0,
    "mean_token_accuracy": 0.5872704654932022,
    "epoch": 0.006803878210580031,
    "step": 20
  },
  {
    "loss": 1.8198,
    "grad_norm": 0.5580214262008667,
    "learning_rate": 2.9e-05,
    "entropy": 1.7635144531726836,
    "num_tokens": 343755.0,
    "mean_token_accuracy": 0.59696926176548,
    "epoch": 0.010205817315870046,
    "step": 30
  },
  {
    "loss": 1.6405,
    "grad_norm": 0.9694721102714539,
    "learning_rate": 3.9000000000000006e-05,
    "entropy": 1.6470328569412231,
    "num_tokens": 462278.0,
    "mean_token_accuracy": 0.6220342874526977,
    "epoch": 0.013607756421160062,
    "step": 40
  },
  {
    "loss": 1.2829,
    "grad_norm": 0.8506905436515808,
    "learning_rate": 4.9e-05,
    "entropy": 1.3260977208614348,
    "num_tokens": 580883.0,
    "mean_token_accuracy": 0.6881176471710205,
    "epoch": 0.017009695526450076,
    "step": 50
  },
  {
    "loss": 1.0615,
    "grad_norm": 0.5665390491485596,
    "learning_rate": 5.9e-05,
    "entropy": 1.0532318294048308,
    "num_tokens": 693719.0,
    "mean_token_accuracy": 0.7430563151836396,
    "epoch": 0.020411634631740092,
    "step": 60
  },
  {
    "loss": 0.9244,
    "grad_norm": 0.5383474230766296,
    "learning_rate": 6.9e-05,
    "entropy": 0.9287566632032395,
    "num_tokens": 800049.0,
    "mean_token_accuracy": 0.7744365096092224,
    "epoch": 0.023813573737030108,
    "step": 70
  },
  {
    "loss": 0.8564,
    "grad_norm": 0.5389857292175293,
    "learning_rate": 7.900000000000001e-05,
    "entropy": 0.8546981632709503,
    "num_tokens": 909347.0,
    "mean_token_accuracy": 0.788225668668747,
    "epoch": 0.027215512842320124,
    "step": 80
  },
  {
    "loss": 0.7884,
    "grad_norm": 0.5869827270507812,
    "learning_rate": 8.900000000000001e-05,
    "entropy": 0.7829572021961212,
    "num_tokens": 1024214.0,
    "mean_token_accuracy": 0.8022073119878769,
    "epoch": 0.030617451947610137,
    "step": 90
  },
  {
    "loss": 0.7827,
    "grad_norm": 0.4458995461463928,
    "learning_rate": 9.900000000000001e-05,
    "entropy": 0.796157357096672,
    "num_tokens": 1145337.0,
    "mean_token_accuracy": 0.8003427386283875,
    "epoch": 0.03401939105290015,
    "step": 100
  },
  {
    "loss": 0.7694,
    "grad_norm": 0.4616456925868988,
    "learning_rate": 9.999752209583493e-05,
    "entropy": 0.7759479373693466,
    "num_tokens": 1255483.0,
    "mean_token_accuracy": 0.8059569209814071,
    "epoch": 0.037421330158190165,
    "step": 110
  },
  {
    "loss": 0.7328,
    "grad_norm": 0.5324557423591614,
    "learning_rate": 9.998895681650353e-05,
    "entropy": 0.7250411480665206,
    "num_tokens": 1376406.0,
    "mean_token_accuracy": 0.8113020449876786,
    "epoch": 0.040823269263480184,
    "step": 120
  },
  {
    "loss": 0.7204,
    "grad_norm": 0.4355196952819824,
    "learning_rate": 9.997427461844063e-05,
    "entropy": 0.7220166236162185,
    "num_tokens": 1487418.0,
    "mean_token_accuracy": 0.8159421294927597,
    "epoch": 0.0442252083687702,
    "step": 130
  },
  {
    "loss": 0.7181,
    "grad_norm": 0.4837367534637451,
    "learning_rate": 9.99534772982393e-05,
    "entropy": 0.7167926013469696,
    "num_tokens": 1603603.0,
    "mean_token_accuracy": 0.8169557631015778,
    "epoch": 0.047627147474060216,
    "step": 140
  },
  {
    "loss": 0.7004,
    "grad_norm": 0.42535701394081116,
    "learning_rate": 9.992656740077193e-05,
    "entropy": 0.7097186267375946,
    "num_tokens": 1719331.0,
    "mean_token_accuracy": 0.8169945359230042,
    "epoch": 0.05102908657935023,
    "step": 150
  },
  {
    "loss": 0.74,
    "grad_norm": 0.4244067966938019,
    "learning_rate": 9.989354821887894e-05,
    "entropy": 0.7415460795164108,
    "num_tokens": 1835165.0,
    "mean_token_accuracy": 0.8098712652921677,
    "epoch": 0.05443102568464025,
    "step": 160
  },
  {
    "loss": 0.7071,
    "grad_norm": 0.46471285820007324,
    "learning_rate": 9.985442379296574e-05,
    "entropy": 0.7142940938472748,
    "num_tokens": 1955878.0,
    "mean_token_accuracy": 0.8144379884004593,
    "epoch": 0.05783296478993026,
    "step": 170
  },
  {
    "loss": 0.6959,
    "grad_norm": 0.48218467831611633,
    "learning_rate": 9.980919891050836e-05,
    "entropy": 0.7002523899078369,
    "num_tokens": 2079862.0,
    "mean_token_accuracy": 0.8176177382469177,
    "epoch": 0.06123490389522027,
    "step": 180
  },
  {
    "loss": 0.6977,
    "grad_norm": 0.43155911564826965,
    "learning_rate": 9.97578791054677e-05,
    "entropy": 0.7015494972467422,
    "num_tokens": 2183469.0,
    "mean_token_accuracy": 0.8218911945819855,
    "epoch": 0.06463684300051029,
    "step": 190
  },
  {
    "loss": 0.7036,
    "grad_norm": 0.4237188398838043,
    "learning_rate": 9.970047065761225e-05,
    "entropy": 0.7062139421701431,
    "num_tokens": 2297979.0,
    "mean_token_accuracy": 0.8169696599245071,
    "epoch": 0.0680387821058003,
    "step": 200
  },
  {
    "loss": 0.6896,
    "grad_norm": 0.4372621178627014,
    "learning_rate": 9.963698059174973e-05,
    "entropy": 0.6874466806650161,
    "num_tokens": 2413698.0,
    "mean_token_accuracy": 0.8205429404973984,
    "epoch": 0.07144072121109032,
    "step": 210
  },
  {
    "loss": 0.6861,
    "grad_norm": 0.42981356382369995,
    "learning_rate": 9.956741667686749e-05,
    "entropy": 0.6936574131250381,
    "num_tokens": 2523689.0,
    "mean_token_accuracy": 0.8207722246646881,
    "epoch": 0.07484266031638033,
    "step": 220
  },
  {
    "loss": 0.6876,
    "grad_norm": 0.4125567078590393,
    "learning_rate": 9.949178742518188e-05,
    "entropy": 0.7051049053668976,
    "num_tokens": 2629062.0,
    "mean_token_accuracy": 0.8210607588291168,
    "epoch": 0.07824459942167035,
    "step": 230
  },
  {
    "loss": 0.7149,
    "grad_norm": 0.3944653868675232,
    "learning_rate": 9.941010209109653e-05,
    "entropy": 0.7121995657682418,
    "num_tokens": 2744221.0,
    "mean_token_accuracy": 0.8171839863061905,
    "epoch": 0.08164653852696037,
    "step": 240
  },
  {
    "loss": 0.7041,
    "grad_norm": 0.4692147970199585,
    "learning_rate": 9.932237067007014e-05,
    "entropy": 0.7010555505752564,
    "num_tokens": 2867672.0,
    "mean_token_accuracy": 0.8154427468776703,
    "epoch": 0.08504847763225039,
    "step": 250
  },
  {
    "loss": 0.6701,
    "grad_norm": 0.36812499165534973,
    "learning_rate": 9.922860389739316e-05,
    "entropy": 0.673400056362152,
    "num_tokens": 2969786.0,
    "mean_token_accuracy": 0.826720854640007,
    "epoch": 0.0884504167375404,
    "step": 260
  },
  {
    "loss": 0.6828,
    "grad_norm": 0.5079509019851685,
    "learning_rate": 9.91288132468743e-05,
    "entropy": 0.6830800861120224,
    "num_tokens": 3078545.0,
    "mean_token_accuracy": 0.8246483653783798,
    "epoch": 0.09185235584283041,
    "step": 270
  },
  {
    "loss": 0.6591,
    "grad_norm": 0.4336896538734436,
    "learning_rate": 9.90230109294365e-05,
    "entropy": 0.6682640433311462,
    "num_tokens": 3202461.0,
    "mean_token_accuracy": 0.8245112657546997,
    "epoch": 0.09525429494812043,
    "step": 280
  },
  {
    "loss": 0.6742,
    "grad_norm": 0.38131922483444214,
    "learning_rate": 9.891120989162277e-05,
    "entropy": 0.6732221603393554,
    "num_tokens": 3317749.0,
    "mean_token_accuracy": 0.8278978168964386,
    "epoch": 0.09865623405341044,
    "step": 290
  },
  {
    "loss": 0.7072,
    "grad_norm": 0.43391740322113037,
    "learning_rate": 9.879342381401183e-05,
    "entropy": 0.719240027666092,
    "num_tokens": 3438813.0,
    "mean_token_accuracy": 0.8140431463718414,
    "epoch": 0.10205817315870046,
    "step": 300
  },
  {
    "loss": 0.6726,
    "grad_norm": 0.37840133905410767,
    "learning_rate": 9.866966710954432e-05,
    "entropy": 0.6736223340034485,
    "num_tokens": 3558578.0,
    "mean_token_accuracy": 0.8215447336435318,
    "epoch": 0.10546011226399048,
    "step": 310
  },
  {
    "loss": 0.6739,
    "grad_norm": 0.36936870217323303,
    "learning_rate": 9.853995492175894e-05,
    "entropy": 0.6741836041212081,
    "num_tokens": 3680797.0,
    "mean_token_accuracy": 0.8207811802625656,
    "epoch": 0.1088620513692805,
    "step": 320
  },
  {
    "loss": 0.6462,
    "grad_norm": 0.3381315767765045,
    "learning_rate": 9.840430312293954e-05,
    "entropy": 0.6454634428024292,
    "num_tokens": 3798203.0,
    "mean_token_accuracy": 0.8290177077054978,
    "epoch": 0.1122639904745705,
    "step": 330
  },
  {
    "loss": 0.6309,
    "grad_norm": 0.4414357841014862,
    "learning_rate": 9.826272831217282e-05,
    "entropy": 0.6358878672122955,
    "num_tokens": 3908285.0,
    "mean_token_accuracy": 0.8346272200345993,
    "epoch": 0.11566592957986052,
    "step": 340
  },
  {
    "loss": 0.662,
    "grad_norm": 0.3412431478500366,
    "learning_rate": 9.811524781331726e-05,
    "entropy": 0.669054564833641,
    "num_tokens": 4027009.0,
    "mean_token_accuracy": 0.823895263671875,
    "epoch": 0.11906786868515054,
    "step": 350
  },
  {
    "loss": 0.6396,
    "grad_norm": 0.4202592670917511,
    "learning_rate": 9.796187967288317e-05,
    "entropy": 0.6534929439425469,
    "num_tokens": 4141379.0,
    "mean_token_accuracy": 0.8286080300807953,
    "epoch": 0.12246980779044055,
    "step": 360
  },
  {
    "loss": 0.6462,
    "grad_norm": 0.4112204313278198,
    "learning_rate": 9.780264265782452e-05,
    "entropy": 0.6516435265541076,
    "num_tokens": 4264298.0,
    "mean_token_accuracy": 0.8284994632005691,
    "epoch": 0.12587174689573058,
    "step": 370
  },
  {
    "loss": 0.6385,
    "grad_norm": 0.4041576683521271,
    "learning_rate": 9.763755625324247e-05,
    "entropy": 0.6425079524517059,
    "num_tokens": 4380764.0,
    "mean_token_accuracy": 0.8293725460767746,
    "epoch": 0.12927368600102057,
    "step": 380
  },
  {
    "loss": 0.6474,
    "grad_norm": 0.40553557872772217,
    "learning_rate": 9.746664066000105e-05,
    "entropy": 0.650575777888298,
    "num_tokens": 4502319.0,
    "mean_token_accuracy": 0.8261306613683701,
    "epoch": 0.1326756251063106,
    "step": 390
  },
  {
    "loss": 0.6208,
    "grad_norm": 0.4033198058605194,
    "learning_rate": 9.728991679225534e-05,
    "entropy": 0.6212592482566833,
    "num_tokens": 4619137.0,
    "mean_token_accuracy": 0.8338732123374939,
    "epoch": 0.1360775642116006,
    "step": 400
  },
  {
    "loss": 0.6588,
    "grad_norm": 0.3973551392555237,
    "learning_rate": 9.71074062748922e-05,
    "entropy": 0.6602439433336258,
    "num_tokens": 4740514.0,
    "mean_token_accuracy": 0.8240435630083084,
    "epoch": 0.13947950331689063,
    "step": 410
  },
  {
    "loss": 0.6231,
    "grad_norm": 0.38988974690437317,
    "learning_rate": 9.691913144088423e-05,
    "entropy": 0.6275951579213143,
    "num_tokens": 4849434.0,
    "mean_token_accuracy": 0.8351284950971604,
    "epoch": 0.14288144242218065,
    "step": 420
  },
  {
    "loss": 0.6486,
    "grad_norm": 0.4000033736228943,
    "learning_rate": 9.672511532855695e-05,
    "entropy": 0.6495797425508499,
    "num_tokens": 4970677.0,
    "mean_token_accuracy": 0.8287895292043685,
    "epoch": 0.14628338152747067,
    "step": 430
  },
  {
    "loss": 0.6369,
    "grad_norm": 0.36982354521751404,
    "learning_rate": 9.652538167876966e-05,
    "entropy": 0.6435147285461426,
    "num_tokens": 5079274.0,
    "mean_token_accuracy": 0.8325352728366852,
    "epoch": 0.14968532063276066,
    "step": 440
  },
  {
    "loss": 0.6589,
    "grad_norm": 0.30258873105049133,
    "learning_rate": 9.63199549320105e-05,
    "entropy": 0.656851077079773,
    "num_tokens": 5199280.0,
    "mean_token_accuracy": 0.8274909108877182,
    "epoch": 0.15308725973805068,
    "step": 450
  },
  {
    "loss": 0.6171,
    "grad_norm": 0.40595975518226624,
    "learning_rate": 9.610886022540558e-05,
    "entropy": 0.6167075544595718,
    "num_tokens": 5314855.0,
    "mean_token_accuracy": 0.8360531657934189,
    "epoch": 0.1564891988433407,
    "step": 460
  },
  {
    "loss": 0.6535,
    "grad_norm": 0.3571644723415375,
    "learning_rate": 9.589212338964331e-05,
    "entropy": 0.6573659986257553,
    "num_tokens": 5427712.0,
    "mean_token_accuracy": 0.826265224814415,
    "epoch": 0.15989113794863072,
    "step": 470
  },
  {
    "loss": 0.6322,
    "grad_norm": 0.3340984880924225,
    "learning_rate": 9.566977094581344e-05,
    "entropy": 0.6363773226737977,
    "num_tokens": 5555811.0,
    "mean_token_accuracy": 0.830034053325653,
    "epoch": 0.16329307705392074,
    "step": 480
  },
  {
    "loss": 0.6505,
    "grad_norm": 0.36710014939308167,
    "learning_rate": 9.544183010216184e-05,
    "entropy": 0.6666623204946518,
    "num_tokens": 5670677.0,
    "mean_token_accuracy": 0.8263821274042129,
    "epoch": 0.16669501615921076,
    "step": 490
  },
  {
    "loss": 0.6484,
    "grad_norm": 0.35856229066848755,
    "learning_rate": 9.520832875076118e-05,
    "entropy": 0.6493088006973267,
    "num_tokens": 5791079.0,
    "mean_token_accuracy": 0.826611652970314,
    "epoch": 0.17009695526450078,
    "step": 500
  },
  {
    "loss": 0.6395,
    "grad_norm": 0.36879998445510864,
    "learning_rate": 9.496929546409791e-05,
    "entropy": 0.6355485945940018,
    "num_tokens": 5915427.0,
    "mean_token_accuracy": 0.8316580206155777,
    "epoch": 0.17349889436979077,
    "step": 510
  },
  {
    "loss": 0.6405,
    "grad_norm": 0.3662184774875641,
    "learning_rate": 9.472475949157591e-05,
    "entropy": 0.6411178290843964,
    "num_tokens": 6042113.0,
    "mean_token_accuracy": 0.8287693798542023,
    "epoch": 0.1769008334750808,
    "step": 520
  },
  {
    "loss": 0.6248,
    "grad_norm": 0.3676609992980957,
    "learning_rate": 9.447475075593746e-05,
    "entropy": 0.6330993086099624,
    "num_tokens": 6158285.0,
    "mean_token_accuracy": 0.8320850521326065,
    "epoch": 0.1803027725803708,
    "step": 530
  },
  {
    "loss": 0.605,
    "grad_norm": 0.3490847945213318,
    "learning_rate": 9.421929984960168e-05,
    "entropy": 0.6043026179075242,
    "num_tokens": 6269273.0,
    "mean_token_accuracy": 0.8386462539434433,
    "epoch": 0.18370471168566083,
    "step": 540
  },
  {
    "loss": 0.6035,
    "grad_norm": 0.3701430857181549,
    "learning_rate": 9.395843803092103e-05,
    "entropy": 0.6117337167263031,
    "num_tokens": 6374978.0,
    "mean_token_accuracy": 0.8389517337083816,
    "epoch": 0.18710665079095085,
    "step": 550
  },
  {
    "loss": 0.667,
    "grad_norm": 0.37965649366378784,
    "learning_rate": 9.369219722035657e-05,
    "entropy": 0.6723731458187103,
    "num_tokens": 6492671.0,
    "mean_token_accuracy": 0.8266532897949219,
    "epoch": 0.19050858989624087,
    "step": 560
  },
  {
    "loss": 0.6473,
    "grad_norm": 0.335111528635025,
    "learning_rate": 9.34206099965717e-05,
    "entropy": 0.6496838480234146,
    "num_tokens": 6605704.0,
    "mean_token_accuracy": 0.8288087368011474,
    "epoch": 0.19391052900153088,
    "step": 570
  },
  {
    "loss": 0.6479,
    "grad_norm": 0.4136720895767212,
    "learning_rate": 9.314370959244589e-05,
    "entropy": 0.6532777503132821,
    "num_tokens": 6724628.0,
    "mean_token_accuracy": 0.8273604065179825,
    "epoch": 0.19731246810682088,
    "step": 580
  },
  {
    "loss": 0.6917,
    "grad_norm": 0.349250465631485,
    "learning_rate": 9.286152989100808e-05,
    "entropy": 0.6956728607416153,
    "num_tokens": 6842793.0,
    "mean_token_accuracy": 0.8188706696033478,
    "epoch": 0.2007144072121109,
    "step": 590
  },
  {
    "loss": 0.6596,
    "grad_norm": 0.37561726570129395,
    "learning_rate": 9.257410542129048e-05,
    "entropy": 0.6637489289045334,
    "num_tokens": 6969269.0,
    "mean_token_accuracy": 0.8233319729566574,
    "epoch": 0.20411634631740092,
    "step": 600
  },
  {
    "loss": 0.6148,
    "grad_norm": 0.35306721925735474,
    "learning_rate": 9.22814713541035e-05,
    "entropy": 0.609418871998787,
    "num_tokens": 7089158.0,
    "mean_token_accuracy": 0.8354996383190155,
    "epoch": 0.20751828542269093,
    "step": 610
  },
  {
    "loss": 0.644,
    "grad_norm": 0.38659054040908813,
    "learning_rate": 9.198366349773205e-05,
    "entropy": 0.6533850967884064,
    "num_tokens": 7214581.0,
    "mean_token_accuracy": 0.8262244701385498,
    "epoch": 0.21092022452798095,
    "step": 620
  },
  {
    "loss": 0.6227,
    "grad_norm": 0.3564472794532776,
    "learning_rate": 9.168071829355376e-05,
    "entropy": 0.6244180202484131,
    "num_tokens": 7332309.0,
    "mean_token_accuracy": 0.8361935257911682,
    "epoch": 0.21432216363327097,
    "step": 630
  },
  {
    "loss": 0.6308,
    "grad_norm": 0.3747089207172394,
    "learning_rate": 9.137267281157999e-05,
    "entropy": 0.6325557798147201,
    "num_tokens": 7450413.0,
    "mean_token_accuracy": 0.8316955536603927,
    "epoch": 0.217724102738561,
    "step": 640
  },
  {
    "loss": 0.616,
    "grad_norm": 0.4263840913772583,
    "learning_rate": 9.105956474591953e-05,
    "entropy": 0.620253998041153,
    "num_tokens": 7557132.0,
    "mean_token_accuracy": 0.8373414397239685,
    "epoch": 0.22112604184385098,
    "step": 650
  },
  {
    "loss": 0.6176,
    "grad_norm": 0.38803404569625854,
    "learning_rate": 9.074143241016631e-05,
    "entropy": 0.6201047956943512,
    "num_tokens": 7671938.0,
    "mean_token_accuracy": 0.8363149344921113,
    "epoch": 0.224527980949141,
    "step": 660
  },
  {
    "loss": 0.6434,
    "grad_norm": 0.3495098054409027,
    "learning_rate": 9.04183147327111e-05,
    "entropy": 0.6501924157142639,
    "num_tokens": 7798607.0,
    "mean_token_accuracy": 0.8257293999195099,
    "epoch": 0.22792992005443102,
    "step": 670
  },
  {
    "loss": 0.6176,
    "grad_norm": 0.34818509221076965,
    "learning_rate": 9.009025125197792e-05,
    "entropy": 0.6221291124820709,
    "num_tokens": 7904767.0,
    "mean_token_accuracy": 0.8381055980920792,
    "epoch": 0.23133185915972104,
    "step": 680
  },
  {
    "loss": 0.6194,
    "grad_norm": 0.35937735438346863,
    "learning_rate": 8.975728211158609e-05,
    "entropy": 0.6223055779933929,
    "num_tokens": 8026821.0,
    "mean_token_accuracy": 0.8331304669380188,
    "epoch": 0.23473379826501106,
    "step": 690
  },
  {
    "loss": 0.6343,
    "grad_norm": 0.3995741307735443,
    "learning_rate": 8.941944805543788e-05,
    "entropy": 0.6313437297940254,
    "num_tokens": 8137346.0,
    "mean_token_accuracy": 0.8347465574741364,
    "epoch": 0.23813573737030108,
    "step": 700
  },
  {
    "loss": 0.6447,
    "grad_norm": 0.38594338297843933,
    "learning_rate": 8.907679042273293e-05,
    "entropy": 0.6474226981401443,
    "num_tokens": 8241289.0,
    "mean_token_accuracy": 0.8319136798381805,
    "epoch": 0.24153767647559107,
    "step": 710
  },
  {
    "loss": 0.605,
    "grad_norm": 0.3661476969718933,
    "learning_rate": 8.872935114290979e-05,
    "entropy": 0.6085086613893509,
    "num_tokens": 8357072.0,
    "mean_token_accuracy": 0.8373961597681046,
    "epoch": 0.2449396155808811,
    "step": 720
  },
  {
    "loss": 0.5994,
    "grad_norm": 0.3534446954727173,
    "learning_rate": 8.837717273051515e-05,
    "entropy": 0.5992837458848953,
    "num_tokens": 8480791.0,
    "mean_token_accuracy": 0.8372073084115982,
    "epoch": 0.2483415546861711,
    "step": 730
  },
  {
    "loss": 0.591,
    "grad_norm": 0.3611646294593811,
    "learning_rate": 8.802029828000156e-05,
    "entropy": 0.5920722305774688,
    "num_tokens": 8593584.0,
    "mean_token_accuracy": 0.840802788734436,
    "epoch": 0.25174349379146116,
    "step": 740
  },
  {
    "loss": 0.6379,
    "grad_norm": 0.3454264998435974,
    "learning_rate": 8.765877146045413e-05,
    "entropy": 0.645842906832695,
    "num_tokens": 8697045.0,
    "mean_token_accuracy": 0.832975035905838,
    "epoch": 0.2551454328967511,
    "step": 750
  },
  {
    "loss": 0.588,
    "grad_norm": 0.36970141530036926,
    "learning_rate": 8.729263651024705e-05,
    "entropy": 0.589533594250679,
    "num_tokens": 8810383.0,
    "mean_token_accuracy": 0.8419877678155899,
    "epoch": 0.25854737200204114,
    "step": 760
  },
  {
    "loss": 0.6126,
    "grad_norm": 0.30673205852508545,
    "learning_rate": 8.692193823163017e-05,
    "entropy": 0.603627099096775,
    "num_tokens": 8926623.0,
    "mean_token_accuracy": 0.8384459555149079,
    "epoch": 0.26194931110733116,
    "step": 770
  },
  {
    "loss": 0.6416,
    "grad_norm": 0.40019306540489197,
    "learning_rate": 8.654672198524692e-05,
    "entropy": 0.6472606122493744,
    "num_tokens": 9037995.0,
    "mean_token_accuracy": 0.8314683675765991,
    "epoch": 0.2653512502126212,
    "step": 780
  },
  {
    "loss": 0.6028,
    "grad_norm": 0.3364422917366028,
    "learning_rate": 8.616703368458361e-05,
    "entropy": 0.6071313232183456,
    "num_tokens": 9159140.0,
    "mean_token_accuracy": 0.8371691346168518,
    "epoch": 0.2687531893179112,
    "step": 790
  },
  {
    "loss": 0.6054,
    "grad_norm": 0.3635218143463135,
    "learning_rate": 8.578291979035132e-05,
    "entropy": 0.615416657924652,
    "num_tokens": 9267853.0,
    "mean_token_accuracy": 0.8379261702299118,
    "epoch": 0.2721551284232012,
    "step": 800
  },
  {
    "loss": 0.6104,
    "grad_norm": 0.33966347575187683,
    "learning_rate": 8.539442730480061e-05,
    "entropy": 0.6133037626743316,
    "num_tokens": 9383976.0,
    "mean_token_accuracy": 0.835821345448494,
    "epoch": 0.27555706752849124,
    "step": 810
  },
  {
    "loss": 0.5883,
    "grad_norm": 0.3680882155895233,
    "learning_rate": 8.500160376597016e-05,
    "entropy": 0.5889941453933716,
    "num_tokens": 9490262.0,
    "mean_token_accuracy": 0.8436674028635025,
    "epoch": 0.27895900663378126,
    "step": 820
  },
  {
    "loss": 0.6011,
    "grad_norm": 0.3962339758872986,
    "learning_rate": 8.46044972418697e-05,
    "entropy": 0.599892795085907,
    "num_tokens": 9607939.0,
    "mean_token_accuracy": 0.8401912897825241,
    "epoch": 0.2823609457390713,
    "step": 830
  },
  {
    "loss": 0.6268,
    "grad_norm": 0.3375062644481659,
    "learning_rate": 8.420315632459817e-05,
    "entropy": 0.6311625123023987,
    "num_tokens": 9722935.0,
    "mean_token_accuracy": 0.8338307291269302,
    "epoch": 0.2857628848443613,
    "step": 840
  },
  {
    "loss": 0.638,
    "grad_norm": 0.36880624294281006,
    "learning_rate": 8.379763012439771e-05,
    "entropy": 0.641839075088501,
    "num_tokens": 9835999.0,
    "mean_token_accuracy": 0.831251859664917,
    "epoch": 0.2891648239496513,
    "step": 850
  },
  {
    "loss": 0.5823,
    "grad_norm": 0.35837697982788086,
    "learning_rate": 8.338796826364437e-05,
    "entropy": 0.5842991501092911,
    "num_tokens": 9946974.0,
    "mean_token_accuracy": 0.8434822201728821,
    "epoch": 0.29256676305494134,
    "step": 860
  },
  {
    "loss": 0.6356,
    "grad_norm": 0.39383432269096375,
    "learning_rate": 8.29742208707758e-05,
    "entropy": 0.6413630664348602,
    "num_tokens": 10067421.0,
    "mean_token_accuracy": 0.8298591256141663,
    "epoch": 0.29596870216023136,
    "step": 870
  },
  {
    "loss": 0.6434,
    "grad_norm": 0.38471469283103943,
    "learning_rate": 8.255643857415758e-05,
    "entropy": 0.6492842510342598,
    "num_tokens": 10172152.0,
    "mean_token_accuracy": 0.8317259013652801,
    "epoch": 0.2993706412655213,
    "step": 880
  },
  {
    "loss": 0.6135,
    "grad_norm": 0.3366308808326721,
    "learning_rate": 8.213467249588783e-05,
    "entropy": 0.6186555951833725,
    "num_tokens": 10290628.0,
    "mean_token_accuracy": 0.833742293715477,
    "epoch": 0.30277258037081134,
    "step": 890
  },
  {
    "loss": 0.5787,
    "grad_norm": 0.40084564685821533,
    "learning_rate": 8.170897424554171e-05,
    "entropy": 0.5826229274272918,
    "num_tokens": 10407951.0,
    "mean_token_accuracy": 0.8439725250005722,
    "epoch": 0.30617451947610136,
    "step": 900
  },
  {
    "loss": 0.628,
    "grad_norm": 0.4093911051750183,
    "learning_rate": 8.127939591385623e-05,
    "entropy": 0.639702819287777,
    "num_tokens": 10516552.0,
    "mean_token_accuracy": 0.8321585863828659,
    "epoch": 0.3095764585813914,
    "step": 910
  },
  {
    "loss": 0.6622,
    "grad_norm": 0.3402785360813141,
    "learning_rate": 8.084599006635602e-05,
    "entropy": 0.6612581133842468,
    "num_tokens": 10622685.0,
    "mean_token_accuracy": 0.8280930787324905,
    "epoch": 0.3129783976866814,
    "step": 920
  },
  {
    "loss": 0.5945,
    "grad_norm": 0.381832093000412,
    "learning_rate": 8.040880973692133e-05,
    "entropy": 0.5946418136358261,
    "num_tokens": 10740084.0,
    "mean_token_accuracy": 0.8418374478816986,
    "epoch": 0.3163803367919714,
    "step": 930
  },
  {
    "loss": 0.6012,
    "grad_norm": 0.36465200781822205,
    "learning_rate": 7.996790842129834e-05,
    "entropy": 0.6030470594763756,
    "num_tokens": 10858519.0,
    "mean_token_accuracy": 0.8368221133947372,
    "epoch": 0.31978227589726144,
    "step": 940
  },
  {
    "loss": 0.6109,
    "grad_norm": 0.34183382987976074,
    "learning_rate": 7.952334007055317e-05,
    "entropy": 0.6180869042873383,
    "num_tokens": 10982717.0,
    "mean_token_accuracy": 0.835748416185379,
    "epoch": 0.32318421500255146,
    "step": 950
  },
  {
    "loss": 0.6427,
    "grad_norm": 0.3375698924064636,
    "learning_rate": 7.907515908447026e-05,
    "entropy": 0.6360474318265915,
    "num_tokens": 11109821.0,
    "mean_token_accuracy": 0.8303357720375061,
    "epoch": 0.3265861541078415,
    "step": 960
  },
  {
    "loss": 0.6125,
    "grad_norm": 0.33526158332824707,
    "learning_rate": 7.862342030489548e-05,
    "entropy": 0.6055434793233871,
    "num_tokens": 11222141.0,
    "mean_token_accuracy": 0.839334124326706,
    "epoch": 0.3299880932131315,
    "step": 970
  },
  {
    "loss": 0.601,
    "grad_norm": 0.40864965319633484,
    "learning_rate": 7.816817900902569e-05,
    "entropy": 0.6122152507305145,
    "num_tokens": 11334721.0,
    "mean_token_accuracy": 0.8393468946218491,
    "epoch": 0.3333900323184215,
    "step": 980
  },
  {
    "loss": 0.6341,
    "grad_norm": 0.35346099734306335,
    "learning_rate": 7.77094909026444e-05,
    "entropy": 0.6331113621592521,
    "num_tokens": 11442226.0,
    "mean_token_accuracy": 0.8354059338569642,
    "epoch": 0.33679197142371153,
    "step": 990
  },
  {
    "loss": 0.6021,
    "grad_norm": 0.3349364101886749,
    "learning_rate": 7.724741211330561e-05,
    "entropy": 0.609099206328392,
    "num_tokens": 11559196.0,
    "mean_token_accuracy": 0.8378647118806839,
    "epoch": 0.34019391052900155,
    "step": 1000
  },
  {
    "loss": 0.596,
    "grad_norm": 0.3482501804828644,
    "learning_rate": 7.67819991834655e-05,
    "entropy": 0.6021771982312203,
    "num_tokens": 11674831.0,
    "mean_token_accuracy": 0.8393041074275971,
    "epoch": 0.3435958496342916,
    "step": 1010
  },
  {
    "loss": 0.5712,
    "grad_norm": 0.3629717528820038,
    "learning_rate": 7.631330906356371e-05,
    "entropy": 0.5844026952981949,
    "num_tokens": 11795718.0,
    "mean_token_accuracy": 0.8434269249439239,
    "epoch": 0.34699778873958154,
    "step": 1020
  },
  {
    "loss": 0.6058,
    "grad_norm": 0.3741706311702728,
    "learning_rate": 7.584139910505458e-05,
    "entropy": 0.6056048899888993,
    "num_tokens": 11911642.0,
    "mean_token_accuracy": 0.8380466908216476,
    "epoch": 0.35039972784487156,
    "step": 1030
  },
  {
    "loss": 0.5877,
    "grad_norm": 0.30570724606513977,
    "learning_rate": 7.536632705338929e-05,
    "entropy": 0.5891546666622162,
    "num_tokens": 12029598.0,
    "mean_token_accuracy": 0.8402005821466446,
    "epoch": 0.3538016669501616,
    "step": 1040
  },
  {
    "loss": 0.6213,
    "grad_norm": 0.362284779548645,
    "learning_rate": 7.488815104094977e-05,
    "entropy": 0.6238945558667183,
    "num_tokens": 12138307.0,
    "mean_token_accuracy": 0.8360639154911041,
    "epoch": 0.3572036060554516,
    "step": 1050
  },
  {
    "loss": 0.575,
    "grad_norm": 0.33959442377090454,
    "learning_rate": 7.440692957993543e-05,
    "entropy": 0.5791127294301986,
    "num_tokens": 12253996.0,
    "mean_token_accuracy": 0.8450855612754822,
    "epoch": 0.3606055451607416,
    "step": 1060
  },
  {
    "loss": 0.6387,
    "grad_norm": 0.35954946279525757,
    "learning_rate": 7.392272155520313e-05,
    "entropy": 0.6434630990028382,
    "num_tokens": 12367363.0,
    "mean_token_accuracy": 0.8307264894247055,
    "epoch": 0.36400748426603163,
    "step": 1070
  },
  {
    "loss": 0.6154,
    "grad_norm": 0.3226819932460785,
    "learning_rate": 7.343558621706186e-05,
    "entropy": 0.6265482395887375,
    "num_tokens": 12487636.0,
    "mean_token_accuracy": 0.8338571727275849,
    "epoch": 0.36740942337132165,
    "step": 1080
  },
  {
    "loss": 0.6088,
    "grad_norm": 0.3557349145412445,
    "learning_rate": 7.294558317402249e-05,
    "entropy": 0.6144857376813888,
    "num_tokens": 12598626.0,
    "mean_token_accuracy": 0.8359712421894073,
    "epoch": 0.3708113624766117,
    "step": 1090
  },
  {
    "loss": 0.6009,
    "grad_norm": 0.3251045346260071,
    "learning_rate": 7.24527723855037e-05,
    "entropy": 0.6028687685728074,
    "num_tokens": 12711531.0,
    "mean_token_accuracy": 0.8390907049179077,
    "epoch": 0.3742133015819017,
    "step": 1100
  },
  {
    "loss": 0.6078,
    "grad_norm": 0.3440212905406952,
    "learning_rate": 7.195721415449508e-05,
    "entropy": 0.61930713057518,
    "num_tokens": 12839137.0,
    "mean_token_accuracy": 0.8350177526473999,
    "epoch": 0.3776152406871917,
    "step": 1110
  },
  {
    "loss": 0.626,
    "grad_norm": 0.3647129535675049,
    "learning_rate": 7.14589691201782e-05,
    "entropy": 0.6314490288496017,
    "num_tokens": 12947358.0,
    "mean_token_accuracy": 0.8336542189121247,
    "epoch": 0.38101717979248173,
    "step": 1120
  },
  {
    "loss": 0.5787,
    "grad_norm": 0.3373214304447174,
    "learning_rate": 7.095809825050626e-05,
    "entropy": 0.5762265086174011,
    "num_tokens": 13066371.0,
    "mean_token_accuracy": 0.8430596113204956,
    "epoch": 0.38441911889777175,
    "step": 1130
  },
  {
    "loss": 0.6219,
    "grad_norm": 0.3417420983314514,
    "learning_rate": 7.045466283474397e-05,
    "entropy": 0.6200504913926125,
    "num_tokens": 13187362.0,
    "mean_token_accuracy": 0.8362764894962311,
    "epoch": 0.38782105800306177,
    "step": 1140
  },
  {
    "loss": 0.6018,
    "grad_norm": 0.37871527671813965,
    "learning_rate": 6.99487244759677e-05,
    "entropy": 0.6063272505998611,
    "num_tokens": 13299146.0,
    "mean_token_accuracy": 0.8369143933057785,
    "epoch": 0.39122299710835173,
    "step": 1150
  },
  {
    "loss": 0.6311,
    "grad_norm": 0.33044958114624023,
    "learning_rate": 6.944034508352745e-05,
    "entropy": 0.6377698928117752,
    "num_tokens": 13418154.0,
    "mean_token_accuracy": 0.831796658039093,
    "epoch": 0.39462493621364175,
    "step": 1160
  },
  {
    "loss": 0.5661,
    "grad_norm": 0.34921640157699585,
    "learning_rate": 6.892958686547129e-05,
    "entropy": 0.5706936493515968,
    "num_tokens": 13532234.0,
    "mean_token_accuracy": 0.8478675246238708,
    "epoch": 0.39802687531893177,
    "step": 1170
  },
  {
    "loss": 0.6017,
    "grad_norm": 0.3906354010105133,
    "learning_rate": 6.841651232093321e-05,
    "entropy": 0.604531979560852,
    "num_tokens": 13647265.0,
    "mean_token_accuracy": 0.8396757930517197,
    "epoch": 0.4014288144242218,
    "step": 1180
  },
  {
    "loss": 0.5594,
    "grad_norm": 0.35730934143066406,
    "learning_rate": 6.790118423248545e-05,
    "entropy": 0.5594036534428597,
    "num_tokens": 13762753.0,
    "mean_token_accuracy": 0.8494072288274765,
    "epoch": 0.4048307535295118,
    "step": 1190
  },
  {
    "loss": 0.6035,
    "grad_norm": 0.40569987893104553,
    "learning_rate": 6.738366565845609e-05,
    "entropy": 0.6062624737620353,
    "num_tokens": 13887573.0,
    "mean_token_accuracy": 0.8375277072191238,
    "epoch": 0.40823269263480183,
    "step": 1200
  },
  {
    "loss": 0.6,
    "grad_norm": 0.3749787509441376,
    "learning_rate": 6.686401992521274e-05,
    "entropy": 0.5981257140636445,
    "num_tokens": 14009093.0,
    "mean_token_accuracy": 0.839872732758522,
    "epoch": 0.41163463174009185,
    "step": 1210
  },
  {
    "loss": 0.5752,
    "grad_norm": 0.34199485182762146,
    "learning_rate": 6.634231061941383e-05,
    "entropy": 0.5763212725520134,
    "num_tokens": 14133014.0,
    "mean_token_accuracy": 0.8439627707004547,
    "epoch": 0.41503657084538187,
    "step": 1220
  },
  {
    "loss": 0.5881,
    "grad_norm": 0.38227760791778564,
    "learning_rate": 6.581860158022757e-05,
    "entropy": 0.5911096960306168,
    "num_tokens": 14243866.0,
    "mean_token_accuracy": 0.8422502607107163,
    "epoch": 0.4184385099506719,
    "step": 1230
  },
  {
    "loss": 0.5924,
    "grad_norm": 0.31732606887817383,
    "learning_rate": 6.529295689152042e-05,
    "entropy": 0.6007756769657135,
    "num_tokens": 14354926.0,
    "mean_token_accuracy": 0.841707193851471,
    "epoch": 0.4218404490559619,
    "step": 1240
  },
  {
    "loss": 0.564,
    "grad_norm": 0.3495018780231476,
    "learning_rate": 6.476544087401532e-05,
    "entropy": 0.5664652720093727,
    "num_tokens": 14472132.0,
    "mean_token_accuracy": 0.845953956246376,
    "epoch": 0.4252423881612519,
    "step": 1250
  },
  {
    "loss": 0.6371,
    "grad_norm": 0.39551210403442383,
    "learning_rate": 6.423611807742116e-05,
    "entropy": 0.6342313542962075,
    "num_tokens": 14595618.0,
    "mean_token_accuracy": 0.8320615291595459,
    "epoch": 0.42864432726654195,
    "step": 1260
  },
  {
    "loss": 0.5862,
    "grad_norm": 0.3622702360153198,
    "learning_rate": 6.3705053272534e-05,
    "entropy": 0.5873361572623252,
    "num_tokens": 14715491.0,
    "mean_token_accuracy": 0.8399784296751023,
    "epoch": 0.43204626637183197,
    "step": 1270
  },
  {
    "loss": 0.5878,
    "grad_norm": 0.3421778082847595,
    "learning_rate": 6.317231144331153e-05,
    "entropy": 0.591562668979168,
    "num_tokens": 14829063.0,
    "mean_token_accuracy": 0.8422017335891724,
    "epoch": 0.435448205477122,
    "step": 1280
  },
  {
    "loss": 0.556,
    "grad_norm": 0.33399707078933716,
    "learning_rate": 6.263795777892115e-05,
    "entropy": 0.5609270349144936,
    "num_tokens": 14942097.0,
    "mean_token_accuracy": 0.8491690754890442,
    "epoch": 0.43885014458241195,
    "step": 1290
  },
  {
    "loss": 0.598,
    "grad_norm": 0.33439555764198303,
    "learning_rate": 6.210205766576308e-05,
    "entropy": 0.6013634830713273,
    "num_tokens": 15063270.0,
    "mean_token_accuracy": 0.8379765659570694,
    "epoch": 0.44225208368770197,
    "step": 1300
  },
  {
    "loss": 0.5739,
    "grad_norm": 0.34527984261512756,
    "learning_rate": 6.156467667946944e-05,
    "entropy": 0.5805606454610824,
    "num_tokens": 15182142.0,
    "mean_token_accuracy": 0.8448573082685471,
    "epoch": 0.445654022792992,
    "step": 1310
  },
  {
    "loss": 0.6355,
    "grad_norm": 0.32778117060661316,
    "learning_rate": 6.1025880576879934e-05,
    "entropy": 0.6364081218838692,
    "num_tokens": 15301327.0,
    "mean_token_accuracy": 0.8308557868003845,
    "epoch": 0.449055961898282,
    "step": 1320
  },
  {
    "loss": 0.6181,
    "grad_norm": 0.34031936526298523,
    "learning_rate": 6.048573528799556e-05,
    "entropy": 0.6205928295850753,
    "num_tokens": 15420327.0,
    "mean_token_accuracy": 0.834335333108902,
    "epoch": 0.452457901003572,
    "step": 1330
  },
  {
    "loss": 0.5708,
    "grad_norm": 0.40162762999534607,
    "learning_rate": 5.994430690791102e-05,
    "entropy": 0.572858938574791,
    "num_tokens": 15527631.0,
    "mean_token_accuracy": 0.8488004267215729,
    "epoch": 0.45585984010886205,
    "step": 1340
  },
  {
    "loss": 0.6002,
    "grad_norm": 0.3677639961242676,
    "learning_rate": 5.9401661688726986e-05,
    "entropy": 0.6025476709008217,
    "num_tokens": 15645227.0,
    "mean_token_accuracy": 0.8378155291080475,
    "epoch": 0.45926177921415207,
    "step": 1350
  },
  {
    "loss": 0.6391,
    "grad_norm": 0.3611242175102234,
    "learning_rate": 5.8857866031443155e-05,
    "entropy": 0.6453688651323318,
    "num_tokens": 15755942.0,
    "mean_token_accuracy": 0.8302028566598892,
    "epoch": 0.4626637183194421,
    "step": 1360
  },
  {
    "loss": 0.5891,
    "grad_norm": 0.3595288097858429,
    "learning_rate": 5.8312986477833035e-05,
    "entropy": 0.6027933269739151,
    "num_tokens": 15862604.0,
    "mean_token_accuracy": 0.842942762374878,
    "epoch": 0.4660656574247321,
    "step": 1370
  },
  {
    "loss": 0.5901,
    "grad_norm": 0.3880114257335663,
    "learning_rate": 5.7767089702301526e-05,
    "entropy": 0.5783761963248253,
    "num_tokens": 15974317.0,
    "mean_token_accuracy": 0.8423945069313049,
    "epoch": 0.4694675965300221,
    "step": 1380
  },
  {
    "loss": 0.5735,
    "grad_norm": 0.33461010456085205,
    "learning_rate": 5.722024250372632e-05,
    "entropy": 0.5808866828680038,
    "num_tokens": 16083489.0,
    "mean_token_accuracy": 0.8458300411701203,
    "epoch": 0.47286953563531214,
    "step": 1390
  },
  {
    "loss": 0.5967,
    "grad_norm": 0.3592625856399536,
    "learning_rate": 5.667251179728398e-05,
    "entropy": 0.5973397344350815,
    "num_tokens": 16209893.0,
    "mean_token_accuracy": 0.8369288891553879,
    "epoch": 0.47627147474060216,
    "step": 1400
  },
  {
    "loss": 0.6051,
    "grad_norm": 0.356492817401886,
    "learning_rate": 5.612396460626188e-05,
    "entropy": 0.6047371000051498,
    "num_tokens": 16328112.0,
    "mean_token_accuracy": 0.8397100865840912,
    "epoch": 0.4796734138458922,
    "step": 1410
  },
  {
    "loss": 0.6526,
    "grad_norm": 0.3834919333457947,
    "learning_rate": 5.557466805385685e-05,
    "entropy": 0.6492822825908661,
    "num_tokens": 16439800.0,
    "mean_token_accuracy": 0.829655796289444,
    "epoch": 0.48307535295118215,
    "step": 1420
  },
  {
    "loss": 0.5926,
    "grad_norm": 0.35866421461105347,
    "learning_rate": 5.502468935496164e-05,
    "entropy": 0.6050204545259475,
    "num_tokens": 16553741.0,
    "mean_token_accuracy": 0.8413259029388428,
    "epoch": 0.48647729205647217,
    "step": 1430
  },
  {
    "loss": 0.5734,
    "grad_norm": 0.3319409191608429,
    "learning_rate": 5.4474095807940094e-05,
    "entropy": 0.5755202546715736,
    "num_tokens": 16673973.0,
    "mean_token_accuracy": 0.8435852289199829,
    "epoch": 0.4898792311617622,
    "step": 1440
  },
  {
    "loss": 0.586,
    "grad_norm": 0.3485705554485321,
    "learning_rate": 5.392295478639225e-05,
    "entropy": 0.5858001261949539,
    "num_tokens": 16783732.0,
    "mean_token_accuracy": 0.8417124092578888,
    "epoch": 0.4932811702670522,
    "step": 1450
  },
  {
    "loss": 0.612,
    "grad_norm": 0.3999135494232178,
    "learning_rate": 5.3371333730910035e-05,
    "entropy": 0.6123771637678146,
    "num_tokens": 16899108.0,
    "mean_token_accuracy": 0.8361808478832244,
    "epoch": 0.4966831093723422,
    "step": 1460
  },
  {
    "loss": 0.5768,
    "grad_norm": 0.37186917662620544,
    "learning_rate": 5.2819300140824924e-05,
    "entropy": 0.5769275650382042,
    "num_tokens": 17015498.0,
    "mean_token_accuracy": 0.8437288701534271,
    "epoch": 0.5000850484776322,
    "step": 1470
  },
  {
    "loss": 0.6068,
    "grad_norm": 0.36763256788253784,
    "learning_rate": 5.22669215659484e-05,
    "entropy": 0.6088968113064765,
    "num_tokens": 17122826.0,
    "mean_token_accuracy": 0.8394175559282303,
    "epoch": 0.5034869875829223,
    "step": 1480
  },
  {
    "loss": 0.6024,
    "grad_norm": 0.33930572867393494,
    "learning_rate": 5.17142655983061e-05,
    "entropy": 0.6036048322916031,
    "num_tokens": 17248073.0,
    "mean_token_accuracy": 0.838006791472435,
    "epoch": 0.5068889266882123,
    "step": 1490
  },
  {
    "loss": 0.5755,
    "grad_norm": 0.355539470911026,
    "learning_rate": 5.116139986386695e-05,
    "entropy": 0.57216976583004,
    "num_tokens": 17353849.0,
    "mean_token_accuracy": 0.8462014138698578,
    "epoch": 0.5102908657935022,
    "step": 1500
  },
  {
    "loss": 0.5785,
    "grad_norm": 0.3130170702934265,
    "learning_rate": 5.0608392014268093e-05,
    "entropy": 0.5841953173279762,
    "num_tokens": 17465362.0,
    "mean_token_accuracy": 0.8455700755119324,
    "epoch": 0.5136928048987923,
    "step": 1510
  },
  {
    "loss": 0.5854,
    "grad_norm": 0.34390658140182495,
    "learning_rate": 5.005530971853661e-05,
    "entropy": 0.588604536652565,
    "num_tokens": 17583990.0,
    "mean_token_accuracy": 0.8393468260765076,
    "epoch": 0.5170947440040823,
    "step": 1520
  },
  {
    "loss": 0.5412,
    "grad_norm": 0.3938881456851959,
    "learning_rate": 4.950222065480926e-05,
    "entropy": 0.5458601832389831,
    "num_tokens": 17695756.0,
    "mean_token_accuracy": 0.8516582667827606,
    "epoch": 0.5204966831093724,
    "step": 1530
  },
  {
    "loss": 0.6066,
    "grad_norm": 0.36709338426589966,
    "learning_rate": 4.894919250205095e-05,
    "entropy": 0.6080822870135307,
    "num_tokens": 17819124.0,
    "mean_token_accuracy": 0.8351925998926163,
    "epoch": 0.5238986222146623,
    "step": 1540
  },
  {
    "loss": 0.603,
    "grad_norm": 0.3591138422489166,
    "learning_rate": 4.8396292931773194e-05,
    "entropy": 0.6067529514431953,
    "num_tokens": 17942151.0,
    "mean_token_accuracy": 0.8357031852006912,
    "epoch": 0.5273005613199524,
    "step": 1550
  },
  {
    "loss": 0.5889,
    "grad_norm": 0.32743504643440247,
    "learning_rate": 4.7843589599753436e-05,
    "entropy": 0.5867392048239708,
    "num_tokens": 18051775.0,
    "mean_token_accuracy": 0.8451169669628144,
    "epoch": 0.5307025004252424,
    "step": 1560
  },
  {
    "loss": 0.5827,
    "grad_norm": 0.34111034870147705,
    "learning_rate": 4.729115013775639e-05,
    "entropy": 0.5817334815859795,
    "num_tokens": 18174207.0,
    "mean_token_accuracy": 0.8418365478515625,
    "epoch": 0.5341044395305324,
    "step": 1570
  },
  {
    "loss": 0.5605,
    "grad_norm": 0.40375614166259766,
    "learning_rate": 4.673904214525817e-05,
    "entropy": 0.5690050527453423,
    "num_tokens": 18296847.0,
    "mean_token_accuracy": 0.8457064211368561,
    "epoch": 0.5375063786358224,
    "step": 1580
  },
  {
    "loss": 0.6214,
    "grad_norm": 0.34886664152145386,
    "learning_rate": 4.618733318117453e-05,
    "entropy": 0.613763065636158,
    "num_tokens": 18411175.0,
    "mean_token_accuracy": 0.8373794168233871,
    "epoch": 0.5409083177411125,
    "step": 1590
  },
  {
    "loss": 0.5932,
    "grad_norm": 0.4052237868309021,
    "learning_rate": 4.56360907555939e-05,
    "entropy": 0.5990845367312432,
    "num_tokens": 18523679.0,
    "mean_token_accuracy": 0.8415322303771973,
    "epoch": 0.5443102568464024,
    "step": 1600
  },
  {
    "loss": 0.5587,
    "grad_norm": 0.33428770303726196,
    "learning_rate": 4.508538232151659e-05,
    "entropy": 0.5686810597777366,
    "num_tokens": 18638795.0,
    "mean_token_accuracy": 0.8485085606575012,
    "epoch": 0.5477121959516925,
    "step": 1610
  },
  {
    "loss": 0.5528,
    "grad_norm": 0.3429810106754303,
    "learning_rate": 4.453527526660079e-05,
    "entropy": 0.5541219785809517,
    "num_tokens": 18757690.0,
    "mean_token_accuracy": 0.8475034803152084,
    "epoch": 0.5511141350569825,
    "step": 1620
  },
  {
    "loss": 0.5896,
    "grad_norm": 0.36002102494239807,
    "learning_rate": 4.398583690491673e-05,
    "entropy": 0.594765892624855,
    "num_tokens": 18870237.0,
    "mean_token_accuracy": 0.841697883605957,
    "epoch": 0.5545160741622724,
    "step": 1630
  },
  {
    "loss": 0.5583,
    "grad_norm": 0.3443513810634613,
    "learning_rate": 4.3437134468709684e-05,
    "entropy": 0.5576672151684761,
    "num_tokens": 18989654.0,
    "mean_token_accuracy": 0.8470352560281753,
    "epoch": 0.5579180132675625,
    "step": 1640
  },
  {
    "loss": 0.5715,
    "grad_norm": 0.3210230767726898,
    "learning_rate": 4.2889235100173117e-05,
    "entropy": 0.5803326085209847,
    "num_tokens": 19111873.0,
    "mean_token_accuracy": 0.8444038182497025,
    "epoch": 0.5613199523728525,
    "step": 1650
  },
  {
    "loss": 0.579,
    "grad_norm": 0.3320671617984772,
    "learning_rate": 4.2342205843232815e-05,
    "entropy": 0.5725411295890808,
    "num_tokens": 19227912.0,
    "mean_token_accuracy": 0.8457438528537751,
    "epoch": 0.5647218914781426,
    "step": 1660
  },
  {
    "loss": 0.587,
    "grad_norm": 0.38086163997650146,
    "learning_rate": 4.1796113635342995e-05,
    "entropy": 0.5874329909682274,
    "num_tokens": 19349977.0,
    "mean_token_accuracy": 0.840248167514801,
    "epoch": 0.5681238305834325,
    "step": 1670
  },
  {
    "loss": 0.5847,
    "grad_norm": 0.3494766652584076,
    "learning_rate": 4.1251025299295484e-05,
    "entropy": 0.5962051749229431,
    "num_tokens": 19463971.0,
    "mean_token_accuracy": 0.8411266356706619,
    "epoch": 0.5715257696887226,
    "step": 1680
  },
  {
    "loss": 0.5886,
    "grad_norm": 0.34530606865882874,
    "learning_rate": 4.0707007535042965e-05,
    "entropy": 0.5932681888341904,
    "num_tokens": 19584720.0,
    "mean_token_accuracy": 0.8420757114887237,
    "epoch": 0.5749277087940126,
    "step": 1690
  },
  {
    "loss": 0.6255,
    "grad_norm": 0.39939698576927185,
    "learning_rate": 4.0164126911537124e-05,
    "entropy": 0.6298390612006187,
    "num_tokens": 19707943.0,
    "mean_token_accuracy": 0.8327343821525574,
    "epoch": 0.5783296478993026,
    "step": 1700
  },
  {
    "loss": 0.5829,
    "grad_norm": 0.39982014894485474,
    "learning_rate": 3.9622449858583e-05,
    "entropy": 0.5805615946650505,
    "num_tokens": 19821092.0,
    "mean_token_accuracy": 0.8441908627748489,
    "epoch": 0.5817315870045926,
    "step": 1710
  },
  {
    "loss": 0.5546,
    "grad_norm": 0.3315586447715759,
    "learning_rate": 3.908204265871024e-05,
    "entropy": 0.5536176040768623,
    "num_tokens": 19926967.0,
    "mean_token_accuracy": 0.8505498200654984,
    "epoch": 0.5851335261098827,
    "step": 1720
  },
  {
    "loss": 0.6144,
    "grad_norm": 0.38606083393096924,
    "learning_rate": 3.8542971439062375e-05,
    "entropy": 0.6170039817690849,
    "num_tokens": 20050273.0,
    "mean_token_accuracy": 0.8357849180698395,
    "epoch": 0.5885354652151726,
    "step": 1730
  },
  {
    "loss": 0.6018,
    "grad_norm": 0.35552018880844116,
    "learning_rate": 3.800530216330522e-05,
    "entropy": 0.6068514689803124,
    "num_tokens": 20168367.0,
    "mean_token_accuracy": 0.8391161382198333,
    "epoch": 0.5919374043204627,
    "step": 1740
  },
  {
    "loss": 0.6014,
    "grad_norm": 0.408333957195282,
    "learning_rate": 3.746910062355514e-05,
    "entropy": 0.6115496993064881,
    "num_tokens": 20275208.0,
    "mean_token_accuracy": 0.837997031211853,
    "epoch": 0.5953393434257527,
    "step": 1750
  },
  {
    "loss": 0.5656,
    "grad_norm": 0.3675805628299713,
    "learning_rate": 3.693443243232839e-05,
    "entropy": 0.5659057840704917,
    "num_tokens": 20395513.0,
    "mean_token_accuracy": 0.846270963549614,
    "epoch": 0.5987412825310426,
    "step": 1760
  },
  {
    "loss": 0.5876,
    "grad_norm": 0.34130120277404785,
    "learning_rate": 3.6401363014512465e-05,
    "entropy": 0.5852897882461547,
    "num_tokens": 20513840.0,
    "mean_token_accuracy": 0.8406417429447174,
    "epoch": 0.6021432216363327,
    "step": 1770
  },
  {
    "loss": 0.5846,
    "grad_norm": 0.3448680341243744,
    "learning_rate": 3.586995759936027e-05,
    "entropy": 0.5861192226409913,
    "num_tokens": 20626418.0,
    "mean_token_accuracy": 0.8443343281745911,
    "epoch": 0.6055451607416227,
    "step": 1780
  },
  {
    "loss": 0.5629,
    "grad_norm": 0.33498626947402954,
    "learning_rate": 3.5340281212508355e-05,
    "entropy": 0.5744142100214958,
    "num_tokens": 20743700.0,
    "mean_token_accuracy": 0.8447044044733047,
    "epoch": 0.6089470998469128,
    "step": 1790
  },
  {
    "loss": 0.5491,
    "grad_norm": 0.3723412752151489,
    "learning_rate": 3.481239866802003e-05,
    "entropy": 0.5503455892205238,
    "num_tokens": 20848890.0,
    "mean_token_accuracy": 0.8521771311759949,
    "epoch": 0.6123490389522027,
    "step": 1800
  },
  {
    "loss": 0.5728,
    "grad_norm": 0.37624675035476685,
    "learning_rate": 3.428637456045438e-05,
    "entropy": 0.5796245783567429,
    "num_tokens": 20959919.0,
    "mean_token_accuracy": 0.8451456815004349,
    "epoch": 0.6157509780574928,
    "step": 1810
  },
  {
    "loss": 0.6005,
    "grad_norm": 0.4184034466743469,
    "learning_rate": 3.3762273256962115e-05,
    "entropy": 0.6072939783334732,
    "num_tokens": 21068427.0,
    "mean_token_accuracy": 0.8402355641126633,
    "epoch": 0.6191529171627828,
    "step": 1820
  },
  {
    "loss": 0.6061,
    "grad_norm": 0.38452163338661194,
    "learning_rate": 3.324015888940925e-05,
    "entropy": 0.6065735876560211,
    "num_tokens": 21184885.0,
    "mean_token_accuracy": 0.8360267549753189,
    "epoch": 0.6225548562680728,
    "step": 1830
  },
  {
    "loss": 0.5824,
    "grad_norm": 0.4377152919769287,
    "learning_rate": 3.2720095346529566e-05,
    "entropy": 0.5836168915033341,
    "num_tokens": 21298789.0,
    "mean_token_accuracy": 0.8420229732990265,
    "epoch": 0.6259567953733628,
    "step": 1840
  },
  {
    "loss": 0.5761,
    "grad_norm": 0.3785684406757355,
    "learning_rate": 3.220214626610689e-05,
    "entropy": 0.5765029028058052,
    "num_tokens": 21424296.0,
    "mean_token_accuracy": 0.8433551669120789,
    "epoch": 0.6293587344786529,
    "step": 1850
  },
  {
    "loss": 0.572,
    "grad_norm": 0.37199345231056213,
    "learning_rate": 3.168637502718798e-05,
    "entropy": 0.5756848677992821,
    "num_tokens": 21534093.0,
    "mean_token_accuracy": 0.8463689625263214,
    "epoch": 0.6327606735839428,
    "step": 1860
  },
  {
    "loss": 0.5863,
    "grad_norm": 0.38243773579597473,
    "learning_rate": 3.117284474232717e-05,
    "entropy": 0.5788663312792778,
    "num_tokens": 21656220.0,
    "mean_token_accuracy": 0.8417115867137909,
    "epoch": 0.6361626126892329,
    "step": 1870
  },
  {
    "loss": 0.5986,
    "grad_norm": 0.3821103274822235,
    "learning_rate": 3.066161824986352e-05,
    "entropy": 0.5999485164880752,
    "num_tokens": 21770257.0,
    "mean_token_accuracy": 0.8380926698446274,
    "epoch": 0.6395645517945229,
    "step": 1880
  },
  {
    "loss": 0.577,
    "grad_norm": 0.33614298701286316,
    "learning_rate": 3.0152758106231642e-05,
    "entropy": 0.5815630614757538,
    "num_tokens": 21896451.0,
    "mean_token_accuracy": 0.8430263727903367,
    "epoch": 0.642966490899813,
    "step": 1890
  },
  {
    "loss": 0.6131,
    "grad_norm": 0.39611414074897766,
    "learning_rate": 2.9646326578306915e-05,
    "entropy": 0.6189764007925987,
    "num_tokens": 22003560.0,
    "mean_token_accuracy": 0.8391340941190719,
    "epoch": 0.6463684300051029,
    "step": 1900
  },
  {
    "loss": 0.5443,
    "grad_norm": 0.33414164185523987,
    "learning_rate": 2.914238563578618e-05,
    "entropy": 0.5429508209228515,
    "num_tokens": 22117872.0,
    "mean_token_accuracy": 0.8526928842067718,
    "epoch": 0.6497703691103929,
    "step": 1910
  },
  {
    "loss": 0.6123,
    "grad_norm": 0.4099610447883606,
    "learning_rate": 2.864099694360479e-05,
    "entropy": 0.6225771173834801,
    "num_tokens": 22222509.0,
    "mean_token_accuracy": 0.8378654181957245,
    "epoch": 0.653172308215683,
    "step": 1920
  },
  {
    "loss": 0.5621,
    "grad_norm": 0.3818487823009491,
    "learning_rate": 2.814222185439098e-05,
    "entropy": 0.5593055367469788,
    "num_tokens": 22335582.0,
    "mean_token_accuracy": 0.8501265943050385,
    "epoch": 0.6565742473209729,
    "step": 1930
  },
  {
    "loss": 0.6008,
    "grad_norm": 0.4001128673553467,
    "learning_rate": 2.764612140095839e-05,
    "entropy": 0.6034526154398918,
    "num_tokens": 22448701.0,
    "mean_token_accuracy": 0.8390175521373748,
    "epoch": 0.659976186426263,
    "step": 1940
  },
  {
    "loss": 0.5682,
    "grad_norm": 0.3717249035835266,
    "learning_rate": 2.715275628883781e-05,
    "entropy": 0.5757593274116516,
    "num_tokens": 22565429.0,
    "mean_token_accuracy": 0.8461737126111984,
    "epoch": 0.663378125531553,
    "step": 1950
  },
  {
    "loss": 0.5819,
    "grad_norm": 0.31234627962112427,
    "learning_rate": 2.6662186888848862e-05,
    "entropy": 0.5898503661155701,
    "num_tokens": 22690809.0,
    "mean_token_accuracy": 0.8421528667211533,
    "epoch": 0.666780064636843,
    "step": 1960
  },
  {
    "loss": 0.5795,
    "grad_norm": 0.43498075008392334,
    "learning_rate": 2.6174473229712758e-05,
    "entropy": 0.5725113779306412,
    "num_tokens": 22808337.0,
    "mean_token_accuracy": 0.8442942827939988,
    "epoch": 0.670182003742133,
    "step": 1970
  },
  {
    "loss": 0.5602,
    "grad_norm": 0.35489723086357117,
    "learning_rate": 2.5689674990706813e-05,
    "entropy": 0.5673380747437478,
    "num_tokens": 22924233.0,
    "mean_token_accuracy": 0.8477986544370651,
    "epoch": 0.6735839428474231,
    "step": 1980
  },
  {
    "loss": 0.616,
    "grad_norm": 0.3546064496040344,
    "learning_rate": 2.520785149436181e-05,
    "entropy": 0.6025599181652069,
    "num_tokens": 23040709.0,
    "mean_token_accuracy": 0.8388166666030884,
    "epoch": 0.676985881952713,
    "step": 1990
  },
  {
    "loss": 0.5866,
    "grad_norm": 0.3466935455799103,
    "learning_rate": 2.4729061699202933e-05,
    "entropy": 0.5931081950664521,
    "num_tokens": 23168281.0,
    "mean_token_accuracy": 0.8412128478288651,
    "epoch": 0.6803878210580031,
    "step": 2000
  },
  {
    "loss": 0.5745,
    "grad_norm": 0.36624687910079956,
    "learning_rate": 2.4253364192535348e-05,
    "entropy": 0.5797276377677918,
    "num_tokens": 23265650.0,
    "mean_token_accuracy": 0.8477521270513535,
    "epoch": 0.6837897601632931,
    "step": 2010
  },
  {
    "loss": 0.6003,
    "grad_norm": 0.3339534401893616,
    "learning_rate": 2.3780817183275057e-05,
    "entropy": 0.5999033793807029,
    "num_tokens": 23380810.0,
    "mean_token_accuracy": 0.8395819336175918,
    "epoch": 0.6871916992685831,
    "step": 2020
  },
  {
    "loss": 0.6001,
    "grad_norm": 0.3682357370853424,
    "learning_rate": 2.331147849482619e-05,
    "entropy": 0.6057513281702995,
    "num_tokens": 23488800.0,
    "mean_token_accuracy": 0.8401612430810929,
    "epoch": 0.6905936383738731,
    "step": 2030
  },
  {
    "loss": 0.5681,
    "grad_norm": 0.40995046496391296,
    "learning_rate": 2.2845405558005428e-05,
    "entropy": 0.5787547707557679,
    "num_tokens": 23603230.0,
    "mean_token_accuracy": 0.8451647937297821,
    "epoch": 0.6939955774791631,
    "step": 2040
  },
  {
    "loss": 0.5821,
    "grad_norm": 0.37340083718299866,
    "learning_rate": 2.2382655404014447e-05,
    "entropy": 0.5805535346269608,
    "num_tokens": 23722605.0,
    "mean_token_accuracy": 0.844177857041359,
    "epoch": 0.6973975165844531,
    "step": 2050
  },
  {
    "loss": 0.5787,
    "grad_norm": 0.4327142536640167,
    "learning_rate": 2.1923284657461258e-05,
    "entropy": 0.5828775107860565,
    "num_tokens": 23835961.0,
    "mean_token_accuracy": 0.842606109380722,
    "epoch": 0.7007994556897431,
    "step": 2060
  },
  {
    "loss": 0.5982,
    "grad_norm": 0.4304671287536621,
    "learning_rate": 2.1467349529431317e-05,
    "entropy": 0.5941863551735878,
    "num_tokens": 23949131.0,
    "mean_token_accuracy": 0.8400673478841781,
    "epoch": 0.7042013947950332,
    "step": 2070
  },
  {
    "loss": 0.5679,
    "grad_norm": 0.37865591049194336,
    "learning_rate": 2.10149058106093e-05,
    "entropy": 0.5765653476119041,
    "num_tokens": 24070739.0,
    "mean_token_accuracy": 0.8449007272720337,
    "epoch": 0.7076033339003232,
    "step": 2080
  },
  {
    "loss": 0.5886,
    "grad_norm": 0.3672359585762024,
    "learning_rate": 2.056600886445213e-05,
    "entropy": 0.5881167098879814,
    "num_tokens": 24178837.0,
    "mean_token_accuracy": 0.8434468775987625,
    "epoch": 0.7110052730056132,
    "step": 2090
  },
  {
    "loss": 0.5956,
    "grad_norm": 0.36181676387786865,
    "learning_rate": 2.01207136204145e-05,
    "entropy": 0.591875921189785,
    "num_tokens": 24297729.0,
    "mean_token_accuracy": 0.8414703428745269,
    "epoch": 0.7144072121109032,
    "step": 2100
  },
  {
    "loss": 0.5599,
    "grad_norm": 0.35588833689689636,
    "learning_rate": 1.967907456722738e-05,
    "entropy": 0.5663163229823113,
    "num_tokens": 24408243.0,
    "mean_token_accuracy": 0.8491099238395691,
    "epoch": 0.7178091512161933,
    "step": 2110
  },
  {
    "loss": 0.5644,
    "grad_norm": 0.38050028681755066,
    "learning_rate": 1.9241145746230478e-05,
    "entropy": 0.5726602435111999,
    "num_tokens": 24512247.0,
    "mean_token_accuracy": 0.8495869427919388,
    "epoch": 0.7212110903214832,
    "step": 2120
  },
  {
    "loss": 0.5693,
    "grad_norm": 0.3819943964481354,
    "learning_rate": 1.8806980744759445e-05,
    "entropy": 0.5744548887014389,
    "num_tokens": 24627020.0,
    "mean_token_accuracy": 0.842121011018753,
    "epoch": 0.7246130294267733,
    "step": 2130
  },
  {
    "loss": 0.6092,
    "grad_norm": 0.3454561233520508,
    "learning_rate": 1.837663268958867e-05,
    "entropy": 0.6237739801406861,
    "num_tokens": 24736398.0,
    "mean_token_accuracy": 0.8370961755514145,
    "epoch": 0.7280149685320633,
    "step": 2140
  },
  {
    "loss": 0.6223,
    "grad_norm": 0.35335883498191833,
    "learning_rate": 1.795015424043035e-05,
    "entropy": 0.6208432152867317,
    "num_tokens": 24861135.0,
    "mean_token_accuracy": 0.8354604303836822,
    "epoch": 0.7314169076373533,
    "step": 2150
  },
  {
    "loss": 0.5919,
    "grad_norm": 0.3348969519138336,
    "learning_rate": 1.7527597583490822e-05,
    "entropy": 0.5919017851352691,
    "num_tokens": 24974197.0,
    "mean_token_accuracy": 0.840225774049759,
    "epoch": 0.7348188467426433,
    "step": 2160
  },
  {
    "loss": 0.5706,
    "grad_norm": 0.3716595470905304,
    "learning_rate": 1.7109014425084725e-05,
    "entropy": 0.5833111733198166,
    "num_tokens": 25091846.0,
    "mean_token_accuracy": 0.8445227384567261,
    "epoch": 0.7382207858479333,
    "step": 2170
  },
  {
    "loss": 0.5679,
    "grad_norm": 0.32632264494895935,
    "learning_rate": 1.669445598530796e-05,
    "entropy": 0.5681742131710052,
    "num_tokens": 25210289.0,
    "mean_token_accuracy": 0.8463738948106766,
    "epoch": 0.7416227249532233,
    "step": 2180
  },
  {
    "loss": 0.6147,
    "grad_norm": 0.3877134621143341,
    "learning_rate": 1.628397299177013e-05,
    "entropy": 0.6165782377123833,
    "num_tokens": 25334926.0,
    "mean_token_accuracy": 0.8343934565782547,
    "epoch": 0.7450246640585133,
    "step": 2190
  },
  {
    "loss": 0.5806,
    "grad_norm": 0.38453352451324463,
    "learning_rate": 1.5877615673387215e-05,
    "entropy": 0.5809252455830574,
    "num_tokens": 25453005.0,
    "mean_token_accuracy": 0.8436485230922699,
    "epoch": 0.7484266031638034,
    "step": 2200
  },
  {
    "loss": 0.5584,
    "grad_norm": 0.3705204725265503,
    "learning_rate": 1.5475433754235312e-05,
    "entropy": 0.5579630300402642,
    "num_tokens": 25568378.0,
    "mean_token_accuracy": 0.8482485413551331,
    "epoch": 0.7518285422690933,
    "step": 2210
  },
  {
    "loss": 0.5846,
    "grad_norm": 0.4126192331314087,
    "learning_rate": 1.5077476447466115e-05,
    "entropy": 0.5949768751859665,
    "num_tokens": 25679312.0,
    "mean_token_accuracy": 0.8417560040950776,
    "epoch": 0.7552304813743834,
    "step": 2220
  },
  {
    "loss": 0.5992,
    "grad_norm": 0.369003564119339,
    "learning_rate": 1.4683792449284922e-05,
    "entropy": 0.5958976864814758,
    "num_tokens": 25802476.0,
    "mean_token_accuracy": 0.8378436207771301,
    "epoch": 0.7586324204796734,
    "step": 2230
  },
  {
    "loss": 0.5845,
    "grad_norm": 0.3608800768852234,
    "learning_rate": 1.429442993299191e-05,
    "entropy": 0.5774736061692238,
    "num_tokens": 25917526.0,
    "mean_token_accuracy": 0.8448121935129166,
    "epoch": 0.7620343595849635,
    "step": 2240
  },
  {
    "loss": 0.533,
    "grad_norm": 0.35715219378471375,
    "learning_rate": 1.3909436543087428e-05,
    "entropy": 0.5421435713768006,
    "num_tokens": 26022101.0,
    "mean_token_accuracy": 0.8561456590890885,
    "epoch": 0.7654362986902534,
    "step": 2250
  },
  {
    "loss": 0.5595,
    "grad_norm": 0.36197429895401,
    "learning_rate": 1.352885938944189e-05,
    "entropy": 0.564231875538826,
    "num_tokens": 26132618.0,
    "mean_token_accuracy": 0.8467835336923599,
    "epoch": 0.7688382377955435,
    "step": 2260
  },
  {
    "loss": 0.5394,
    "grad_norm": 0.34464767575263977,
    "learning_rate": 1.3152745041531201e-05,
    "entropy": 0.5447842225432395,
    "num_tokens": 26251170.0,
    "mean_token_accuracy": 0.852738481760025,
    "epoch": 0.7722401769008335,
    "step": 2270
  },
  {
    "loss": 0.6081,
    "grad_norm": 0.34223926067352295,
    "learning_rate": 1.2781139522738239e-05,
    "entropy": 0.6099189713597297,
    "num_tokens": 26366238.0,
    "mean_token_accuracy": 0.8368343204259873,
    "epoch": 0.7756421160061235,
    "step": 2280
  },
  {
    "loss": 0.5608,
    "grad_norm": 0.40201887488365173,
    "learning_rate": 1.2414088304721234e-05,
    "entropy": 0.5682180404663086,
    "num_tokens": 26481121.0,
    "mean_token_accuracy": 0.8468090683221817,
    "epoch": 0.7790440551114135,
    "step": 2290
  },
  {
    "loss": 0.5831,
    "grad_norm": 0.3707852363586426,
    "learning_rate": 1.2051636301849539e-05,
    "entropy": 0.5816125243902206,
    "num_tokens": 26591544.0,
    "mean_token_accuracy": 0.8441601723432541,
    "epoch": 0.7824459942167035,
    "step": 2300
  },
  {
    "loss": 0.644,
    "grad_norm": 0.42641323804855347,
    "learning_rate": 1.1693827865707729e-05,
    "entropy": 0.6465028315782547,
    "num_tokens": 26701423.0,
    "mean_token_accuracy": 0.8323942631483078,
    "epoch": 0.7858479333219935,
    "step": 2310
  },
  {
    "loss": 0.5624,
    "grad_norm": 0.3720209300518036,
    "learning_rate": 1.1340706779668464e-05,
    "entropy": 0.568995788693428,
    "num_tokens": 26821407.0,
    "mean_token_accuracy": 0.844905823469162,
    "epoch": 0.7892498724272835,
    "step": 2320
  },
  {
    "loss": 0.562,
    "grad_norm": 0.33470767736434937,
    "learning_rate": 1.099231625353494e-05,
    "entropy": 0.5680811256170273,
    "num_tokens": 26942232.0,
    "mean_token_accuracy": 0.8472270101308823,
    "epoch": 0.7926518115325736,
    "step": 2330
  },
  {
    "loss": 0.5871,
    "grad_norm": 0.35473379492759705,
    "learning_rate": 1.0648698918253474e-05,
    "entropy": 0.5944129660725593,
    "num_tokens": 27054063.0,
    "mean_token_accuracy": 0.8423294246196746,
    "epoch": 0.7960537506378635,
    "step": 2340
  },
  {
    "loss": 0.5813,
    "grad_norm": 0.336117684841156,
    "learning_rate": 1.0309896820697002e-05,
    "entropy": 0.5846550479531288,
    "num_tokens": 27170010.0,
    "mean_token_accuracy": 0.8445548832416534,
    "epoch": 0.7994556897431536,
    "step": 2350
  },
  {
    "loss": 0.5587,
    "grad_norm": 0.36509907245635986,
    "learning_rate": 9.975951418519941e-06,
    "entropy": 0.565837450325489,
    "num_tokens": 27288934.0,
    "mean_token_accuracy": 0.8460572957992554,
    "epoch": 0.8028576288484436,
    "step": 2360
  },
  {
    "loss": 0.5744,
    "grad_norm": 0.37850871682167053,
    "learning_rate": 9.646903575085236e-06,
    "entropy": 0.5862236008048057,
    "num_tokens": 27405172.0,
    "mean_token_accuracy": 0.8447532653808594,
    "epoch": 0.8062595679537337,
    "step": 2370
  },
  {
    "loss": 0.5813,
    "grad_norm": 0.3511059582233429,
    "learning_rate": 9.32279355446411e-06,
    "entropy": 0.5788961902260781,
    "num_tokens": 27518997.0,
    "mean_token_accuracy": 0.8418603748083114,
    "epoch": 0.8096615070590236,
    "step": 2380
  },
  {
    "loss": 0.5855,
    "grad_norm": 0.3253343999385834,
    "learning_rate": 9.003661016509102e-06,
    "entropy": 0.5858073383569717,
    "num_tokens": 27644298.0,
    "mean_token_accuracy": 0.8415432572364807,
    "epoch": 0.8130634461643137,
    "step": 2390
  },
  {
    "loss": 0.5595,
    "grad_norm": 0.36334091424942017,
    "learning_rate": 8.689545012001083e-06,
    "entropy": 0.5582752287387848,
    "num_tokens": 27753242.0,
    "mean_token_accuracy": 0.8509968429803848,
    "epoch": 0.8164653852696037,
    "step": 2400
  },
  {
    "loss": 0.5706,
    "grad_norm": 0.32456502318382263,
    "learning_rate": 8.380483977870834e-06,
    "entropy": 0.5666798010468483,
    "num_tokens": 27882376.0,
    "mean_token_accuracy": 0.84336639046669,
    "epoch": 0.8198673243748937,
    "step": 2410
  },
  {
    "loss": 0.5394,
    "grad_norm": 0.3463776707649231,
    "learning_rate": 8.076515732495626e-06,
    "entropy": 0.5584284573793411,
    "num_tokens": 28000308.0,
    "mean_token_accuracy": 0.8512558758258819,
    "epoch": 0.8232692634801837,
    "step": 2420
  },
  {
    "loss": 0.6075,
    "grad_norm": 0.35784080624580383,
    "learning_rate": 7.777677471071614e-06,
    "entropy": 0.6100872978568077,
    "num_tokens": 28115648.0,
    "mean_token_accuracy": 0.8378289490938187,
    "epoch": 0.8266712025854738,
    "step": 2430
  },
  {
    "loss": 0.5815,
    "grad_norm": 0.351990282535553,
    "learning_rate": 7.484005761062391e-06,
    "entropy": 0.5808347776532173,
    "num_tokens": 28222592.0,
    "mean_token_accuracy": 0.8474200069904327,
    "epoch": 0.8300731416907637,
    "step": 2440
  },
  {
    "loss": 0.5889,
    "grad_norm": 0.3908417224884033,
    "learning_rate": 7.195536537724429e-06,
    "entropy": 0.5958179652690887,
    "num_tokens": 28325239.0,
    "mean_token_accuracy": 0.845181006193161,
    "epoch": 0.8334750807960537,
    "step": 2450
  },
  {
    "loss": 0.5846,
    "grad_norm": 0.38566815853118896,
    "learning_rate": 6.912305099709831e-06,
    "entropy": 0.5907109215855598,
    "num_tokens": 28445952.0,
    "mean_token_accuracy": 0.84410340487957,
    "epoch": 0.8368770199013438,
    "step": 2460
  },
  {
    "loss": 0.5876,
    "grad_norm": 0.4085163474082947,
    "learning_rate": 6.634346104746997e-06,
    "entropy": 0.5830874457955361,
    "num_tokens": 28554611.0,
    "mean_token_accuracy": 0.8452066659927369,
    "epoch": 0.8402789590066337,
    "step": 2470
  },
  {
    "loss": 0.5575,
    "grad_norm": 0.352285236120224,
    "learning_rate": 6.36169356539974e-06,
    "entropy": 0.5569544598460198,
    "num_tokens": 28659685.0,
    "mean_token_accuracy": 0.851721253991127,
    "epoch": 0.8436808981119238,
    "step": 2480
  },
  {
    "loss": 0.5667,
    "grad_norm": 0.365980863571167,
    "learning_rate": 6.094380844905278e-06,
    "entropy": 0.5703560188412666,
    "num_tokens": 28768833.0,
    "mean_token_accuracy": 0.8493399202823639,
    "epoch": 0.8470828372172138,
    "step": 2490
  },
  {
    "loss": 0.6135,
    "grad_norm": 0.4027390480041504,
    "learning_rate": 5.832440653091775e-06,
    "entropy": 0.6129578873515129,
    "num_tokens": 28885843.0,
    "mean_token_accuracy": 0.8367623895406723,
    "epoch": 0.8504847763225039,
    "step": 2500
  },
  {
    "loss": 0.559,
    "grad_norm": 0.37476834654808044,
    "learning_rate": 5.575905042375751e-06,
    "entropy": 0.5681184902787209,
    "num_tokens": 29003923.0,
    "mean_token_accuracy": 0.8465311348438262,
    "epoch": 0.8538867154277938,
    "step": 2510
  },
  {
    "loss": 0.5571,
    "grad_norm": 0.3746230900287628,
    "learning_rate": 5.324805403840006e-06,
    "entropy": 0.5609279617667198,
    "num_tokens": 29114669.0,
    "mean_token_accuracy": 0.8491218000650406,
    "epoch": 0.8572886545330839,
    "step": 2520
  },
  {
    "loss": 0.6017,
    "grad_norm": 0.3442933261394501,
    "learning_rate": 5.079172463392434e-06,
    "entropy": 0.6076221778988838,
    "num_tokens": 29237873.0,
    "mean_token_accuracy": 0.8369300574064255,
    "epoch": 0.8606905936383739,
    "step": 2530
  },
  {
    "loss": 0.5838,
    "grad_norm": 0.34637150168418884,
    "learning_rate": 4.839036278006215e-06,
    "entropy": 0.5847443833947181,
    "num_tokens": 29356084.0,
    "mean_token_accuracy": 0.8435145944356919,
    "epoch": 0.8640925327436639,
    "step": 2540
  },
  {
    "loss": 0.5735,
    "grad_norm": 0.3903183937072754,
    "learning_rate": 4.6044262320418915e-06,
    "entropy": 0.5809785157442093,
    "num_tokens": 29466199.0,
    "mean_token_accuracy": 0.8461235493421555,
    "epoch": 0.8674944718489539,
    "step": 2550
  },
  {
    "loss": 0.5468,
    "grad_norm": 0.38433846831321716,
    "learning_rate": 4.375371033651754e-06,
    "entropy": 0.5522323742508888,
    "num_tokens": 29585003.0,
    "mean_token_accuracy": 0.8522881597280503,
    "epoch": 0.870896410954244,
    "step": 2560
  },
  {
    "loss": 0.5747,
    "grad_norm": 0.38316574692726135,
    "learning_rate": 4.151898711266932e-06,
    "entropy": 0.5752411767840385,
    "num_tokens": 29697951.0,
    "mean_token_accuracy": 0.8458108186721802,
    "epoch": 0.8742983500595339,
    "step": 2570
  },
  {
    "loss": 0.5561,
    "grad_norm": 0.37862345576286316,
    "learning_rate": 3.934036610167696e-06,
    "entropy": 0.5635033801198006,
    "num_tokens": 29814806.0,
    "mean_token_accuracy": 0.8486339598894119,
    "epoch": 0.8777002891648239,
    "step": 2580
  },
  {
    "loss": 0.5879,
    "grad_norm": 0.40011560916900635,
    "learning_rate": 3.721811389137353e-06,
    "entropy": 0.5901739224791527,
    "num_tokens": 29938190.0,
    "mean_token_accuracy": 0.8400143891572952,
    "epoch": 0.881102228270114,
    "step": 2590
  },
  {
    "loss": 0.5741,
    "grad_norm": 0.36540430784225464,
    "learning_rate": 3.5152490172001116e-06,
    "entropy": 0.5728347212076187,
    "num_tokens": 30057305.0,
    "mean_token_accuracy": 0.8454901963472367,
    "epoch": 0.8845041673754039,
    "step": 2600
  },
  {
    "loss": 0.6221,
    "grad_norm": 0.3713492751121521,
    "learning_rate": 3.314374770443379e-06,
    "entropy": 0.6258899614214897,
    "num_tokens": 30172939.0,
    "mean_token_accuracy": 0.8357624322175979,
    "epoch": 0.887906106480694,
    "step": 2610
  },
  {
    "loss": 0.5946,
    "grad_norm": 0.3445519208908081,
    "learning_rate": 3.1192132289248497e-06,
    "entropy": 0.596692156791687,
    "num_tokens": 30300574.0,
    "mean_token_accuracy": 0.8411296546459198,
    "epoch": 0.891308045585984,
    "step": 2620
  },
  {
    "loss": 0.5774,
    "grad_norm": 0.37687185406684875,
    "learning_rate": 2.9297882736647574e-06,
    "entropy": 0.5851211875677109,
    "num_tokens": 30411366.0,
    "mean_token_accuracy": 0.8454822093248368,
    "epoch": 0.894709984691274,
    "step": 2630
  },
  {
    "loss": 0.6138,
    "grad_norm": 0.3559032678604126,
    "learning_rate": 2.746123083723656e-06,
    "entropy": 0.612051984667778,
    "num_tokens": 30533680.0,
    "mean_token_accuracy": 0.8369055807590484,
    "epoch": 0.898111923796564,
    "step": 2640
  },
  {
    "loss": 0.5639,
    "grad_norm": 0.37055504322052,
    "learning_rate": 2.5682401333661067e-06,
    "entropy": 0.5725692048668861,
    "num_tokens": 30643544.0,
    "mean_token_accuracy": 0.8459548681974411,
    "epoch": 0.9015138629018541,
    "step": 2650
  },
  {
    "loss": 0.5329,
    "grad_norm": 0.3881671130657196,
    "learning_rate": 2.396161189310603e-06,
    "entropy": 0.5357650727033615,
    "num_tokens": 30753201.0,
    "mean_token_accuracy": 0.8559219360351562,
    "epoch": 0.904915802007144,
    "step": 2660
  },
  {
    "loss": 0.5647,
    "grad_norm": 0.3779488205909729,
    "learning_rate": 2.2299073080660926e-06,
    "entropy": 0.5651383563876152,
    "num_tokens": 30866789.0,
    "mean_token_accuracy": 0.8465316832065582,
    "epoch": 0.9083177411124341,
    "step": 2670
  },
  {
    "loss": 0.5846,
    "grad_norm": 0.3553813397884369,
    "learning_rate": 2.069498833355371e-06,
    "entropy": 0.5878252282738685,
    "num_tokens": 30980359.0,
    "mean_token_accuracy": 0.846244278550148,
    "epoch": 0.9117196802177241,
    "step": 2680
  },
  {
    "loss": 0.5718,
    "grad_norm": 0.36092498898506165,
    "learning_rate": 1.914955393625717e-06,
    "entropy": 0.5731380596756935,
    "num_tokens": 31096112.0,
    "mean_token_accuracy": 0.8443791717290878,
    "epoch": 0.9151216193230142,
    "step": 2690
  },
  {
    "loss": 0.5533,
    "grad_norm": 0.39117369055747986,
    "learning_rate": 1.7662958996470635e-06,
    "entropy": 0.5589064419269562,
    "num_tokens": 31211039.0,
    "mean_token_accuracy": 0.8504908055067062,
    "epoch": 0.9185235584283041,
    "step": 2700
  },
  {
    "loss": 0.5676,
    "grad_norm": 0.37049877643585205,
    "learning_rate": 1.6235385421979555e-06,
    "entropy": 0.5745102554559708,
    "num_tokens": 31323209.0,
    "mean_token_accuracy": 0.8450854837894439,
    "epoch": 0.9219254975335941,
    "step": 2710
  },
  {
    "loss": 0.5449,
    "grad_norm": 0.36490964889526367,
    "learning_rate": 1.4867007898396457e-06,
    "entropy": 0.5510489344596863,
    "num_tokens": 31436438.0,
    "mean_token_accuracy": 0.8506697177886963,
    "epoch": 0.9253274366388842,
    "step": 2720
  },
  {
    "loss": 0.5585,
    "grad_norm": 0.3628520965576172,
    "learning_rate": 1.3557993867785335e-06,
    "entropy": 0.5683807358145714,
    "num_tokens": 31553624.0,
    "mean_token_accuracy": 0.8480296969413758,
    "epoch": 0.9287293757441741,
    "step": 2730
  },
  {
    "loss": 0.5956,
    "grad_norm": 0.34894460439682007,
    "learning_rate": 1.230850350817253e-06,
    "entropy": 0.6040398806333542,
    "num_tokens": 31669977.0,
    "mean_token_accuracy": 0.8400009006261826,
    "epoch": 0.9321313148494642,
    "step": 2740
  },
  {
    "loss": 0.5967,
    "grad_norm": 0.3939131200313568,
    "learning_rate": 1.1118689713946529e-06,
    "entropy": 0.5958102196455002,
    "num_tokens": 31790332.0,
    "mean_token_accuracy": 0.8398540407419205,
    "epoch": 0.9355332539547542,
    "step": 2750
  },
  {
    "loss": 0.5696,
    "grad_norm": 0.38736289739608765,
    "learning_rate": 9.98869807714914e-07,
    "entropy": 0.573664478957653,
    "num_tokens": 31897948.0,
    "mean_token_accuracy": 0.8483406931161881,
    "epoch": 0.9389351930600442,
    "step": 2760
  },
  {
    "loss": 0.6001,
    "grad_norm": 0.35055264830589294,
    "learning_rate": 8.918666869659642e-07,
    "entropy": 0.6001556605100632,
    "num_tokens": 32026504.0,
    "mean_token_accuracy": 0.8378526538610458,
    "epoch": 0.9423371321653342,
    "step": 2770
  },
  {
    "loss": 0.5759,
    "grad_norm": 0.39700019359588623,
    "learning_rate": 7.908727026275597e-07,
    "entropy": 0.5858756378293037,
    "num_tokens": 32140548.0,
    "mean_token_accuracy": 0.8406315892934799,
    "epoch": 0.9457390712706243,
    "step": 2780
  },
  {
    "loss": 0.5991,
    "grad_norm": 0.39454156160354614,
    "learning_rate": 6.959002128690606e-07,
    "entropy": 0.5967293009161949,
    "num_tokens": 32258820.0,
    "mean_token_accuracy": 0.8434277713298798,
    "epoch": 0.9491410103759143,
    "step": 2790
  },
  {
    "loss": 0.5897,
    "grad_norm": 0.42352262139320374,
    "learning_rate": 6.069608390372294e-07,
    "entropy": 0.5910201713442802,
    "num_tokens": 32370310.0,
    "mean_token_accuracy": 0.8437776923179626,
    "epoch": 0.9525429494812043,
    "step": 2800
  },
  {
    "loss": 0.5704,
    "grad_norm": 0.33332422375679016,
    "learning_rate": 5.240654642341913e-07,
    "entropy": 0.5700356841087342,
    "num_tokens": 32488521.0,
    "mean_token_accuracy": 0.8457501322031021,
    "epoch": 0.9559448885864943,
    "step": 2810
  },
  {
    "loss": 0.5863,
    "grad_norm": 0.3759448826313019,
    "learning_rate": 4.4722423198568785e-07,
    "entropy": 0.5873485445976258,
    "num_tokens": 32604586.0,
    "mean_token_accuracy": 0.8435898423194885,
    "epoch": 0.9593468276917844,
    "step": 2820
  },
  {
    "loss": 0.5712,
    "grad_norm": 0.3626145124435425,
    "learning_rate": 3.764465449999033e-07,
    "entropy": 0.5771880641579628,
    "num_tokens": 32720534.0,
    "mean_token_accuracy": 0.8457781076431274,
    "epoch": 0.9627487667970743,
    "step": 2830
  },
  {
    "loss": 0.5665,
    "grad_norm": 0.3542146384716034,
    "learning_rate": 3.1174106401686274e-07,
    "entropy": 0.5678144678473472,
    "num_tokens": 32842877.0,
    "mean_token_accuracy": 0.8448899626731873,
    "epoch": 0.9661507059023643,
    "step": 2840
  },
  {
    "loss": 0.5471,
    "grad_norm": 0.33239731192588806,
    "learning_rate": 2.5311570674867444e-07,
    "entropy": 0.5530853852629661,
    "num_tokens": 32951936.0,
    "mean_token_accuracy": 0.850860059261322,
    "epoch": 0.9695526450076544,
    "step": 2850
  },
  {
    "loss": 0.5992,
    "grad_norm": 0.3345488905906677,
    "learning_rate": 2.0057764691067127e-07,
    "entropy": 0.5983069583773613,
    "num_tokens": 33071421.0,
    "mean_token_accuracy": 0.839891204237938,
    "epoch": 0.9729545841129443,
    "step": 2860
  },
  {
    "loss": 0.5621,
    "grad_norm": 0.3342396020889282,
    "learning_rate": 1.5413331334360182e-07,
    "entropy": 0.5691981092095375,
    "num_tokens": 33189424.0,
    "mean_token_accuracy": 0.8471813201904297,
    "epoch": 0.9763565232182344,
    "step": 2870
  },
  {
    "loss": 0.6094,
    "grad_norm": 0.3459620475769043,
    "learning_rate": 1.1378838922694868e-07,
    "entropy": 0.6119711980223655,
    "num_tokens": 33306193.0,
    "mean_token_accuracy": 0.8373264789581298,
    "epoch": 0.9797584623235244,
    "step": 2880
  },
  {
    "loss": 0.5924,
    "grad_norm": 0.39813292026519775,
    "learning_rate": 7.954781138351241e-08,
    "entropy": 0.5946994125843048,
    "num_tokens": 33423283.0,
    "mean_token_accuracy": 0.8413377672433853,
    "epoch": 0.9831604014288144,
    "step": 2890
  },
  {
    "loss": 0.559,
    "grad_norm": 0.38787519931793213,
    "learning_rate": 5.141576967533368e-08,
    "entropy": 0.5609085857868195,
    "num_tokens": 33538638.0,
    "mean_token_accuracy": 0.8481513082981109,
    "epoch": 0.9865623405341044,
    "step": 2900
  },
  {
    "loss": 0.5985,
    "grad_norm": 0.4220837354660034,
    "learning_rate": 2.9395706490958907e-08,
    "entropy": 0.5978995203971863,
    "num_tokens": 33653798.0,
    "mean_token_accuracy": 0.8422107011079788,
    "epoch": 0.9899642796393945,
    "step": 2910
  },
  {
    "loss": 0.5694,
    "grad_norm": 0.3267785906791687,
    "learning_rate": 1.3490316324249463e-08,
    "entropy": 0.5705890819430351,
    "num_tokens": 33766910.0,
    "mean_token_accuracy": 0.8483996599912643,
    "epoch": 0.9933662187446844,
    "step": 2920
  },
  {
    "loss": 0.5656,
    "grad_norm": 0.38315078616142273,
    "learning_rate": 3.7015454446398536e-09,
    "entropy": 0.563205997645855,
    "num_tokens": 33880183.0,
    "mean_token_accuracy": 0.8477971255779266,
    "epoch": 0.9967681578499745,
    "step": 2930
  },
  {
    "loss": 0.565,
    "grad_norm": 0.5500030517578125,
    "learning_rate": 3.059165900598515e-11,
    "entropy": 0.5684452543133184,
    "num_tokens": 33988110.0,
    "mean_token_accuracy": 0.8478908413334897,
    "epoch": 1.0,
    "step": 2940
  },
  {
    "train_runtime": 22288.508,
    "train_samples_per_second": 4.22,
    "train_steps_per_second": 0.132,
    "total_flos": 2.46979541812514e+19,
    "train_loss": 0.6280506309197873,
    "epoch": 1.0,
    "step": 2940
  }
]