{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9995732696082615,
  "eval_steps": 100,
  "global_step": 732,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0013655372535631989,
      "grad_norm": 12.694538866354412,
      "learning_rate": 2.702702702702703e-07,
      "loss": 1.349,
      "mean_token_accuracy": 0.6731206055166955,
      "step": 1
    },
    {
      "epoch": 0.0027310745071263977,
      "grad_norm": 14.28682858099926,
      "learning_rate": 5.405405405405406e-07,
      "loss": 1.3695,
      "mean_token_accuracy": 0.6685968866221867,
      "step": 2
    },
    {
      "epoch": 0.004096611760689596,
      "grad_norm": 12.153556394325475,
      "learning_rate": 8.108108108108109e-07,
      "loss": 1.2534,
      "mean_token_accuracy": 0.6932581184973134,
      "step": 3
    },
    {
      "epoch": 0.0054621490142527955,
      "grad_norm": 10.014906326374183,
      "learning_rate": 1.0810810810810812e-06,
      "loss": 1.1814,
      "mean_token_accuracy": 0.7059116849514391,
      "step": 4
    },
    {
      "epoch": 0.006827686267815994,
      "grad_norm": 13.154435035689092,
      "learning_rate": 1.3513513513513515e-06,
      "loss": 1.3152,
      "mean_token_accuracy": 0.6801386295402803,
      "step": 5
    },
    {
      "epoch": 0.008193223521379193,
      "grad_norm": 12.656442197272979,
      "learning_rate": 1.6216216216216219e-06,
      "loss": 1.3519,
      "mean_token_accuracy": 0.6725742496834655,
      "step": 6
    },
    {
      "epoch": 0.009558760774942391,
      "grad_norm": 9.154457220968185,
      "learning_rate": 1.8918918918918922e-06,
      "loss": 1.3201,
      "mean_token_accuracy": 0.6776771426562855,
      "step": 7
    },
    {
      "epoch": 0.010924298028505591,
      "grad_norm": 12.926796361133281,
      "learning_rate": 2.1621621621621623e-06,
      "loss": 1.3668,
      "mean_token_accuracy": 0.6739682822033727,
      "step": 8
    },
    {
      "epoch": 0.01228983528206879,
      "grad_norm": 13.702550349182166,
      "learning_rate": 2.432432432432433e-06,
      "loss": 1.3398,
      "mean_token_accuracy": 0.675676714000192,
      "step": 9
    },
    {
      "epoch": 0.013655372535631987,
      "grad_norm": 12.100638571387348,
      "learning_rate": 2.702702702702703e-06,
      "loss": 1.237,
      "mean_token_accuracy": 0.6897704789682599,
      "step": 10
    },
    {
      "epoch": 0.015020909789195187,
      "grad_norm": 9.02772882068927,
      "learning_rate": 2.9729729729729736e-06,
      "loss": 1.237,
      "mean_token_accuracy": 0.6967972223171133,
      "step": 11
    },
    {
      "epoch": 0.016386447042758386,
      "grad_norm": 7.751826273673437,
      "learning_rate": 3.2432432432432437e-06,
      "loss": 1.2499,
      "mean_token_accuracy": 0.6873925007889158,
      "step": 12
    },
    {
      "epoch": 0.017751984296321584,
      "grad_norm": 14.605567306288991,
      "learning_rate": 3.513513513513514e-06,
      "loss": 1.1656,
      "mean_token_accuracy": 0.702033566686862,
      "step": 13
    },
    {
      "epoch": 0.019117521549884782,
      "grad_norm": 16.69580523349115,
      "learning_rate": 3.7837837837837844e-06,
      "loss": 1.1057,
      "mean_token_accuracy": 0.7166825564328566,
      "step": 14
    },
    {
      "epoch": 0.02048305880344798,
      "grad_norm": 4.277348708917889,
      "learning_rate": 4.0540540540540545e-06,
      "loss": 1.1453,
      "mean_token_accuracy": 0.7064730765093167,
      "step": 15
    },
    {
      "epoch": 0.021848596057011182,
      "grad_norm": 5.501675262398263,
      "learning_rate": 4.324324324324325e-06,
      "loss": 1.0609,
      "mean_token_accuracy": 0.7263837797527011,
      "step": 16
    },
    {
      "epoch": 0.02321413331057438,
      "grad_norm": 4.594618515621041,
      "learning_rate": 4.594594594594596e-06,
      "loss": 1.1575,
      "mean_token_accuracy": 0.6986264355169661,
      "step": 17
    },
    {
      "epoch": 0.02457967056413758,
      "grad_norm": 4.370254196578077,
      "learning_rate": 4.864864864864866e-06,
      "loss": 0.9763,
      "mean_token_accuracy": 0.7371238560229174,
      "step": 18
    },
    {
      "epoch": 0.025945207817700777,
      "grad_norm": 12.420426104801459,
      "learning_rate": 5.135135135135135e-06,
      "loss": 0.9947,
      "mean_token_accuracy": 0.7336301337158547,
      "step": 19
    },
    {
      "epoch": 0.027310745071263975,
      "grad_norm": 7.306932902857305,
      "learning_rate": 5.405405405405406e-06,
      "loss": 1.0089,
      "mean_token_accuracy": 0.7298717119207753,
      "step": 20
    },
    {
      "epoch": 0.028676282324827173,
      "grad_norm": 3.792544000354483,
      "learning_rate": 5.675675675675676e-06,
      "loss": 0.9992,
      "mean_token_accuracy": 0.7355598065427686,
      "step": 21
    },
    {
      "epoch": 0.030041819578390375,
      "grad_norm": 5.599819351248164,
      "learning_rate": 5.945945945945947e-06,
      "loss": 0.938,
      "mean_token_accuracy": 0.7446242908841856,
      "step": 22
    },
    {
      "epoch": 0.03140735683195357,
      "grad_norm": 3.2303878160644106,
      "learning_rate": 6.2162162162162164e-06,
      "loss": 0.842,
      "mean_token_accuracy": 0.7625513938625904,
      "step": 23
    },
    {
      "epoch": 0.03277289408551677,
      "grad_norm": 7.560111146519516,
      "learning_rate": 6.486486486486487e-06,
      "loss": 0.8541,
      "mean_token_accuracy": 0.7603489427240392,
      "step": 24
    },
    {
      "epoch": 0.03413843133907997,
      "grad_norm": 8.994000664594445,
      "learning_rate": 6.7567567567567575e-06,
      "loss": 0.8033,
      "mean_token_accuracy": 0.7710385708849239,
      "step": 25
    },
    {
      "epoch": 0.03550396859264317,
      "grad_norm": 11.174112139759627,
      "learning_rate": 7.027027027027028e-06,
      "loss": 0.7972,
      "mean_token_accuracy": 0.7709743153956785,
      "step": 26
    },
    {
      "epoch": 0.036869505846206366,
      "grad_norm": 4.4474620188546465,
      "learning_rate": 7.297297297297298e-06,
      "loss": 0.7779,
      "mean_token_accuracy": 0.7762167832548105,
      "step": 27
    },
    {
      "epoch": 0.038235043099769564,
      "grad_norm": 6.420798005839511,
      "learning_rate": 7.567567567567569e-06,
      "loss": 0.7364,
      "mean_token_accuracy": 0.7872430296163124,
      "step": 28
    },
    {
      "epoch": 0.03960058035333276,
      "grad_norm": 3.505076403243893,
      "learning_rate": 7.837837837837838e-06,
      "loss": 0.7366,
      "mean_token_accuracy": 0.7846892672832243,
      "step": 29
    },
    {
      "epoch": 0.04096611760689596,
      "grad_norm": 4.257419834855069,
      "learning_rate": 8.108108108108109e-06,
      "loss": 0.7276,
      "mean_token_accuracy": 0.7832414543493696,
      "step": 30
    },
    {
      "epoch": 0.04233165486045916,
      "grad_norm": 1.8596248429959361,
      "learning_rate": 8.378378378378378e-06,
      "loss": 0.6917,
      "mean_token_accuracy": 0.7949227280414939,
      "step": 31
    },
    {
      "epoch": 0.043697192114022364,
      "grad_norm": 5.998640085410845,
      "learning_rate": 8.64864864864865e-06,
      "loss": 0.74,
      "mean_token_accuracy": 0.7826185949965834,
      "step": 32
    },
    {
      "epoch": 0.04506272936758556,
      "grad_norm": 1.6136645601879687,
      "learning_rate": 8.91891891891892e-06,
      "loss": 0.682,
      "mean_token_accuracy": 0.7971229334178439,
      "step": 33
    },
    {
      "epoch": 0.04642826662114876,
      "grad_norm": 4.813218130166523,
      "learning_rate": 9.189189189189191e-06,
      "loss": 0.6818,
      "mean_token_accuracy": 0.7956781396448597,
      "step": 34
    },
    {
      "epoch": 0.04779380387471196,
      "grad_norm": 1.6648957199273453,
      "learning_rate": 9.45945945945946e-06,
      "loss": 0.6536,
      "mean_token_accuracy": 0.8039919273243677,
      "step": 35
    },
    {
      "epoch": 0.04915934112827516,
      "grad_norm": 2.727767929908723,
      "learning_rate": 9.729729729729732e-06,
      "loss": 0.5991,
      "mean_token_accuracy": 0.8157655755998052,
      "step": 36
    },
    {
      "epoch": 0.050524878381838355,
      "grad_norm": 3.307683813511232,
      "learning_rate": 1e-05,
      "loss": 0.6039,
      "mean_token_accuracy": 0.8131568470367118,
      "step": 37
    },
    {
      "epoch": 0.05189041563540155,
      "grad_norm": 1.6551722305811238,
      "learning_rate": 1.027027027027027e-05,
      "loss": 0.6005,
      "mean_token_accuracy": 0.8140848694651097,
      "step": 38
    },
    {
      "epoch": 0.05325595288896475,
      "grad_norm": 1.447375367061217,
      "learning_rate": 1.0540540540540541e-05,
      "loss": 0.6002,
      "mean_token_accuracy": 0.8150523180034766,
      "step": 39
    },
    {
      "epoch": 0.05462149014252795,
      "grad_norm": 1.072024342448014,
      "learning_rate": 1.0810810810810812e-05,
      "loss": 0.6008,
      "mean_token_accuracy": 0.8159713680250943,
      "step": 40
    },
    {
      "epoch": 0.05598702739609115,
      "grad_norm": 0.7603023981378255,
      "learning_rate": 1.1081081081081081e-05,
      "loss": 0.5885,
      "mean_token_accuracy": 0.8169645259893636,
      "step": 41
    },
    {
      "epoch": 0.057352564649654346,
      "grad_norm": 0.9581646192295539,
      "learning_rate": 1.1351351351351352e-05,
      "loss": 0.561,
      "mean_token_accuracy": 0.8244735432096343,
      "step": 42
    },
    {
      "epoch": 0.058718101903217544,
      "grad_norm": 1.2036888425863506,
      "learning_rate": 1.1621621621621622e-05,
      "loss": 0.5912,
      "mean_token_accuracy": 0.8190549345072387,
      "step": 43
    },
    {
      "epoch": 0.06008363915678075,
      "grad_norm": 0.7718687528534887,
      "learning_rate": 1.1891891891891894e-05,
      "loss": 0.5749,
      "mean_token_accuracy": 0.819959293524251,
      "step": 44
    },
    {
      "epoch": 0.06144917641034395,
      "grad_norm": 1.682073620529749,
      "learning_rate": 1.2162162162162164e-05,
      "loss": 0.5584,
      "mean_token_accuracy": 0.825639299275228,
      "step": 45
    },
    {
      "epoch": 0.06281471366390715,
      "grad_norm": 0.7001166412928403,
      "learning_rate": 1.2432432432432433e-05,
      "loss": 0.5581,
      "mean_token_accuracy": 0.8246738540448362,
      "step": 46
    },
    {
      "epoch": 0.06418025091747034,
      "grad_norm": 1.0537050693773666,
      "learning_rate": 1.2702702702702702e-05,
      "loss": 0.5551,
      "mean_token_accuracy": 0.8259298934704246,
      "step": 47
    },
    {
      "epoch": 0.06554578817103354,
      "grad_norm": 0.7920322967117208,
      "learning_rate": 1.2972972972972975e-05,
      "loss": 0.5507,
      "mean_token_accuracy": 0.8265237150428488,
      "step": 48
    },
    {
      "epoch": 0.06691132542459674,
      "grad_norm": 0.562181333813763,
      "learning_rate": 1.3243243243243244e-05,
      "loss": 0.5688,
      "mean_token_accuracy": 0.8204618404616847,
      "step": 49
    },
    {
      "epoch": 0.06827686267815994,
      "grad_norm": 0.5763859640174397,
      "learning_rate": 1.3513513513513515e-05,
      "loss": 0.5127,
      "mean_token_accuracy": 0.8361940201766735,
      "step": 50
    },
    {
      "epoch": 0.06964239993172314,
      "grad_norm": 0.8441858338744881,
      "learning_rate": 1.3783783783783784e-05,
      "loss": 0.5605,
      "mean_token_accuracy": 0.8224741152107983,
      "step": 51
    },
    {
      "epoch": 0.07100793718528634,
      "grad_norm": 0.623060068151014,
      "learning_rate": 1.4054054054054055e-05,
      "loss": 0.5154,
      "mean_token_accuracy": 0.8358351656376076,
      "step": 52
    },
    {
      "epoch": 0.07237347443884953,
      "grad_norm": 0.487364049768217,
      "learning_rate": 1.4324324324324326e-05,
      "loss": 0.5508,
      "mean_token_accuracy": 0.824997923142241,
      "step": 53
    },
    {
      "epoch": 0.07373901169241273,
      "grad_norm": 0.4710640541449941,
      "learning_rate": 1.4594594594594596e-05,
      "loss": 0.5121,
      "mean_token_accuracy": 0.8361596705359315,
      "step": 54
    },
    {
      "epoch": 0.07510454894597593,
      "grad_norm": 0.5307868639543387,
      "learning_rate": 1.4864864864864865e-05,
      "loss": 0.5388,
      "mean_token_accuracy": 0.8287927367224052,
      "step": 55
    },
    {
      "epoch": 0.07647008619953913,
      "grad_norm": 0.502790636865642,
      "learning_rate": 1.5135135135135138e-05,
      "loss": 0.5002,
      "mean_token_accuracy": 0.8393840192332627,
      "step": 56
    },
    {
      "epoch": 0.07783562345310233,
      "grad_norm": 0.43346690723829867,
      "learning_rate": 1.540540540540541e-05,
      "loss": 0.5125,
      "mean_token_accuracy": 0.8363432033977718,
      "step": 57
    },
    {
      "epoch": 0.07920116070666552,
      "grad_norm": 0.4993605510803232,
      "learning_rate": 1.5675675675675676e-05,
      "loss": 0.5188,
      "mean_token_accuracy": 0.8324141327632871,
      "step": 58
    },
    {
      "epoch": 0.08056669796022872,
      "grad_norm": 0.4277849657264807,
      "learning_rate": 1.5945945945945947e-05,
      "loss": 0.5099,
      "mean_token_accuracy": 0.8360334578638953,
      "step": 59
    },
    {
      "epoch": 0.08193223521379192,
      "grad_norm": 0.4591464651505517,
      "learning_rate": 1.6216216216216218e-05,
      "loss": 0.5046,
      "mean_token_accuracy": 0.8377397858533894,
      "step": 60
    },
    {
      "epoch": 0.08329777246735512,
      "grad_norm": 0.4510027457654597,
      "learning_rate": 1.648648648648649e-05,
      "loss": 0.511,
      "mean_token_accuracy": 0.8336544170905339,
      "step": 61
    },
    {
      "epoch": 0.08466330972091832,
      "grad_norm": 0.4050262343002075,
      "learning_rate": 1.6756756756756757e-05,
      "loss": 0.5051,
      "mean_token_accuracy": 0.8365473990294894,
      "step": 62
    },
    {
      "epoch": 0.08602884697448153,
      "grad_norm": 0.44615719951666044,
      "learning_rate": 1.7027027027027028e-05,
      "loss": 0.5265,
      "mean_token_accuracy": 0.8290703723078953,
      "step": 63
    },
    {
      "epoch": 0.08739438422804473,
      "grad_norm": 0.38700566099417927,
      "learning_rate": 1.72972972972973e-05,
      "loss": 0.4888,
      "mean_token_accuracy": 0.8415804018657317,
      "step": 64
    },
    {
      "epoch": 0.08875992148160793,
      "grad_norm": 0.45966773746839323,
      "learning_rate": 1.756756756756757e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8367421080881654,
      "step": 65
    },
    {
      "epoch": 0.09012545873517112,
      "grad_norm": 0.3462174672127141,
      "learning_rate": 1.783783783783784e-05,
      "loss": 0.5056,
      "mean_token_accuracy": 0.8362969110791204,
      "step": 66
    },
    {
      "epoch": 0.09149099598873432,
      "grad_norm": 0.5378368089497744,
      "learning_rate": 1.8108108108108108e-05,
      "loss": 0.4845,
      "mean_token_accuracy": 0.841530221609978,
      "step": 67
    },
    {
      "epoch": 0.09285653324229752,
      "grad_norm": 0.3927908242857095,
      "learning_rate": 1.8378378378378383e-05,
      "loss": 0.5017,
      "mean_token_accuracy": 0.8354575522429637,
      "step": 68
    },
    {
      "epoch": 0.09422207049586072,
      "grad_norm": 0.48195021375325325,
      "learning_rate": 1.864864864864865e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8440971672690992,
      "step": 69
    },
    {
      "epoch": 0.09558760774942392,
      "grad_norm": 0.42743054642794015,
      "learning_rate": 1.891891891891892e-05,
      "loss": 0.5054,
      "mean_token_accuracy": 0.8351946171032623,
      "step": 70
    },
    {
      "epoch": 0.09695314500298712,
      "grad_norm": 0.5073725399698784,
      "learning_rate": 1.918918918918919e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8416211405529895,
      "step": 71
    },
    {
      "epoch": 0.09831868225655031,
      "grad_norm": 0.32802476428146243,
      "learning_rate": 1.9459459459459463e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8410772826528182,
      "step": 72
    },
    {
      "epoch": 0.09968421951011351,
      "grad_norm": 0.44056399284971365,
      "learning_rate": 1.972972972972973e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8444900562879512,
      "step": 73
    },
    {
      "epoch": 0.10104975676367671,
      "grad_norm": 0.46074182462916863,
      "learning_rate": 2e-05,
      "loss": 0.5052,
      "mean_token_accuracy": 0.8356767677438455,
      "step": 74
    },
    {
      "epoch": 0.10241529401723991,
      "grad_norm": 0.43060352923187173,
      "learning_rate": 1.999988602302209e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8357036382011872,
      "step": 75
    },
    {
      "epoch": 0.1037808312708031,
      "grad_norm": 0.5392177840776831,
      "learning_rate": 1.9999544094686517e-05,
      "loss": 0.464,
      "mean_token_accuracy": 0.8469077059106851,
      "step": 76
    },
    {
      "epoch": 0.1051463685243663,
      "grad_norm": 0.4383769298059935,
      "learning_rate": 1.999897422278767e-05,
      "loss": 0.475,
      "mean_token_accuracy": 0.8436229064865175,
      "step": 77
    },
    {
      "epoch": 0.1065119057779295,
      "grad_norm": 0.4904266229445523,
      "learning_rate": 1.9998176420316002e-05,
      "loss": 0.4853,
      "mean_token_accuracy": 0.8408816529254115,
      "step": 78
    },
    {
      "epoch": 0.1078774430314927,
      "grad_norm": 0.36607390424605935,
      "learning_rate": 1.9997150705457738e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8487558375701372,
      "step": 79
    },
    {
      "epoch": 0.1092429802850559,
      "grad_norm": 0.4551168695964319,
      "learning_rate": 1.9995897101594454e-05,
      "loss": 0.4876,
      "mean_token_accuracy": 0.8396327622489336,
      "step": 80
    },
    {
      "epoch": 0.1106085175386191,
      "grad_norm": 0.42442283856925495,
      "learning_rate": 1.9994415637302545e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8483948181491869,
      "step": 81
    },
    {
      "epoch": 0.1119740547921823,
      "grad_norm": 0.6393625908948564,
      "learning_rate": 1.999270634635258e-05,
      "loss": 0.4606,
      "mean_token_accuracy": 0.8486694050596694,
      "step": 82
    },
    {
      "epoch": 0.1133395920457455,
      "grad_norm": 0.3938759824609981,
      "learning_rate": 1.9990769267708517e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.843921182173024,
      "step": 83
    },
    {
      "epoch": 0.11470512929930869,
      "grad_norm": 0.46323080288240137,
      "learning_rate": 1.998860444552683e-05,
      "loss": 0.4579,
      "mean_token_accuracy": 0.8487062802081777,
      "step": 84
    },
    {
      "epoch": 0.11607066655287189,
      "grad_norm": 0.4100887075069859,
      "learning_rate": 1.99862119291555e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8426625139694884,
      "step": 85
    },
    {
      "epoch": 0.11743620380643509,
      "grad_norm": 0.460266156523643,
      "learning_rate": 1.9983591773132885e-05,
      "loss": 0.4634,
      "mean_token_accuracy": 0.8479961635199413,
      "step": 86
    },
    {
      "epoch": 0.11880174105999829,
      "grad_norm": 0.4093999187610122,
      "learning_rate": 1.998074403718647e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8413903297828744,
      "step": 87
    },
    {
      "epoch": 0.1201672783135615,
      "grad_norm": 0.4529979351460269,
      "learning_rate": 1.9977668786231536e-05,
      "loss": 0.4646,
      "mean_token_accuracy": 0.8439730016702047,
      "step": 88
    },
    {
      "epoch": 0.1215328155671247,
      "grad_norm": 0.31940153601185445,
      "learning_rate": 1.997436609036963e-05,
      "loss": 0.4929,
      "mean_token_accuracy": 0.8389086531135729,
      "step": 89
    },
    {
      "epoch": 0.1228983528206879,
      "grad_norm": 0.36879956082602044,
      "learning_rate": 1.997083602488702e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8442029592930739,
      "step": 90
    },
    {
      "epoch": 0.1242638900742511,
      "grad_norm": 0.31625144322456566,
      "learning_rate": 1.9967078670252947e-05,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8414532739667411,
      "step": 91
    },
    {
      "epoch": 0.1256294273278143,
      "grad_norm": 0.37572795515302265,
      "learning_rate": 1.9963094112117786e-05,
      "loss": 0.4731,
      "mean_token_accuracy": 0.8458833864797164,
      "step": 92
    },
    {
      "epoch": 0.1269949645813775,
      "grad_norm": 0.32729110107644405,
      "learning_rate": 1.995888244131113e-05,
      "loss": 0.4729,
      "mean_token_accuracy": 0.844289710669155,
      "step": 93
    },
    {
      "epoch": 0.1283605018349407,
      "grad_norm": 0.4508003392275939,
      "learning_rate": 1.9954443753839666e-05,
      "loss": 0.4844,
      "mean_token_accuracy": 0.8410096036799376,
      "step": 94
    },
    {
      "epoch": 0.12972603908850389,
      "grad_norm": 0.34367729445854445,
      "learning_rate": 1.994977815088504e-05,
      "loss": 0.4536,
      "mean_token_accuracy": 0.8496317958214457,
      "step": 95
    },
    {
      "epoch": 0.13109157634206708,
      "grad_norm": 0.39368016078712464,
      "learning_rate": 1.994488573880152e-05,
      "loss": 0.4427,
      "mean_token_accuracy": 0.8527215734701595,
      "step": 96
    },
    {
      "epoch": 0.13245711359563028,
      "grad_norm": 0.39400450065152637,
      "learning_rate": 1.9939766629113568e-05,
      "loss": 0.4718,
      "mean_token_accuracy": 0.8439487535684244,
      "step": 97
    },
    {
      "epoch": 0.13382265084919348,
      "grad_norm": 0.3644413048898344,
      "learning_rate": 1.9934420938513313e-05,
      "loss": 0.4451,
      "mean_token_accuracy": 0.8515555292296904,
      "step": 98
    },
    {
      "epoch": 0.13518818810275668,
      "grad_norm": 0.33820519416057315,
      "learning_rate": 1.9928848788857887e-05,
      "loss": 0.4687,
      "mean_token_accuracy": 0.8435826255747995,
      "step": 99
    },
    {
      "epoch": 0.13655372535631988,
      "grad_norm": 0.35474640844985367,
      "learning_rate": 1.9923050307166655e-05,
      "loss": 0.4694,
      "mean_token_accuracy": 0.8443660349778408,
      "step": 100
    },
    {
      "epoch": 0.13791926260988308,
      "grad_norm": 0.3787756690185754,
      "learning_rate": 1.9917025625618295e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8515359704739787,
      "step": 101
    },
    {
      "epoch": 0.13928479986344627,
      "grad_norm": 0.3362568869450033,
      "learning_rate": 1.9910774881547803e-05,
      "loss": 0.4697,
      "mean_token_accuracy": 0.8443949120016512,
      "step": 102
    },
    {
      "epoch": 0.14065033711700947,
      "grad_norm": 0.33478608195399623,
      "learning_rate": 1.9904298217443366e-05,
      "loss": 0.4655,
      "mean_token_accuracy": 0.8454621512591025,
      "step": 103
    },
    {
      "epoch": 0.14201587437057267,
      "grad_norm": 0.3157906568874303,
      "learning_rate": 1.9897595780943104e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8479587393969411,
      "step": 104
    },
    {
      "epoch": 0.14338141162413587,
      "grad_norm": 0.33967710856761196,
      "learning_rate": 1.989066772483171e-05,
      "loss": 0.4584,
      "mean_token_accuracy": 0.8502480645030208,
      "step": 105
    },
    {
      "epoch": 0.14474694887769907,
      "grad_norm": 0.3625458211243482,
      "learning_rate": 1.988351420703696e-05,
      "loss": 0.4449,
      "mean_token_accuracy": 0.8503154268653385,
      "step": 106
    },
    {
      "epoch": 0.14611248613126226,
      "grad_norm": 0.33343733476340787,
      "learning_rate": 1.9876135390626123e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.8469237180257693,
      "step": 107
    },
    {
      "epoch": 0.14747802338482546,
      "grad_norm": 0.3402896753249496,
      "learning_rate": 1.986853144380224e-05,
      "loss": 0.4737,
      "mean_token_accuracy": 0.8415288812333515,
      "step": 108
    },
    {
      "epoch": 0.14884356063838866,
      "grad_norm": 0.30321741460606055,
      "learning_rate": 1.9860702539900288e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8479057193176431,
      "step": 109
    },
    {
      "epoch": 0.15020909789195186,
      "grad_norm": 0.34336551349752786,
      "learning_rate": 1.9852648857383224e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.8476062469011421,
      "step": 110
    },
    {
      "epoch": 0.15157463514551506,
      "grad_norm": 0.3096450131783167,
      "learning_rate": 1.984437057983793e-05,
      "loss": 0.4388,
      "mean_token_accuracy": 0.8534628120170084,
      "step": 111
    },
    {
      "epoch": 0.15294017239907826,
      "grad_norm": 0.36874309540738714,
      "learning_rate": 1.9835867895971015e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8526718507556617,
      "step": 112
    },
    {
      "epoch": 0.15430570965264145,
      "grad_norm": 0.3605796560605681,
      "learning_rate": 1.982714099960452e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.846226495136954,
      "step": 113
    },
    {
      "epoch": 0.15567124690620465,
      "grad_norm": 0.4627441704849731,
      "learning_rate": 1.981819008967151e-05,
      "loss": 0.4858,
      "mean_token_accuracy": 0.8394183495047153,
      "step": 114
    },
    {
      "epoch": 0.15703678415976785,
      "grad_norm": 0.34714642141003565,
      "learning_rate": 1.9809015370211505e-05,
      "loss": 0.4667,
      "mean_token_accuracy": 0.8454976903304268,
      "step": 115
    },
    {
      "epoch": 0.15840232141333105,
      "grad_norm": 0.34086378030044884,
      "learning_rate": 1.979961705036587e-05,
      "loss": 0.4483,
      "mean_token_accuracy": 0.8508435940254968,
      "step": 116
    },
    {
      "epoch": 0.15976785866689425,
      "grad_norm": 0.4140107037865621,
      "learning_rate": 1.9789995344373027e-05,
      "loss": 0.4449,
      "mean_token_accuracy": 0.8518490808801514,
      "step": 117
    },
    {
      "epoch": 0.16113339592045745,
      "grad_norm": 0.34331795050708586,
      "learning_rate": 1.978015047156356e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8451241258921095,
      "step": 118
    },
    {
      "epoch": 0.16249893317402064,
      "grad_norm": 0.3380747151761631,
      "learning_rate": 1.977008265635525e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8551851359088176,
      "step": 119
    },
    {
      "epoch": 0.16386447042758384,
      "grad_norm": 0.39246892074574696,
      "learning_rate": 1.9759792128247922e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8475478494544696,
      "step": 120
    },
    {
      "epoch": 0.16523000768114704,
      "grad_norm": 0.3077478512701279,
      "learning_rate": 1.9749279121818235e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8506487627564587,
      "step": 121
    },
    {
      "epoch": 0.16659554493471024,
      "grad_norm": 0.3128694161488127,
      "learning_rate": 1.9738543876714335e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8477825790742587,
      "step": 122
    },
    {
      "epoch": 0.16796108218827344,
      "grad_norm": 0.33718369881101784,
      "learning_rate": 1.9727586637650373e-05,
      "loss": 0.4381,
      "mean_token_accuracy": 0.8531846568153524,
      "step": 123
    },
    {
      "epoch": 0.16932661944183663,
      "grad_norm": 0.31111696829985575,
      "learning_rate": 1.9716407654400954e-05,
      "loss": 0.4715,
      "mean_token_accuracy": 0.8433811011221168,
      "step": 124
    },
    {
      "epoch": 0.17069215669539983,
      "grad_norm": 0.3250655447526398,
      "learning_rate": 1.9705007181795416e-05,
      "loss": 0.4594,
      "mean_token_accuracy": 0.847703186503929,
      "step": 125
    },
    {
      "epoch": 0.17205769394896306,
      "grad_norm": 0.31423640348811155,
      "learning_rate": 1.9693385479712047e-05,
      "loss": 0.442,
      "mean_token_accuracy": 0.8514727417072737,
      "step": 126
    },
    {
      "epoch": 0.17342323120252626,
      "grad_norm": 0.35736412097514153,
      "learning_rate": 1.9681542813072147e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8411854317148191,
      "step": 127
    },
    {
      "epoch": 0.17478876845608946,
      "grad_norm": 0.2902859082859034,
      "learning_rate": 1.9669479451833976e-05,
      "loss": 0.4388,
      "mean_token_accuracy": 0.8539477536048056,
      "step": 128
    },
    {
      "epoch": 0.17615430570965265,
      "grad_norm": 0.31967638765399986,
      "learning_rate": 1.9657195670986638e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8475490129873101,
      "step": 129
    },
    {
      "epoch": 0.17751984296321585,
      "grad_norm": 0.3019630526325269,
      "learning_rate": 1.964469175054377e-05,
      "loss": 0.4508,
      "mean_token_accuracy": 0.8500409132192751,
      "step": 130
    },
    {
      "epoch": 0.17888538021677905,
      "grad_norm": 0.28471378493835015,
      "learning_rate": 1.963196797553718e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.8465730185761348,
      "step": 131
    },
    {
      "epoch": 0.18025091747034225,
      "grad_norm": 0.2976461319068471,
      "learning_rate": 1.961902463601036e-05,
      "loss": 0.4483,
      "mean_token_accuracy": 0.8504603311676011,
      "step": 132
    },
    {
      "epoch": 0.18161645472390545,
      "grad_norm": 0.3420853043630677,
      "learning_rate": 1.9605862027011858e-05,
      "loss": 0.446,
      "mean_token_accuracy": 0.8522397868056077,
      "step": 133
    },
    {
      "epoch": 0.18298199197746864,
      "grad_norm": 0.30068683893569925,
      "learning_rate": 1.959248044858854e-05,
      "loss": 0.4538,
      "mean_token_accuracy": 0.8479476016652441,
      "step": 134
    },
    {
      "epoch": 0.18434752923103184,
      "grad_norm": 0.337968764863526,
      "learning_rate": 1.9578880205778793e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8544472005484545,
      "step": 135
    },
    {
      "epoch": 0.18571306648459504,
      "grad_norm": 0.3187881918567848,
      "learning_rate": 1.9565061608605526e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8479640485502516,
      "step": 136
    },
    {
      "epoch": 0.18707860373815824,
      "grad_norm": 0.35269232923831895,
      "learning_rate": 1.9551024972069127e-05,
      "loss": 0.4509,
      "mean_token_accuracy": 0.8508397906934503,
      "step": 137
    },
    {
      "epoch": 0.18844414099172144,
      "grad_norm": 0.3058471460955213,
      "learning_rate": 1.9536770616140277e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.8494591928231272,
      "step": 138
    },
    {
      "epoch": 0.18980967824528464,
      "grad_norm": 0.2923803706742491,
      "learning_rate": 1.9522298865752662e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8497101586020829,
      "step": 139
    },
    {
      "epoch": 0.19117521549884783,
      "grad_norm": 0.35715418556836526,
      "learning_rate": 1.950761005079556e-05,
      "loss": 0.4779,
      "mean_token_accuracy": 0.8411692267142432,
      "step": 140
    },
    {
      "epoch": 0.19254075275241103,
      "grad_norm": 0.27357483270279737,
      "learning_rate": 1.949270450610631e-05,
      "loss": 0.4116,
      "mean_token_accuracy": 0.8609656246183783,
      "step": 141
    },
    {
      "epoch": 0.19390629000597423,
      "grad_norm": 0.3713092433646525,
      "learning_rate": 1.9477582571462706e-05,
      "loss": 0.4513,
      "mean_token_accuracy": 0.8494811710899883,
      "step": 142
    },
    {
      "epoch": 0.19527182725953743,
      "grad_norm": 0.34866780781386153,
      "learning_rate": 1.9462244591575222e-05,
      "loss": 0.4447,
      "mean_token_accuracy": 0.8526829842758383,
      "step": 143
    },
    {
      "epoch": 0.19663736451310063,
      "grad_norm": 0.2949930109269014,
      "learning_rate": 1.944669091607919e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8562389055851354,
      "step": 144
    },
    {
      "epoch": 0.19800290176666382,
      "grad_norm": 0.374752731020472,
      "learning_rate": 1.9430921899526786e-05,
      "loss": 0.4478,
      "mean_token_accuracy": 0.8499277623582386,
      "step": 145
    },
    {
      "epoch": 0.19936843902022702,
      "grad_norm": 0.3651538698048502,
      "learning_rate": 1.941493790137898e-05,
      "loss": 0.4442,
      "mean_token_accuracy": 0.8526649985454055,
      "step": 146
    },
    {
      "epoch": 0.20073397627379022,
      "grad_norm": 0.36636578170953166,
      "learning_rate": 1.9398739285997342e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8504179631544849,
      "step": 147
    },
    {
      "epoch": 0.20209951352735342,
      "grad_norm": 0.40761326956885474,
      "learning_rate": 1.9382326422635705e-05,
      "loss": 0.4642,
      "mean_token_accuracy": 0.8452167499746224,
      "step": 148
    },
    {
      "epoch": 0.20346505078091662,
      "grad_norm": 0.31422691233021155,
      "learning_rate": 1.936569968543179e-05,
      "loss": 0.4368,
      "mean_token_accuracy": 0.8547842065414144,
      "step": 149
    },
    {
      "epoch": 0.20483058803447982,
      "grad_norm": 0.35446348871617744,
      "learning_rate": 1.934885945339865e-05,
      "loss": 0.4402,
      "mean_token_accuracy": 0.8503691031564117,
      "step": 150
    },
    {
      "epoch": 0.20619612528804301,
      "grad_norm": 0.32285449024202295,
      "learning_rate": 1.9331806110416027e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8497348879856765,
      "step": 151
    },
    {
      "epoch": 0.2075616625416062,
      "grad_norm": 0.2974246522701837,
      "learning_rate": 1.9314540045221628e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8566265295867389,
      "step": 152
    },
    {
      "epoch": 0.2089271997951694,
      "grad_norm": 0.34647196924654977,
      "learning_rate": 1.9297061651402237e-05,
      "loss": 0.4326,
      "mean_token_accuracy": 0.8539450355140382,
      "step": 153
    },
    {
      "epoch": 0.2102927370487326,
      "grad_norm": 0.386258719161046,
      "learning_rate": 1.927937132738476e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.8532191960011701,
      "step": 154
    },
    {
      "epoch": 0.2116582743022958,
      "grad_norm": 0.34454258298504326,
      "learning_rate": 1.9261469476427122e-05,
      "loss": 0.4258,
      "mean_token_accuracy": 0.8568923616497963,
      "step": 155
    },
    {
      "epoch": 0.213023811555859,
      "grad_norm": 0.3541361157616604,
      "learning_rate": 1.92433565066091e-05,
      "loss": 0.4529,
      "mean_token_accuracy": 0.8479736960845543,
      "step": 156
    },
    {
      "epoch": 0.2143893488094222,
      "grad_norm": 0.35191182105790164,
      "learning_rate": 1.922503283082301e-05,
      "loss": 0.4202,
      "mean_token_accuracy": 0.8578095481186783,
      "step": 157
    },
    {
      "epoch": 0.2157548860629854,
      "grad_norm": 0.42204829019945844,
      "learning_rate": 1.920649886676429e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8431488494361127,
      "step": 158
    },
    {
      "epoch": 0.2171204233165486,
      "grad_norm": 0.282563065166861,
      "learning_rate": 1.9187755036921976e-05,
      "loss": 0.4481,
      "mean_token_accuracy": 0.8492420629486095,
      "step": 159
    },
    {
      "epoch": 0.2184859605701118,
      "grad_norm": 0.3294030087715004,
      "learning_rate": 1.916880176856909e-05,
      "loss": 0.4239,
      "mean_token_accuracy": 0.856072747237234,
      "step": 160
    },
    {
      "epoch": 0.219851497823675,
      "grad_norm": 0.31721191426888634,
      "learning_rate": 1.914963949375288e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8534969611431503,
      "step": 161
    },
    {
      "epoch": 0.2212170350772382,
      "grad_norm": 0.34061976253510645,
      "learning_rate": 1.9130268649284982e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8489997065725852,
      "step": 162
    },
    {
      "epoch": 0.2225825723308014,
      "grad_norm": 0.319599698534074,
      "learning_rate": 1.9110689676731454e-05,
      "loss": 0.4524,
      "mean_token_accuracy": 0.8484242719179398,
      "step": 163
    },
    {
      "epoch": 0.2239481095843646,
      "grad_norm": 0.38523458399821875,
      "learning_rate": 1.909090302240273e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8499845577594263,
      "step": 164
    },
    {
      "epoch": 0.2253136468379278,
      "grad_norm": 0.3082646414065079,
      "learning_rate": 1.907090913734341e-05,
      "loss": 0.4228,
      "mean_token_accuracy": 0.8565411948866354,
      "step": 165
    },
    {
      "epoch": 0.226679184091491,
      "grad_norm": 0.4281504609323445,
      "learning_rate": 1.905070847732202e-05,
      "loss": 0.4231,
      "mean_token_accuracy": 0.8567314407107315,
      "step": 166
    },
    {
      "epoch": 0.22804472134505419,
      "grad_norm": 0.3122301354550845,
      "learning_rate": 1.9030301502820597e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8526059868064779,
      "step": 167
    },
    {
      "epoch": 0.22941025859861738,
      "grad_norm": 0.3600671344222034,
      "learning_rate": 1.900968867902419e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8509005211327412,
      "step": 168
    },
    {
      "epoch": 0.23077579585218058,
      "grad_norm": 0.3335012451792573,
      "learning_rate": 1.8988870475810284e-05,
      "loss": 0.4323,
      "mean_token_accuracy": 0.8542063476327129,
      "step": 169
    },
    {
      "epoch": 0.23214133310574378,
      "grad_norm": 0.3416637272262734,
      "learning_rate": 1.896784736773805e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8489287380113998,
      "step": 170
    },
    {
      "epoch": 0.23350687035930698,
      "grad_norm": 0.4300545403831784,
      "learning_rate": 1.894661983403755e-05,
      "loss": 0.441,
      "mean_token_accuracy": 0.8535810843521906,
      "step": 171
    },
    {
      "epoch": 0.23487240761287018,
      "grad_norm": 0.3258898174006088,
      "learning_rate": 1.8925188358598815e-05,
      "loss": 0.4376,
      "mean_token_accuracy": 0.8517990481723685,
      "step": 172
    },
    {
      "epoch": 0.23623794486643337,
      "grad_norm": 0.2988219576519335,
      "learning_rate": 1.8903553429960803e-05,
      "loss": 0.4694,
      "mean_token_accuracy": 0.8419916214618768,
      "step": 173
    },
    {
      "epoch": 0.23760348211999657,
      "grad_norm": 0.32171167299283115,
      "learning_rate": 1.8881715541300278e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8526490817200958,
      "step": 174
    },
    {
      "epoch": 0.23896901937355977,
      "grad_norm": 0.36549760624914024,
      "learning_rate": 1.885967519042054e-05,
      "loss": 0.4575,
      "mean_token_accuracy": 0.8467209578442905,
      "step": 175
    },
    {
      "epoch": 0.240334556627123,
      "grad_norm": 0.3206347251456072,
      "learning_rate": 1.8837432879740113e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8562328690343533,
      "step": 176
    },
    {
      "epoch": 0.2417000938806862,
      "grad_norm": 0.36277383760815624,
      "learning_rate": 1.881498911628127e-05,
      "loss": 0.4317,
      "mean_token_accuracy": 0.8544379338426664,
      "step": 177
    },
    {
      "epoch": 0.2430656311342494,
      "grad_norm": 0.3634699354146494,
      "learning_rate": 1.879234441165847e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.8562673113043472,
      "step": 178
    },
    {
      "epoch": 0.2444311683878126,
      "grad_norm": 0.3364593897580825,
      "learning_rate": 1.8769499282066716e-05,
      "loss": 0.4288,
      "mean_token_accuracy": 0.8545978179267162,
      "step": 179
    },
    {
      "epoch": 0.2457967056413758,
      "grad_norm": 0.36736450922616337,
      "learning_rate": 1.8746454248269777e-05,
      "loss": 0.4547,
      "mean_token_accuracy": 0.8489931015745901,
      "step": 180
    },
    {
      "epoch": 0.247162242894939,
      "grad_norm": 0.2842529030222029,
      "learning_rate": 1.872320983558831e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.8537670434938108,
      "step": 181
    },
    {
      "epoch": 0.2485277801485022,
      "grad_norm": 0.3503501258863951,
      "learning_rate": 1.8699766573887902e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8555725755068402,
      "step": 182
    },
    {
      "epoch": 0.24989331740206538,
      "grad_norm": 0.32115728482149525,
      "learning_rate": 1.867612499756697e-05,
      "loss": 0.4349,
      "mean_token_accuracy": 0.8538767592466098,
      "step": 183
    },
    {
      "epoch": 0.2512588546556286,
      "grad_norm": 0.37284059913436857,
      "learning_rate": 1.8652285645544602e-05,
      "loss": 0.4229,
      "mean_token_accuracy": 0.8565206996350189,
      "step": 184
    },
    {
      "epoch": 0.2526243919091918,
      "grad_norm": 0.27914377362605364,
      "learning_rate": 1.862824906124826e-05,
      "loss": 0.4389,
      "mean_token_accuracy": 0.8533956082213735,
      "step": 185
    },
    {
      "epoch": 0.253989929162755,
      "grad_norm": 0.33047332583495903,
      "learning_rate": 1.8604015792601395e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.8587885669486044,
      "step": 186
    },
    {
      "epoch": 0.2553554664163182,
      "grad_norm": 0.3098565583156071,
      "learning_rate": 1.8579586392010943e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.841463056808317,
      "step": 187
    },
    {
      "epoch": 0.2567210036698814,
      "grad_norm": 0.32188400036801423,
      "learning_rate": 1.855496141635476e-05,
      "loss": 0.4488,
      "mean_token_accuracy": 0.8494999156308464,
      "step": 188
    },
    {
      "epoch": 0.2580865409234446,
      "grad_norm": 0.3130726973042191,
      "learning_rate": 1.8530141426968905e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8566668881226269,
      "step": 189
    },
    {
      "epoch": 0.25945207817700777,
      "grad_norm": 0.3957651044446383,
      "learning_rate": 1.850512698963485e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.8556154103093908,
      "step": 190
    },
    {
      "epoch": 0.26081761543057097,
      "grad_norm": 0.3031069025669304,
      "learning_rate": 1.8479918674566602e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8541440562465102,
      "step": 191
    },
    {
      "epoch": 0.26218315268413417,
      "grad_norm": 0.3643754085496684,
      "learning_rate": 1.8454517056397663e-05,
      "loss": 0.4324,
      "mean_token_accuracy": 0.855023699771463,
      "step": 192
    },
    {
      "epoch": 0.26354868993769737,
      "grad_norm": 0.3003852707955295,
      "learning_rate": 1.842892271416797e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.8553599004543545,
      "step": 193
    },
    {
      "epoch": 0.26491422719126057,
      "grad_norm": 0.34484824975763084,
      "learning_rate": 1.8403136231310686e-05,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8566094556118384,
      "step": 194
    },
    {
      "epoch": 0.26627976444482376,
      "grad_norm": 0.3248230314864982,
      "learning_rate": 1.8377158195638877e-05,
      "loss": 0.4411,
      "mean_token_accuracy": 0.850191582672607,
      "step": 195
    },
    {
      "epoch": 0.26764530169838696,
      "grad_norm": 0.3425796212560859,
      "learning_rate": 1.8350989199332156e-05,
      "loss": 0.4467,
      "mean_token_accuracy": 0.8491048527905652,
      "step": 196
    },
    {
      "epoch": 0.26901083895195016,
      "grad_norm": 0.34768568039417325,
      "learning_rate": 1.8324629838923132e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.8511677206305359,
      "step": 197
    },
    {
      "epoch": 0.27037637620551336,
      "grad_norm": 0.34080656327066544,
      "learning_rate": 1.829808071528386e-05,
      "loss": 0.4281,
      "mean_token_accuracy": 0.8569207458025948,
      "step": 198
    },
    {
      "epoch": 0.27174191345907656,
      "grad_norm": 0.3367911094729273,
      "learning_rate": 1.8271342433612114e-05,
      "loss": 0.4395,
      "mean_token_accuracy": 0.8523639031650138,
      "step": 199
    },
    {
      "epoch": 0.27310745071263975,
      "grad_norm": 0.3151512003460036,
      "learning_rate": 1.8244415603417603e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8558035079566513,
      "step": 200
    },
    {
      "epoch": 0.27447298796620295,
      "grad_norm": 0.3220180875156976,
      "learning_rate": 1.8217300838508075e-05,
      "loss": 0.403,
      "mean_token_accuracy": 0.8631310614267218,
      "step": 201
    },
    {
      "epoch": 0.27583852521976615,
      "grad_norm": 0.28929076526601605,
      "learning_rate": 1.8189998756975318e-05,
      "loss": 0.4282,
      "mean_token_accuracy": 0.8555874931968974,
      "step": 202
    },
    {
      "epoch": 0.27720406247332935,
      "grad_norm": 0.36330635065555233,
      "learning_rate": 1.8162509981181084e-05,
      "loss": 0.4479,
      "mean_token_accuracy": 0.8496229750925639,
      "step": 203
    },
    {
      "epoch": 0.27856959972689255,
      "grad_norm": 0.27638019683036386,
      "learning_rate": 1.813483513774289e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8529176942517751,
      "step": 204
    },
    {
      "epoch": 0.27993513698045575,
      "grad_norm": 0.35506296866931764,
      "learning_rate": 1.8106974857519737e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8514268242896852,
      "step": 205
    },
    {
      "epoch": 0.28130067423401894,
      "grad_norm": 0.3286137389758072,
      "learning_rate": 1.807892977559774e-05,
      "loss": 0.4417,
      "mean_token_accuracy": 0.8510960913771437,
      "step": 206
    },
    {
      "epoch": 0.28266621148758214,
      "grad_norm": 0.25518635308883814,
      "learning_rate": 1.8050700531275632e-05,
      "loss": 0.418,
      "mean_token_accuracy": 0.8594128726528723,
      "step": 207
    },
    {
      "epoch": 0.28403174874114534,
      "grad_norm": 0.34031577918112327,
      "learning_rate": 1.80222877680502e-05,
      "loss": 0.421,
      "mean_token_accuracy": 0.8569013665570591,
      "step": 208
    },
    {
      "epoch": 0.28539728599470854,
      "grad_norm": 0.39038645873424477,
      "learning_rate": 1.799369213360163e-05,
      "loss": 0.4389,
      "mean_token_accuracy": 0.851740201076106,
      "step": 209
    },
    {
      "epoch": 0.28676282324827174,
      "grad_norm": 0.34488640211714505,
      "learning_rate": 1.7964914279778716e-05,
      "loss": 0.4449,
      "mean_token_accuracy": 0.8504230733766642,
      "step": 210
    },
    {
      "epoch": 0.28812836050183493,
      "grad_norm": 0.2935740182091683,
      "learning_rate": 1.7935954862584018e-05,
      "loss": 0.4321,
      "mean_token_accuracy": 0.8557445828098919,
      "step": 211
    },
    {
      "epoch": 0.28949389775539813,
      "grad_norm": 0.3228921123428817,
      "learning_rate": 1.7906814542158913e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.8537493570872144,
      "step": 212
    },
    {
      "epoch": 0.29085943500896133,
      "grad_norm": 0.28556153848386795,
      "learning_rate": 1.7877493982768527e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8495622520508093,
      "step": 213
    },
    {
      "epoch": 0.29222497226252453,
      "grad_norm": 0.31694287799599025,
      "learning_rate": 1.7847993852786612e-05,
      "loss": 0.4461,
      "mean_token_accuracy": 0.8494224622463822,
      "step": 214
    },
    {
      "epoch": 0.2935905095160877,
      "grad_norm": 0.3092018814084449,
      "learning_rate": 1.78183148246803e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.8543694459213177,
      "step": 215
    },
    {
      "epoch": 0.2949560467696509,
      "grad_norm": 0.2902008097484682,
      "learning_rate": 1.7788457574994777e-05,
      "loss": 0.4151,
      "mean_token_accuracy": 0.8582394007580811,
      "step": 216
    },
    {
      "epoch": 0.2963215840232141,
      "grad_norm": 0.3328026971189236,
      "learning_rate": 1.775842278433786e-05,
      "loss": 0.4255,
      "mean_token_accuracy": 0.8564833688598898,
      "step": 217
    },
    {
      "epoch": 0.2976871212767773,
      "grad_norm": 0.27435410637394864,
      "learning_rate": 1.772821113736449e-05,
      "loss": 0.4262,
      "mean_token_accuracy": 0.8559848881104419,
      "step": 218
    },
    {
      "epoch": 0.2990526585303405,
      "grad_norm": 0.3632335496079091,
      "learning_rate": 1.76978233227611e-05,
      "loss": 0.4107,
      "mean_token_accuracy": 0.8612579363381194,
      "step": 219
    },
    {
      "epoch": 0.3004181957839037,
      "grad_norm": 0.3167909582527359,
      "learning_rate": 1.7667260033229953e-05,
      "loss": 0.4342,
      "mean_token_accuracy": 0.8534075264867609,
      "step": 220
    },
    {
      "epoch": 0.3017837330374669,
      "grad_norm": 0.31022395154637067,
      "learning_rate": 1.7636521965473324e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8542319208452556,
      "step": 221
    },
    {
      "epoch": 0.3031492702910301,
      "grad_norm": 0.33282321447951396,
      "learning_rate": 1.760560982017762e-05,
      "loss": 0.4529,
      "mean_token_accuracy": 0.8473886646532048,
      "step": 222
    },
    {
      "epoch": 0.3045148075445933,
      "grad_norm": 0.3703839580647503,
      "learning_rate": 1.7574524301997425e-05,
      "loss": 0.4418,
      "mean_token_accuracy": 0.8520835295773124,
      "step": 223
    },
    {
      "epoch": 0.3058803447981565,
      "grad_norm": 0.29499276522228773,
      "learning_rate": 1.7543266119539424e-05,
      "loss": 0.4222,
      "mean_token_accuracy": 0.8572304378636991,
      "step": 224
    },
    {
      "epoch": 0.3072458820517197,
      "grad_norm": 0.36404736107381586,
      "learning_rate": 1.751183598534625e-05,
      "loss": 0.417,
      "mean_token_accuracy": 0.8592393267568642,
      "step": 225
    },
    {
      "epoch": 0.3086114193052829,
      "grad_norm": 0.31900880116039887,
      "learning_rate": 1.7480234615880247e-05,
      "loss": 0.4221,
      "mean_token_accuracy": 0.8570558484506429,
      "step": 226
    },
    {
      "epoch": 0.3099769565588461,
      "grad_norm": 0.39339547941721115,
      "learning_rate": 1.7448462731507133e-05,
      "loss": 0.4351,
      "mean_token_accuracy": 0.8529953241096725,
      "step": 227
    },
    {
      "epoch": 0.3113424938124093,
      "grad_norm": 0.38617429480416876,
      "learning_rate": 1.7416521056479577e-05,
      "loss": 0.4352,
      "mean_token_accuracy": 0.8529279155103227,
      "step": 228
    },
    {
      "epoch": 0.3127080310659725,
      "grad_norm": 0.3239118787173082,
      "learning_rate": 1.7384410318920698e-05,
      "loss": 0.4387,
      "mean_token_accuracy": 0.8516904522581532,
      "step": 229
    },
    {
      "epoch": 0.3140735683195357,
      "grad_norm": 0.3142116046355029,
      "learning_rate": 1.7352131250807466e-05,
      "loss": 0.4164,
      "mean_token_accuracy": 0.8592193212873477,
      "step": 230
    },
    {
      "epoch": 0.3154391055730989,
      "grad_norm": 0.3581861836859027,
      "learning_rate": 1.7319684587954e-05,
      "loss": 0.4175,
      "mean_token_accuracy": 0.8585542096028473,
      "step": 231
    },
    {
      "epoch": 0.3168046428266621,
      "grad_norm": 0.3011586770676647,
      "learning_rate": 1.728707106999482e-05,
      "loss": 0.4034,
      "mean_token_accuracy": 0.8631680964286239,
      "step": 232
    },
    {
      "epoch": 0.3181701800802253,
      "grad_norm": 0.3215575152723449,
      "learning_rate": 1.725429144036797e-05,
      "loss": 0.4146,
      "mean_token_accuracy": 0.8578385115822528,
      "step": 233
    },
    {
      "epoch": 0.3195357173337885,
      "grad_norm": 0.34263046362886657,
      "learning_rate": 1.722134644629807e-05,
      "loss": 0.4142,
      "mean_token_accuracy": 0.8586846673999058,
      "step": 234
    },
    {
      "epoch": 0.3209012545873517,
      "grad_norm": 0.33342256217650906,
      "learning_rate": 1.7188236838779297e-05,
      "loss": 0.44,
      "mean_token_accuracy": 0.851783206820397,
      "step": 235
    },
    {
      "epoch": 0.3222667918409149,
      "grad_norm": 0.3442067446188413,
      "learning_rate": 1.7154963372558246e-05,
      "loss": 0.4415,
      "mean_token_accuracy": 0.8494488715634262,
      "step": 236
    },
    {
      "epoch": 0.3236323290944781,
      "grad_norm": 0.39286054840274137,
      "learning_rate": 1.712152680611675e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8519838146425381,
      "step": 237
    },
    {
      "epoch": 0.3249978663480413,
      "grad_norm": 0.3707489334417565,
      "learning_rate": 1.7087927901654558e-05,
      "loss": 0.4018,
      "mean_token_accuracy": 0.8628199801152938,
      "step": 238
    },
    {
      "epoch": 0.3263634036016045,
      "grad_norm": 0.32915317985042997,
      "learning_rate": 1.7054167425071995e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8583234741171639,
      "step": 239
    },
    {
      "epoch": 0.3277289408551677,
      "grad_norm": 0.3582433826817124,
      "learning_rate": 1.702024614595248e-05,
      "loss": 0.4245,
      "mean_token_accuracy": 0.8573367967440888,
      "step": 240
    },
    {
      "epoch": 0.3290944781087309,
      "grad_norm": 0.33842853808026,
      "learning_rate": 1.6986164837544987e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8586180123428547,
      "step": 241
    },
    {
      "epoch": 0.3304600153622941,
      "grad_norm": 0.2994720944821025,
      "learning_rate": 1.6951924276746425e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8523116286165904,
      "step": 242
    },
    {
      "epoch": 0.3318255526158573,
      "grad_norm": 0.3387086926170212,
      "learning_rate": 1.6917525244083918e-05,
      "loss": 0.4512,
      "mean_token_accuracy": 0.8479398316838589,
      "step": 243
    },
    {
      "epoch": 0.3331910898694205,
      "grad_norm": 0.28444793823353953,
      "learning_rate": 1.688296852369703e-05,
      "loss": 0.4227,
      "mean_token_accuracy": 0.857218350086272,
      "step": 244
    },
    {
      "epoch": 0.3345566271229837,
      "grad_norm": 0.32234365149946503,
      "learning_rate": 1.6848254903319866e-05,
      "loss": 0.433,
      "mean_token_accuracy": 0.8537039316837896,
      "step": 245
    },
    {
      "epoch": 0.3359221643765469,
      "grad_norm": 0.3091069266476509,
      "learning_rate": 1.6813385174263137e-05,
      "loss": 0.4195,
      "mean_token_accuracy": 0.8581039754478647,
      "step": 246
    },
    {
      "epoch": 0.33728770163011007,
      "grad_norm": 0.3640194501208851,
      "learning_rate": 1.677836013139611e-05,
      "loss": 0.4255,
      "mean_token_accuracy": 0.856087771880514,
      "step": 247
    },
    {
      "epoch": 0.33865323888367327,
      "grad_norm": 0.2842912100028283,
      "learning_rate": 1.6743180573128494e-05,
      "loss": 0.4063,
      "mean_token_accuracy": 0.8612077361066686,
      "step": 248
    },
    {
      "epoch": 0.34001877613723647,
      "grad_norm": 0.33927830284172017,
      "learning_rate": 1.6707847301392237e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.854018667863303,
      "step": 249
    },
    {
      "epoch": 0.34138431339079967,
      "grad_norm": 0.3380064283112177,
      "learning_rate": 1.6672361121623238e-05,
      "loss": 0.4196,
      "mean_token_accuracy": 0.858289578175996,
      "step": 250
    },
    {
      "epoch": 0.3427498506443629,
      "grad_norm": 0.2921545005723523,
      "learning_rate": 1.6636722842743013e-05,
      "loss": 0.418,
      "mean_token_accuracy": 0.8576834083602248,
      "step": 251
    },
    {
      "epoch": 0.3441153878979261,
      "grad_norm": 0.30696667285073703,
      "learning_rate": 1.660093327714022e-05,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8571891667514102,
      "step": 252
    },
    {
      "epoch": 0.3454809251514893,
      "grad_norm": 0.2692563224404543,
      "learning_rate": 1.656499324065217e-05,
      "loss": 0.424,
      "mean_token_accuracy": 0.8564762224897754,
      "step": 253
    },
    {
      "epoch": 0.3468464624050525,
      "grad_norm": 0.2902797695424939,
      "learning_rate": 1.6528903552546207e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8531339042109244,
      "step": 254
    },
    {
      "epoch": 0.3482119996586157,
      "grad_norm": 0.3054595022905831,
      "learning_rate": 1.6492665035501048e-05,
      "loss": 0.4304,
      "mean_token_accuracy": 0.8550912353650719,
      "step": 255
    },
    {
      "epoch": 0.3495775369121789,
      "grad_norm": 0.30207364986816204,
      "learning_rate": 1.6456278515588023e-05,
      "loss": 0.4185,
      "mean_token_accuracy": 0.8576527365281009,
      "step": 256
    },
    {
      "epoch": 0.3509430741657421,
      "grad_norm": 0.29635598975960087,
      "learning_rate": 1.6419744822252255e-05,
      "loss": 0.4078,
      "mean_token_accuracy": 0.8610621124104776,
      "step": 257
    },
    {
      "epoch": 0.3523086114193053,
      "grad_norm": 0.31915252694298046,
      "learning_rate": 1.638306478829373e-05,
      "loss": 0.4231,
      "mean_token_accuracy": 0.8561889856733563,
      "step": 258
    },
    {
      "epoch": 0.3536741486728685,
      "grad_norm": 0.31923467215480067,
      "learning_rate": 1.634623924984833e-05,
      "loss": 0.4135,
      "mean_token_accuracy": 0.8592538002988109,
      "step": 259
    },
    {
      "epoch": 0.3550396859264317,
      "grad_norm": 0.30692282494391426,
      "learning_rate": 1.6309269046368777e-05,
      "loss": 0.3966,
      "mean_token_accuracy": 0.8640884833628784,
      "step": 260
    },
    {
      "epoch": 0.3564052231799949,
      "grad_norm": 0.27482122390015884,
      "learning_rate": 1.627215502060548e-05,
      "loss": 0.4216,
      "mean_token_accuracy": 0.8559256932307063,
      "step": 261
    },
    {
      "epoch": 0.3577707604335581,
      "grad_norm": 0.34179806060032714,
      "learning_rate": 1.6234898018587336e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8491678521649152,
      "step": 262
    },
    {
      "epoch": 0.3591362976871213,
      "grad_norm": 0.3011617462318224,
      "learning_rate": 1.619749888960245e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.8554996248647118,
      "step": 263
    },
    {
      "epoch": 0.3605018349406845,
      "grad_norm": 0.31807503774612333,
      "learning_rate": 1.615995848617876e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8559916893442254,
      "step": 264
    },
    {
      "epoch": 0.3618673721942477,
      "grad_norm": 0.3316772571952849,
      "learning_rate": 1.612227766406461e-05,
      "loss": 0.4114,
      "mean_token_accuracy": 0.8605685266181821,
      "step": 265
    },
    {
      "epoch": 0.3632329094478109,
      "grad_norm": 0.2831512502783301,
      "learning_rate": 1.6084457282209244e-05,
      "loss": 0.4182,
      "mean_token_accuracy": 0.8572316390854738,
      "step": 266
    },
    {
      "epoch": 0.3645984467013741,
      "grad_norm": 0.3020965586894018,
      "learning_rate": 1.6046498202743232e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8552705974923581,
      "step": 267
    },
    {
      "epoch": 0.3659639839549373,
      "grad_norm": 0.34579468143149206,
      "learning_rate": 1.6008401290958806e-05,
      "loss": 0.4224,
      "mean_token_accuracy": 0.8574742487392688,
      "step": 268
    },
    {
      "epoch": 0.3673295212085005,
      "grad_norm": 0.3213456748377343,
      "learning_rate": 1.5970167415290142e-05,
      "loss": 0.4191,
      "mean_token_accuracy": 0.8592830241547558,
      "step": 269
    },
    {
      "epoch": 0.3686950584620637,
      "grad_norm": 0.27169537891651463,
      "learning_rate": 1.5931797447293553e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.8565419088799743,
      "step": 270
    },
    {
      "epoch": 0.3700605957156269,
      "grad_norm": 0.31485256379338494,
      "learning_rate": 1.5893292261627644e-05,
      "loss": 0.4211,
      "mean_token_accuracy": 0.8583665894914854,
      "step": 271
    },
    {
      "epoch": 0.3714261329691901,
      "grad_norm": 0.34197779647959986,
      "learning_rate": 1.5854652736033353e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8586362735400643,
      "step": 272
    },
    {
      "epoch": 0.3727916702227533,
      "grad_norm": 0.3463622744135729,
      "learning_rate": 1.5815879751313957e-05,
      "loss": 0.4153,
      "mean_token_accuracy": 0.8590972009223808,
      "step": 273
    },
    {
      "epoch": 0.3741572074763165,
      "grad_norm": 0.2854754568932129,
      "learning_rate": 1.577697419131497e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8549681291103441,
      "step": 274
    },
    {
      "epoch": 0.3755227447298797,
      "grad_norm": 0.3612800362541057,
      "learning_rate": 1.5737936942904025e-05,
      "loss": 0.4164,
      "mean_token_accuracy": 0.8589556245901419,
      "step": 275
    },
    {
      "epoch": 0.3768882819834429,
      "grad_norm": 0.3095104128297478,
      "learning_rate": 1.5698768895950644e-05,
      "loss": 0.4196,
      "mean_token_accuracy": 0.8573392516346519,
      "step": 276
    },
    {
      "epoch": 0.3782538192370061,
      "grad_norm": 0.28011182927413847,
      "learning_rate": 1.5659470943305956e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8575836185111236,
      "step": 277
    },
    {
      "epoch": 0.37961935649056927,
      "grad_norm": 0.38665599972326675,
      "learning_rate": 1.5620043980782327e-05,
      "loss": 0.4228,
      "mean_token_accuracy": 0.8577991932053891,
      "step": 278
    },
    {
      "epoch": 0.38098489374413247,
      "grad_norm": 0.3045801906729731,
      "learning_rate": 1.5580488907132972e-05,
      "loss": 0.4127,
      "mean_token_accuracy": 0.859320001053601,
      "step": 279
    },
    {
      "epoch": 0.38235043099769567,
      "grad_norm": 0.3027464188711677,
      "learning_rate": 1.554080662403144e-05,
      "loss": 0.4344,
      "mean_token_accuracy": 0.8530298389418411,
      "step": 280
    },
    {
      "epoch": 0.38371596825125887,
      "grad_norm": 0.30735481784619834,
      "learning_rate": 1.5500998036051075e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8564421780464974,
      "step": 281
    },
    {
      "epoch": 0.38508150550482206,
      "grad_norm": 0.28386427155284427,
      "learning_rate": 1.546106405064438e-05,
      "loss": 0.4447,
      "mean_token_accuracy": 0.8507831934070248,
      "step": 282
    },
    {
      "epoch": 0.38644704275838526,
      "grad_norm": 0.3157137742956653,
      "learning_rate": 1.5421005578122356e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8566646495893814,
      "step": 283
    },
    {
      "epoch": 0.38781258001194846,
      "grad_norm": 0.3109770970705045,
      "learning_rate": 1.5380823531633727e-05,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8575001524672929,
      "step": 284
    },
    {
      "epoch": 0.38917811726551166,
      "grad_norm": 0.3467377136844463,
      "learning_rate": 1.5340518827144145e-05,
      "loss": 0.4233,
      "mean_token_accuracy": 0.8582951077732086,
      "step": 285
    },
    {
      "epoch": 0.39054365451907486,
      "grad_norm": 0.3381885844162203,
      "learning_rate": 1.5300092383415282e-05,
      "loss": 0.4289,
      "mean_token_accuracy": 0.8549039858412707,
      "step": 286
    },
    {
      "epoch": 0.39190919177263805,
      "grad_norm": 0.3297678208614477,
      "learning_rate": 1.525954512198392e-05,
      "loss": 0.4096,
      "mean_token_accuracy": 0.8613219951830501,
      "step": 287
    },
    {
      "epoch": 0.39327472902620125,
      "grad_norm": 0.3368037995093397,
      "learning_rate": 1.5218877967140921e-05,
      "loss": 0.4224,
      "mean_token_accuracy": 0.8548496161946588,
      "step": 288
    },
    {
      "epoch": 0.39464026627976445,
      "grad_norm": 0.3212698070828666,
      "learning_rate": 1.517809184591017e-05,
      "loss": 0.4263,
      "mean_token_accuracy": 0.8574614580441292,
      "step": 289
    },
    {
      "epoch": 0.39600580353332765,
      "grad_norm": 0.40537049798129476,
      "learning_rate": 1.5137187688027437e-05,
      "loss": 0.43,
      "mean_token_accuracy": 0.8550197749735438,
      "step": 290
    },
    {
      "epoch": 0.39737134078689085,
      "grad_norm": 0.31835336382101076,
      "learning_rate": 1.5096166425919176e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8548674541736617,
      "step": 291
    },
    {
      "epoch": 0.39873687804045405,
      "grad_norm": 0.34618990419302037,
      "learning_rate": 1.5055028994681284e-05,
      "loss": 0.409,
      "mean_token_accuracy": 0.8607840254015119,
      "step": 292
    },
    {
      "epoch": 0.40010241529401724,
      "grad_norm": 0.350507807135204,
      "learning_rate": 1.5013776332057786e-05,
      "loss": 0.4151,
      "mean_token_accuracy": 0.8577964045490754,
      "step": 293
    },
    {
      "epoch": 0.40146795254758044,
      "grad_norm": 0.30189871647045075,
      "learning_rate": 1.4972409378419439e-05,
      "loss": 0.4094,
      "mean_token_accuracy": 0.8607436168222747,
      "step": 294
    },
    {
      "epoch": 0.40283348980114364,
      "grad_norm": 0.3016208540740829,
      "learning_rate": 1.4930929076742317e-05,
      "loss": 0.4151,
      "mean_token_accuracy": 0.8595533954555142,
      "step": 295
    },
    {
      "epoch": 0.40419902705470684,
      "grad_norm": 0.29033881446311255,
      "learning_rate": 1.4889336372586305e-05,
      "loss": 0.3928,
      "mean_token_accuracy": 0.8637799769336061,
      "step": 296
    },
    {
      "epoch": 0.40556456430827004,
      "grad_norm": 0.3668162312007593,
      "learning_rate": 1.4847632214073548e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8499288824377587,
      "step": 297
    },
    {
      "epoch": 0.40693010156183324,
      "grad_norm": 0.32449852074777696,
      "learning_rate": 1.4805817551866839e-05,
      "loss": 0.4084,
      "mean_token_accuracy": 0.8605696761410078,
      "step": 298
    },
    {
      "epoch": 0.40829563881539643,
      "grad_norm": 0.2970564472354985,
      "learning_rate": 1.4763893339147942e-05,
      "loss": 0.418,
      "mean_token_accuracy": 0.8573568392411037,
      "step": 299
    },
    {
      "epoch": 0.40966117606895963,
      "grad_norm": 0.32582221983564175,
      "learning_rate": 1.4721860531595868e-05,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8602134585041119,
      "step": 300
    },
    {
      "epoch": 0.41102671332252283,
      "grad_norm": 0.3089086009639669,
      "learning_rate": 1.4679720087365097e-05,
      "loss": 0.4099,
      "mean_token_accuracy": 0.8610357533905291,
      "step": 301
    },
    {
      "epoch": 0.41239225057608603,
      "grad_norm": 0.32593412647065384,
      "learning_rate": 1.4637472967063721e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8527080209938287,
      "step": 302
    },
    {
      "epoch": 0.4137577878296492,
      "grad_norm": 0.3277327156387932,
      "learning_rate": 1.4595120133731564e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8569046290075678,
      "step": 303
    },
    {
      "epoch": 0.4151233250832124,
      "grad_norm": 0.3730986070237096,
      "learning_rate": 1.4552662552818211e-05,
      "loss": 0.4433,
      "mean_token_accuracy": 0.8504104581423189,
      "step": 304
    },
    {
      "epoch": 0.4164888623367756,
      "grad_norm": 0.28719200698202685,
      "learning_rate": 1.451010119216102e-05,
      "loss": 0.4369,
      "mean_token_accuracy": 0.8519207939028978,
      "step": 305
    },
    {
      "epoch": 0.4178543995903388,
      "grad_norm": 0.2597878117331174,
      "learning_rate": 1.446743702196304e-05,
      "loss": 0.4187,
      "mean_token_accuracy": 0.8574448859814668,
      "step": 306
    },
    {
      "epoch": 0.419219936843902,
      "grad_norm": 0.34783364098922354,
      "learning_rate": 1.4424671014770906e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8523371430850553,
      "step": 307
    },
    {
      "epoch": 0.4205854740974652,
      "grad_norm": 0.33054277050758557,
      "learning_rate": 1.4381804145452672e-05,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8585607772789662,
      "step": 308
    },
    {
      "epoch": 0.4219510113510284,
      "grad_norm": 0.2495797590927163,
      "learning_rate": 1.4338837391175582e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.855114414007386,
      "step": 309
    },
    {
      "epoch": 0.4233165486045916,
      "grad_norm": 0.35728528385485553,
      "learning_rate": 1.4295771731383799e-05,
      "loss": 0.434,
      "mean_token_accuracy": 0.8536842164352073,
      "step": 310
    },
    {
      "epoch": 0.4246820858581548,
      "grad_norm": 0.3288240381017802,
      "learning_rate": 1.4252608147776067e-05,
      "loss": 0.4051,
      "mean_token_accuracy": 0.8619649344224427,
      "step": 311
    },
    {
      "epoch": 0.426047623111718,
      "grad_norm": 0.3883949212610513,
      "learning_rate": 1.4209347624283352e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8541025576014111,
      "step": 312
    },
    {
      "epoch": 0.4274131603652812,
      "grad_norm": 0.3012394552571007,
      "learning_rate": 1.4165991147046404e-05,
      "loss": 0.4121,
      "mean_token_accuracy": 0.8598195582006549,
      "step": 313
    },
    {
      "epoch": 0.4287786976188444,
      "grad_norm": 0.3197771206984412,
      "learning_rate": 1.4122539704393265e-05,
      "loss": 0.4237,
      "mean_token_accuracy": 0.8560034937668209,
      "step": 314
    },
    {
      "epoch": 0.4301442348724076,
      "grad_norm": 0.3057046597401257,
      "learning_rate": 1.4078994286816768e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.853573226005123,
      "step": 315
    },
    {
      "epoch": 0.4315097721259708,
      "grad_norm": 0.3422003725649723,
      "learning_rate": 1.4035355886951924e-05,
      "loss": 0.4327,
      "mean_token_accuracy": 0.8533252072862426,
      "step": 316
    },
    {
      "epoch": 0.432875309379534,
      "grad_norm": 0.2909937625605861,
      "learning_rate": 1.3991625499553325e-05,
      "loss": 0.4373,
      "mean_token_accuracy": 0.8526268972423322,
      "step": 317
    },
    {
      "epoch": 0.4342408466330972,
      "grad_norm": 0.2810097508624511,
      "learning_rate": 1.3947804121472453e-05,
      "loss": 0.4209,
      "mean_token_accuracy": 0.8580495358481994,
      "step": 318
    },
    {
      "epoch": 0.4356063838866604,
      "grad_norm": 0.29587543321646376,
      "learning_rate": 1.3903892751634949e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8580909047258651,
      "step": 319
    },
    {
      "epoch": 0.4369719211402236,
      "grad_norm": 0.2803397206271039,
      "learning_rate": 1.3859892391017867e-05,
      "loss": 0.398,
      "mean_token_accuracy": 0.8636022295266738,
      "step": 320
    },
    {
      "epoch": 0.4383374583937868,
      "grad_norm": 0.3031649571676904,
      "learning_rate": 1.3815804042626828e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8540270034472657,
      "step": 321
    },
    {
      "epoch": 0.43970299564735,
      "grad_norm": 0.2868053671287889,
      "learning_rate": 1.3771628711473173e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.8500793417904773,
      "step": 322
    },
    {
      "epoch": 0.4410685329009132,
      "grad_norm": 0.2607596548746336,
      "learning_rate": 1.3727367404551055e-05,
      "loss": 0.4194,
      "mean_token_accuracy": 0.85756931958572,
      "step": 323
    },
    {
      "epoch": 0.4424340701544764,
      "grad_norm": 0.31970960349713734,
      "learning_rate": 1.368302113081447e-05,
      "loss": 0.3988,
      "mean_token_accuracy": 0.8640139707624799,
      "step": 324
    },
    {
      "epoch": 0.4437996074080396,
      "grad_norm": 0.3073106580181061,
      "learning_rate": 1.3638590901154276e-05,
      "loss": 0.4158,
      "mean_token_accuracy": 0.8588068782465295,
      "step": 325
    },
    {
      "epoch": 0.4451651446616028,
      "grad_norm": 0.30286141069120004,
      "learning_rate": 1.3594077728375129e-05,
      "loss": 0.4201,
      "mean_token_accuracy": 0.8569773919052446,
      "step": 326
    },
    {
      "epoch": 0.446530681915166,
      "grad_norm": 0.279965967756229,
      "learning_rate": 1.3549482627172412e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8531155705252534,
      "step": 327
    },
    {
      "epoch": 0.4478962191687292,
      "grad_norm": 0.30169596805937576,
      "learning_rate": 1.3504806614109098e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8567625209561366,
      "step": 328
    },
    {
      "epoch": 0.4492617564222924,
      "grad_norm": 0.3459438236471543,
      "learning_rate": 1.3460050707592581e-05,
      "loss": 0.4336,
      "mean_token_accuracy": 0.8529489776321926,
      "step": 329
    },
    {
      "epoch": 0.4506272936758556,
      "grad_norm": 0.30726581139747894,
      "learning_rate": 1.341521592785145e-05,
      "loss": 0.399,
      "mean_token_accuracy": 0.8630228365259139,
      "step": 330
    },
    {
      "epoch": 0.4519928309294188,
      "grad_norm": 0.3053218799713375,
      "learning_rate": 1.3370303296912248e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8519843865306225,
      "step": 331
    },
    {
      "epoch": 0.453358368182982,
      "grad_norm": 0.34629772646315893,
      "learning_rate": 1.332531383857616e-05,
      "loss": 0.4112,
      "mean_token_accuracy": 0.860751373812898,
      "step": 332
    },
    {
      "epoch": 0.4547239054365452,
      "grad_norm": 0.35416501044407706,
      "learning_rate": 1.328024857839569e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.857242064978725,
      "step": 333
    },
    {
      "epoch": 0.45608944269010837,
      "grad_norm": 0.30654844603106524,
      "learning_rate": 1.3235108543651272e-05,
      "loss": 0.4152,
      "mean_token_accuracy": 0.8594253890394505,
      "step": 334
    },
    {
      "epoch": 0.45745497994367157,
      "grad_norm": 0.35789751735364733,
      "learning_rate": 1.3189894763327851e-05,
      "loss": 0.4352,
      "mean_token_accuracy": 0.8524388902413612,
      "step": 335
    },
    {
      "epoch": 0.45882051719723477,
      "grad_norm": 0.27296525421750206,
      "learning_rate": 1.3144608268091435e-05,
      "loss": 0.4125,
      "mean_token_accuracy": 0.8600694337472964,
      "step": 336
    },
    {
      "epoch": 0.46018605445079797,
      "grad_norm": 0.35167998845661236,
      "learning_rate": 1.3099250090265599e-05,
      "loss": 0.4212,
      "mean_token_accuracy": 0.8573687132989843,
      "step": 337
    },
    {
      "epoch": 0.46155159170436116,
      "grad_norm": 0.3399733274636042,
      "learning_rate": 1.3053821263807947e-05,
      "loss": 0.4133,
      "mean_token_accuracy": 0.8588422955112552,
      "step": 338
    },
    {
      "epoch": 0.46291712895792436,
      "grad_norm": 0.2843523201867168,
      "learning_rate": 1.3008322824286554e-05,
      "loss": 0.4217,
      "mean_token_accuracy": 0.8581068469248813,
      "step": 339
    },
    {
      "epoch": 0.46428266621148756,
      "grad_norm": 0.25579241148803916,
      "learning_rate": 1.2962755808856341e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8531418182041834,
      "step": 340
    },
    {
      "epoch": 0.46564820346505076,
      "grad_norm": 0.2849802262084464,
      "learning_rate": 1.2917121256235454e-05,
      "loss": 0.4063,
      "mean_token_accuracy": 0.8619580010780344,
      "step": 341
    },
    {
      "epoch": 0.46701374071861396,
      "grad_norm": 0.28376439896546823,
      "learning_rate": 1.2871420206681573e-05,
      "loss": 0.4126,
      "mean_token_accuracy": 0.8584155141530699,
      "step": 342
    },
    {
      "epoch": 0.46837927797217715,
      "grad_norm": 0.27659846392534093,
      "learning_rate": 1.2825653701968199e-05,
      "loss": 0.4028,
      "mean_token_accuracy": 0.8630971886050622,
      "step": 343
    },
    {
      "epoch": 0.46974481522574035,
      "grad_norm": 0.2680308561203238,
      "learning_rate": 1.2779822785360913e-05,
      "loss": 0.4106,
      "mean_token_accuracy": 0.8594322782022719,
      "step": 344
    },
    {
      "epoch": 0.47111035247930355,
      "grad_norm": 0.31082123251549165,
      "learning_rate": 1.2733928501593587e-05,
      "loss": 0.4409,
      "mean_token_accuracy": 0.8504463201078798,
      "step": 345
    },
    {
      "epoch": 0.47247588973286675,
      "grad_norm": 0.28196867432168504,
      "learning_rate": 1.2687971896844575e-05,
      "loss": 0.4143,
      "mean_token_accuracy": 0.8599393422041681,
      "step": 346
    },
    {
      "epoch": 0.47384142698642995,
      "grad_norm": 0.30055986798006457,
      "learning_rate": 1.2641954018712863e-05,
      "loss": 0.427,
      "mean_token_accuracy": 0.8553718131914751,
      "step": 347
    },
    {
      "epoch": 0.47520696423999315,
      "grad_norm": 0.3360073322586856,
      "learning_rate": 1.2595875916194188e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8513545891221347,
      "step": 348
    },
    {
      "epoch": 0.47657250149355634,
      "grad_norm": 0.27942894694642584,
      "learning_rate": 1.2549738639657117e-05,
      "loss": 0.4069,
      "mean_token_accuracy": 0.8616844269198048,
      "step": 349
    },
    {
      "epoch": 0.47793803874711954,
      "grad_norm": 0.25876338591566067,
      "learning_rate": 1.2503543240819127e-05,
      "loss": 0.4189,
      "mean_token_accuracy": 0.8584846388067386,
      "step": 350
    },
    {
      "epoch": 0.4793035760006828,
      "grad_norm": 0.38300218448302764,
      "learning_rate": 1.2457290772722607e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8563944060284168,
      "step": 351
    },
    {
      "epoch": 0.480669113254246,
      "grad_norm": 0.29865375211927264,
      "learning_rate": 1.2410982289710865e-05,
      "loss": 0.4274,
      "mean_token_accuracy": 0.8550406509222035,
      "step": 352
    },
    {
      "epoch": 0.4820346505078092,
      "grad_norm": 0.29262760726515985,
      "learning_rate": 1.2364618847404088e-05,
      "loss": 0.4195,
      "mean_token_accuracy": 0.856524809953863,
      "step": 353
    },
    {
      "epoch": 0.4834001877613724,
      "grad_norm": 0.33358761083999955,
      "learning_rate": 1.2318201502675285e-05,
      "loss": 0.4376,
      "mean_token_accuracy": 0.8536589150399464,
      "step": 354
    },
    {
      "epoch": 0.4847657250149356,
      "grad_norm": 0.24446626288123202,
      "learning_rate": 1.227173131362619e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8556876461790799,
      "step": 355
    },
    {
      "epoch": 0.4861312622684988,
      "grad_norm": 0.2998978259766614,
      "learning_rate": 1.2225209339563144e-05,
      "loss": 0.4244,
      "mean_token_accuracy": 0.8561228530745463,
      "step": 356
    },
    {
      "epoch": 0.487496799522062,
      "grad_norm": 0.3206206105210003,
      "learning_rate": 1.2178636640972954e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8534380165849528,
      "step": 357
    },
    {
      "epoch": 0.4888623367756252,
      "grad_norm": 0.3084930416049475,
      "learning_rate": 1.2132014279498702e-05,
      "loss": 0.4288,
      "mean_token_accuracy": 0.8533390096208622,
      "step": 358
    },
    {
      "epoch": 0.4902278740291884,
      "grad_norm": 0.28357516576751085,
      "learning_rate": 1.2085343317915565e-05,
      "loss": 0.4243,
      "mean_token_accuracy": 0.8562038691159706,
      "step": 359
    },
    {
      "epoch": 0.4915934112827516,
      "grad_norm": 0.33187103925787603,
      "learning_rate": 1.2038624820106572e-05,
      "loss": 0.4333,
      "mean_token_accuracy": 0.852135547957135,
      "step": 360
    },
    {
      "epoch": 0.4929589485363148,
      "grad_norm": 0.2747099173609463,
      "learning_rate": 1.1991859851038362e-05,
      "loss": 0.4007,
      "mean_token_accuracy": 0.8638977286823286,
      "step": 361
    },
    {
      "epoch": 0.494324485789878,
      "grad_norm": 0.3635110977798377,
      "learning_rate": 1.1945049476736905e-05,
      "loss": 0.3958,
      "mean_token_accuracy": 0.8647084604295199,
      "step": 362
    },
    {
      "epoch": 0.4956900230434412,
      "grad_norm": 0.28529172664511565,
      "learning_rate": 1.1898194764263198e-05,
      "loss": 0.4233,
      "mean_token_accuracy": 0.8569803575165137,
      "step": 363
    },
    {
      "epoch": 0.4970555602970044,
      "grad_norm": 0.27537692377864037,
      "learning_rate": 1.1851296781688952e-05,
      "loss": 0.4073,
      "mean_token_accuracy": 0.8600902266025654,
      "step": 364
    },
    {
      "epoch": 0.49842109755056757,
      "grad_norm": 0.32325185164148623,
      "learning_rate": 1.1804356598072223e-05,
      "loss": 0.4176,
      "mean_token_accuracy": 0.8587095805096491,
      "step": 365
    },
    {
      "epoch": 0.49978663480413077,
      "grad_norm": 0.2940543368854904,
      "learning_rate": 1.1757375283433077e-05,
      "loss": 0.421,
      "mean_token_accuracy": 0.8564938642066868,
      "step": 366
    },
    {
      "epoch": 0.5011521720576939,
      "grad_norm": 0.28550857526720214,
      "learning_rate": 1.1710353908729157e-05,
      "loss": 0.4032,
      "mean_token_accuracy": 0.8619838649329472,
      "step": 367
    },
    {
      "epoch": 0.5025177093112572,
      "grad_norm": 0.29699607198851946,
      "learning_rate": 1.1663293545831302e-05,
      "loss": 0.4146,
      "mean_token_accuracy": 0.8589313902193905,
      "step": 368
    },
    {
      "epoch": 0.5038832465648203,
      "grad_norm": 0.25709714555087276,
      "learning_rate": 1.1616195267499102e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8524904329384897,
      "step": 369
    },
    {
      "epoch": 0.5052487838183836,
      "grad_norm": 0.2623470396493795,
      "learning_rate": 1.1569060147356441e-05,
      "loss": 0.4004,
      "mean_token_accuracy": 0.8624617116263626,
      "step": 370
    },
    {
      "epoch": 0.5066143210719467,
      "grad_norm": 0.27349918771238607,
      "learning_rate": 1.1521889259867032e-05,
      "loss": 0.3875,
      "mean_token_accuracy": 0.8656588419393155,
      "step": 371
    },
    {
      "epoch": 0.50797985832551,
      "grad_norm": 0.2659088919583038,
      "learning_rate": 1.1474683680309913e-05,
      "loss": 0.422,
      "mean_token_accuracy": 0.8569432102568713,
      "step": 372
    },
    {
      "epoch": 0.5093453955790731,
      "grad_norm": 0.24245036453653618,
      "learning_rate": 1.1427444484754942e-05,
      "loss": 0.4345,
      "mean_token_accuracy": 0.8533666506270393,
      "step": 373
    },
    {
      "epoch": 0.5107109328326364,
      "grad_norm": 0.3003745935393995,
      "learning_rate": 1.138017275003827e-05,
      "loss": 0.4141,
      "mean_token_accuracy": 0.859278335907566,
      "step": 374
    },
    {
      "epoch": 0.5120764700861995,
      "grad_norm": 0.26886003162184047,
      "learning_rate": 1.133286955373779e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8502673324909059,
      "step": 375
    },
    {
      "epoch": 0.5134420073397628,
      "grad_norm": 0.2390413269524559,
      "learning_rate": 1.1285535974148576e-05,
      "loss": 0.3997,
      "mean_token_accuracy": 0.8639690222498632,
      "step": 376
    },
    {
      "epoch": 0.5148075445933259,
      "grad_norm": 0.2544175262225982,
      "learning_rate": 1.1238173090258292e-05,
      "loss": 0.3998,
      "mean_token_accuracy": 0.8636630321983972,
      "step": 377
    },
    {
      "epoch": 0.5161730818468891,
      "grad_norm": 0.23579375374714043,
      "learning_rate": 1.1190781981722622e-05,
      "loss": 0.4213,
      "mean_token_accuracy": 0.8574136326780977,
      "step": 378
    },
    {
      "epoch": 0.5175386191004523,
      "grad_norm": 0.3009505232094782,
      "learning_rate": 1.1143363728840626e-05,
      "loss": 0.4048,
      "mean_token_accuracy": 0.8630919807826921,
      "step": 379
    },
    {
      "epoch": 0.5189041563540155,
      "grad_norm": 0.3008130036826914,
      "learning_rate": 1.1095919412530136e-05,
      "loss": 0.4162,
      "mean_token_accuracy": 0.8576130359385502,
      "step": 380
    },
    {
      "epoch": 0.5202696936075787,
      "grad_norm": 0.2952343133789482,
      "learning_rate": 1.1048450114303111e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.850824829194507,
      "step": 381
    },
    {
      "epoch": 0.5216352308611419,
      "grad_norm": 0.258661138214629,
      "learning_rate": 1.1000956916240985e-05,
      "loss": 0.4086,
      "mean_token_accuracy": 0.8620580382687907,
      "step": 382
    },
    {
      "epoch": 0.5230007681147051,
      "grad_norm": 0.25494116388793103,
      "learning_rate": 1.0953440900969993e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8565843185391879,
      "step": 383
    },
    {
      "epoch": 0.5243663053682683,
      "grad_norm": 0.27361519707795223,
      "learning_rate": 1.09059031516365e-05,
      "loss": 0.4102,
      "mean_token_accuracy": 0.8609546120002327,
      "step": 384
    },
    {
      "epoch": 0.5257318426218315,
      "grad_norm": 0.2576222350047768,
      "learning_rate": 1.0858344751882304e-05,
      "loss": 0.4069,
      "mean_token_accuracy": 0.860766222892137,
      "step": 385
    },
    {
      "epoch": 0.5270973798753947,
      "grad_norm": 0.23779312285209903,
      "learning_rate": 1.0810766785819947e-05,
      "loss": 0.4228,
      "mean_token_accuracy": 0.8588812441258473,
      "step": 386
    },
    {
      "epoch": 0.5284629171289579,
      "grad_norm": 0.28934124430046615,
      "learning_rate": 1.0763170338007978e-05,
      "loss": 0.4107,
      "mean_token_accuracy": 0.8604884551869458,
      "step": 387
    },
    {
      "epoch": 0.5298284543825211,
      "grad_norm": 0.2466549449805162,
      "learning_rate": 1.0715556493426263e-05,
      "loss": 0.4263,
      "mean_token_accuracy": 0.8557152597091608,
      "step": 388
    },
    {
      "epoch": 0.5311939916360843,
      "grad_norm": 0.27325712567200283,
      "learning_rate": 1.0667926337451217e-05,
      "loss": 0.4149,
      "mean_token_accuracy": 0.8584272502154298,
      "step": 389
    },
    {
      "epoch": 0.5325595288896475,
      "grad_norm": 0.25477351099504564,
      "learning_rate": 1.0620280955831088e-05,
      "loss": 0.4138,
      "mean_token_accuracy": 0.8593247446112272,
      "step": 390
    },
    {
      "epoch": 0.5339250661432107,
      "grad_norm": 0.25978114932856317,
      "learning_rate": 1.0572621434661201e-05,
      "loss": 0.4189,
      "mean_token_accuracy": 0.8573288866024139,
      "step": 391
    },
    {
      "epoch": 0.5352906033967739,
      "grad_norm": 0.26074134187169073,
      "learning_rate": 1.0524948860359194e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8560287289485933,
      "step": 392
    },
    {
      "epoch": 0.5366561406503371,
      "grad_norm": 0.2606859248765302,
      "learning_rate": 1.0477264319640253e-05,
      "loss": 0.4117,
      "mean_token_accuracy": 0.8600141371640406,
      "step": 393
    },
    {
      "epoch": 0.5380216779039003,
      "grad_norm": 0.2260622200890888,
      "learning_rate": 1.0429568899492349e-05,
      "loss": 0.3984,
      "mean_token_accuracy": 0.8646881070534751,
      "step": 394
    },
    {
      "epoch": 0.5393872151574635,
      "grad_norm": 0.2270692517916632,
      "learning_rate": 1.038186368715145e-05,
      "loss": 0.3935,
      "mean_token_accuracy": 0.8645081124206024,
      "step": 395
    },
    {
      "epoch": 0.5407527524110267,
      "grad_norm": 0.25880192379351896,
      "learning_rate": 1.0334149770076747e-05,
      "loss": 0.4238,
      "mean_token_accuracy": 0.8556846129868956,
      "step": 396
    },
    {
      "epoch": 0.5421182896645899,
      "grad_norm": 0.28051700201569574,
      "learning_rate": 1.0286428235925849e-05,
      "loss": 0.4347,
      "mean_token_accuracy": 0.8522094587514498,
      "step": 397
    },
    {
      "epoch": 0.5434838269181531,
      "grad_norm": 0.25125772452792333,
      "learning_rate": 1.0238700172530009e-05,
      "loss": 0.4049,
      "mean_token_accuracy": 0.8625896021400488,
      "step": 398
    },
    {
      "epoch": 0.5448493641717163,
      "grad_norm": 0.268673980738851,
      "learning_rate": 1.019096666786931e-05,
      "loss": 0.4203,
      "mean_token_accuracy": 0.8571522385547277,
      "step": 399
    },
    {
      "epoch": 0.5462149014252795,
      "grad_norm": 0.2890895729277909,
      "learning_rate": 1.0143228810047877e-05,
      "loss": 0.4131,
      "mean_token_accuracy": 0.8595778477918344,
      "step": 400
    },
    {
      "epoch": 0.5475804386788428,
      "grad_norm": 0.2542593184479374,
      "learning_rate": 1.0095487687269055e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8549465126076115,
      "step": 401
    },
    {
      "epoch": 0.5489459759324059,
      "grad_norm": 0.2899474295442422,
      "learning_rate": 1.0047744387810632e-05,
      "loss": 0.4303,
      "mean_token_accuracy": 0.8538087426441767,
      "step": 402
    },
    {
      "epoch": 0.5503115131859692,
      "grad_norm": 0.2869992392687126,
      "learning_rate": 1e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8565849096399576,
      "step": 403
    },
    {
      "epoch": 0.5516770504395323,
      "grad_norm": 0.2804072799948344,
      "learning_rate": 9.95225561218937e-06,
      "loss": 0.4108,
      "mean_token_accuracy": 0.8595849788206829,
      "step": 404
    },
    {
      "epoch": 0.5530425876930956,
      "grad_norm": 0.2643969180366311,
      "learning_rate": 9.904512312730948e-06,
      "loss": 0.4137,
      "mean_token_accuracy": 0.8603534103336365,
      "step": 405
    },
    {
      "epoch": 0.5544081249466587,
      "grad_norm": 0.27062716368889733,
      "learning_rate": 9.856771189952127e-06,
      "loss": 0.4015,
      "mean_token_accuracy": 0.8625163495869607,
      "step": 406
    },
    {
      "epoch": 0.555773662200222,
      "grad_norm": 0.2929202457665895,
      "learning_rate": 9.809033332130694e-06,
      "loss": 0.4245,
      "mean_token_accuracy": 0.8562819218107298,
      "step": 407
    },
    {
      "epoch": 0.5571391994537851,
      "grad_norm": 0.2541385240616914,
      "learning_rate": 9.761299827469993e-06,
      "loss": 0.4152,
      "mean_token_accuracy": 0.8585790240492156,
      "step": 408
    },
    {
      "epoch": 0.5585047367073483,
      "grad_norm": 0.24317703441372845,
      "learning_rate": 9.713571764074153e-06,
      "loss": 0.4156,
      "mean_token_accuracy": 0.8592119050098765,
      "step": 409
    },
    {
      "epoch": 0.5598702739609115,
      "grad_norm": 0.2671919366662445,
      "learning_rate": 9.665850229923258e-06,
      "loss": 0.4166,
      "mean_token_accuracy": 0.8583581070445756,
      "step": 410
    },
    {
      "epoch": 0.5612358112144747,
      "grad_norm": 0.2366025343403895,
      "learning_rate": 9.618136312848552e-06,
      "loss": 0.4167,
      "mean_token_accuracy": 0.8573873504387994,
      "step": 411
    },
    {
      "epoch": 0.5626013484680379,
      "grad_norm": 0.2531493471961078,
      "learning_rate": 9.570431100507653e-06,
      "loss": 0.3986,
      "mean_token_accuracy": 0.8631825906862007,
      "step": 412
    },
    {
      "epoch": 0.5639668857216011,
      "grad_norm": 0.2604649082143078,
      "learning_rate": 9.522735680359752e-06,
      "loss": 0.419,
      "mean_token_accuracy": 0.8566154458082337,
      "step": 413
    },
    {
      "epoch": 0.5653324229751643,
      "grad_norm": 0.24389282767116163,
      "learning_rate": 9.47505113964081e-06,
      "loss": 0.4239,
      "mean_token_accuracy": 0.8564176977638592,
      "step": 414
    },
    {
      "epoch": 0.5666979602287275,
      "grad_norm": 0.2838315749775247,
      "learning_rate": 9.4273785653388e-06,
      "loss": 0.4044,
      "mean_token_accuracy": 0.8619406759583335,
      "step": 415
    },
    {
      "epoch": 0.5680634974822907,
      "grad_norm": 0.24792685107778176,
      "learning_rate": 9.379719044168914e-06,
      "loss": 0.4245,
      "mean_token_accuracy": 0.8563493408114575,
      "step": 416
    },
    {
      "epoch": 0.5694290347358539,
      "grad_norm": 0.252380204387106,
      "learning_rate": 9.332073662548785e-06,
      "loss": 0.4192,
      "mean_token_accuracy": 0.8552596982254129,
      "step": 417
    },
    {
      "epoch": 0.5707945719894171,
      "grad_norm": 0.2633944099843147,
      "learning_rate": 9.28444350657374e-06,
      "loss": 0.3925,
      "mean_token_accuracy": 0.8656064042939838,
      "step": 418
    },
    {
      "epoch": 0.5721601092429803,
      "grad_norm": 0.31236207476517847,
      "learning_rate": 9.236829661992024e-06,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8508478194007478,
      "step": 419
    },
    {
      "epoch": 0.5735256464965435,
      "grad_norm": 0.25934793423863595,
      "learning_rate": 9.189233214180057e-06,
      "loss": 0.4029,
      "mean_token_accuracy": 0.8626796404616222,
      "step": 420
    },
    {
      "epoch": 0.5748911837501067,
      "grad_norm": 0.2602440450848244,
      "learning_rate": 9.1416552481177e-06,
      "loss": 0.4011,
      "mean_token_accuracy": 0.8620878476008459,
      "step": 421
    },
    {
      "epoch": 0.5762567210036699,
      "grad_norm": 0.3026174301084034,
      "learning_rate": 9.094096848363503e-06,
      "loss": 0.4132,
      "mean_token_accuracy": 0.8589640787061215,
      "step": 422
    },
    {
      "epoch": 0.5776222582572331,
      "grad_norm": 0.23000512937699333,
      "learning_rate": 9.046559099030012e-06,
      "loss": 0.411,
      "mean_token_accuracy": 0.8593097755828117,
      "step": 423
    },
    {
      "epoch": 0.5789877955107963,
      "grad_norm": 0.28198364889742744,
      "learning_rate": 8.999043083759016e-06,
      "loss": 0.426,
      "mean_token_accuracy": 0.8541148022282637,
      "step": 424
    },
    {
      "epoch": 0.5803533327643595,
      "grad_norm": 0.261123335266971,
      "learning_rate": 8.951549885696889e-06,
      "loss": 0.4008,
      "mean_token_accuracy": 0.863803099113423,
      "step": 425
    },
    {
      "epoch": 0.5817188700179227,
      "grad_norm": 0.2646118229304491,
      "learning_rate": 8.904080587469869e-06,
      "loss": 0.3884,
      "mean_token_accuracy": 0.8656690716827113,
      "step": 426
    },
    {
      "epoch": 0.5830844072714859,
      "grad_norm": 0.23877427883247976,
      "learning_rate": 8.856636271159378e-06,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8549220531025736,
      "step": 427
    },
    {
      "epoch": 0.5844499445250491,
      "grad_norm": 0.27389213494804404,
      "learning_rate": 8.80921801827738e-06,
      "loss": 0.4139,
      "mean_token_accuracy": 0.8583442992618325,
      "step": 428
    },
    {
      "epoch": 0.5858154817786123,
      "grad_norm": 0.3009404820928479,
      "learning_rate": 8.76182690974171e-06,
      "loss": 0.4325,
      "mean_token_accuracy": 0.854122146316857,
      "step": 429
    },
    {
      "epoch": 0.5871810190321755,
      "grad_norm": 0.2723790212614728,
      "learning_rate": 8.714464025851428e-06,
      "loss": 0.4058,
      "mean_token_accuracy": 0.8629916382485093,
      "step": 430
    },
    {
      "epoch": 0.5885465562857387,
      "grad_norm": 0.25219631499358963,
      "learning_rate": 8.667130446262214e-06,
      "loss": 0.4198,
      "mean_token_accuracy": 0.856574897278309,
      "step": 431
    },
    {
      "epoch": 0.5899120935393019,
      "grad_norm": 0.2772798105097925,
      "learning_rate": 8.619827249961732e-06,
      "loss": 0.4047,
      "mean_token_accuracy": 0.8614651709337668,
      "step": 432
    },
    {
      "epoch": 0.5912776307928651,
      "grad_norm": 0.2561843385637697,
      "learning_rate": 8.57255551524506e-06,
      "loss": 0.4058,
      "mean_token_accuracy": 0.8620459681672171,
      "step": 433
    },
    {
      "epoch": 0.5926431680464282,
      "grad_norm": 0.24701843749006636,
      "learning_rate": 8.525316319690092e-06,
      "loss": 0.389,
      "mean_token_accuracy": 0.8674125095943455,
      "step": 434
    },
    {
      "epoch": 0.5940087052999915,
      "grad_norm": 0.2916250794706418,
      "learning_rate": 8.478110740132971e-06,
      "loss": 0.4222,
      "mean_token_accuracy": 0.8566364817285297,
      "step": 435
    },
    {
      "epoch": 0.5953742425535546,
      "grad_norm": 0.2806760751496244,
      "learning_rate": 8.430939852643559e-06,
      "loss": 0.4169,
      "mean_token_accuracy": 0.8584203185837236,
      "step": 436
    },
    {
      "epoch": 0.5967397798071179,
      "grad_norm": 0.27137099893978883,
      "learning_rate": 8.383804732500901e-06,
      "loss": 0.4103,
      "mean_token_accuracy": 0.8591519595394866,
      "step": 437
    },
    {
      "epoch": 0.598105317060681,
      "grad_norm": 0.29739700836059735,
      "learning_rate": 8.336706454168701e-06,
      "loss": 0.4091,
      "mean_token_accuracy": 0.8611807240702429,
      "step": 438
    },
    {
      "epoch": 0.5994708543142443,
      "grad_norm": 0.2571864812668034,
      "learning_rate": 8.289646091270848e-06,
      "loss": 0.4015,
      "mean_token_accuracy": 0.8637355095867093,
      "step": 439
    },
    {
      "epoch": 0.6008363915678074,
      "grad_norm": 0.23339459771292811,
      "learning_rate": 8.242624716566928e-06,
      "loss": 0.3887,
      "mean_token_accuracy": 0.8657256008172974,
      "step": 440
    },
    {
      "epoch": 0.6022019288213707,
      "grad_norm": 0.2992953835452048,
      "learning_rate": 8.195643401927777e-06,
      "loss": 0.3916,
      "mean_token_accuracy": 0.8653041052199545,
      "step": 441
    },
    {
      "epoch": 0.6035674660749338,
      "grad_norm": 0.2585326431716542,
      "learning_rate": 8.148703218311053e-06,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8596556131686985,
      "step": 442
    },
    {
      "epoch": 0.6049330033284971,
      "grad_norm": 0.2503365003487542,
      "learning_rate": 8.101805235736804e-06,
      "loss": 0.4265,
      "mean_token_accuracy": 0.8549398314920128,
      "step": 443
    },
    {
      "epoch": 0.6062985405820602,
      "grad_norm": 0.2820273966033343,
      "learning_rate": 8.054950523263097e-06,
      "loss": 0.4136,
      "mean_token_accuracy": 0.8588294366875941,
      "step": 444
    },
    {
      "epoch": 0.6076640778356235,
      "grad_norm": 0.25413507701062166,
      "learning_rate": 8.008140148961642e-06,
      "loss": 0.4044,
      "mean_token_accuracy": 0.8625195591348364,
      "step": 445
    },
    {
      "epoch": 0.6090296150891866,
      "grad_norm": 0.23238499595613277,
      "learning_rate": 7.96137517989343e-06,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8561338059337695,
      "step": 446
    },
    {
      "epoch": 0.6103951523427499,
      "grad_norm": 0.23798509908211196,
      "learning_rate": 7.914656682084436e-06,
      "loss": 0.4169,
      "mean_token_accuracy": 0.8577032958934101,
      "step": 447
    },
    {
      "epoch": 0.611760689596313,
      "grad_norm": 0.25296096161140275,
      "learning_rate": 7.867985720501301e-06,
      "loss": 0.399,
      "mean_token_accuracy": 0.863544434056401,
      "step": 448
    },
    {
      "epoch": 0.6131262268498763,
      "grad_norm": 0.22798661470070672,
      "learning_rate": 7.821363359027047e-06,
      "loss": 0.4028,
      "mean_token_accuracy": 0.8615431059914728,
      "step": 449
    },
    {
      "epoch": 0.6144917641034394,
      "grad_norm": 0.2405890996462141,
      "learning_rate": 7.774790660436857e-06,
      "loss": 0.4279,
      "mean_token_accuracy": 0.8547220263658766,
      "step": 450
    },
    {
      "epoch": 0.6158573013570027,
      "grad_norm": 0.29357731632515965,
      "learning_rate": 7.728268686373814e-06,
      "loss": 0.4077,
      "mean_token_accuracy": 0.8615052330629221,
      "step": 451
    },
    {
      "epoch": 0.6172228386105658,
      "grad_norm": 0.28038212690179937,
      "learning_rate": 7.681798497324717e-06,
      "loss": 0.4379,
      "mean_token_accuracy": 0.8519818024977281,
      "step": 452
    },
    {
      "epoch": 0.6185883758641291,
      "grad_norm": 0.25548937081717665,
      "learning_rate": 7.635381152595916e-06,
      "loss": 0.4175,
      "mean_token_accuracy": 0.857195676423713,
      "step": 453
    },
    {
      "epoch": 0.6199539131176922,
      "grad_norm": 0.25398512091441905,
      "learning_rate": 7.5890177102891395e-06,
      "loss": 0.4039,
      "mean_token_accuracy": 0.8617371650977873,
      "step": 454
    },
    {
      "epoch": 0.6213194503712555,
      "grad_norm": 0.26543560160540447,
      "learning_rate": 7.542709227277396e-06,
      "loss": 0.4164,
      "mean_token_accuracy": 0.857181019142036,
      "step": 455
    },
    {
      "epoch": 0.6226849876248186,
      "grad_norm": 0.2459878250772752,
      "learning_rate": 7.496456759180876e-06,
      "loss": 0.4228,
      "mean_token_accuracy": 0.8573935855915926,
      "step": 456
    },
    {
      "epoch": 0.6240505248783819,
      "grad_norm": 0.23786580710281324,
      "learning_rate": 7.4502613603428875e-06,
      "loss": 0.4088,
      "mean_token_accuracy": 0.8607801487917315,
      "step": 457
    },
    {
      "epoch": 0.625416062131945,
      "grad_norm": 0.29589767964465397,
      "learning_rate": 7.404124083805819e-06,
      "loss": 0.4093,
      "mean_token_accuracy": 0.8612231610118724,
      "step": 458
    },
    {
      "epoch": 0.6267815993855083,
      "grad_norm": 0.2887590659400021,
      "learning_rate": 7.358045981287141e-06,
      "loss": 0.4209,
      "mean_token_accuracy": 0.8560159432518372,
      "step": 459
    },
    {
      "epoch": 0.6281471366390714,
      "grad_norm": 0.21799038062254464,
      "learning_rate": 7.312028103155426e-06,
      "loss": 0.3799,
      "mean_token_accuracy": 0.8704455922016139,
      "step": 460
    },
    {
      "epoch": 0.6295126738926347,
      "grad_norm": 0.24324879715514983,
      "learning_rate": 7.266071498406417e-06,
      "loss": 0.4076,
      "mean_token_accuracy": 0.8610840750554581,
      "step": 461
    },
    {
      "epoch": 0.6308782111461978,
      "grad_norm": 0.25428302492825916,
      "learning_rate": 7.220177214639088e-06,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8587935610875843,
      "step": 462
    },
    {
      "epoch": 0.632243748399761,
      "grad_norm": 0.27202575353290637,
      "learning_rate": 7.1743462980318045e-06,
      "loss": 0.4164,
      "mean_token_accuracy": 0.8589937783595583,
      "step": 463
    },
    {
      "epoch": 0.6336092856533242,
      "grad_norm": 0.23695566303733537,
      "learning_rate": 7.128579793318429e-06,
      "loss": 0.4024,
      "mean_token_accuracy": 0.8627954415233714,
      "step": 464
    },
    {
      "epoch": 0.6349748229068874,
      "grad_norm": 0.25513688392749734,
      "learning_rate": 7.0828787437645455e-06,
      "loss": 0.4093,
      "mean_token_accuracy": 0.8600665306531674,
      "step": 465
    },
    {
      "epoch": 0.6363403601604506,
      "grad_norm": 0.26912983946640745,
      "learning_rate": 7.037244191143662e-06,
      "loss": 0.3995,
      "mean_token_accuracy": 0.8619904346664145,
      "step": 466
    },
    {
      "epoch": 0.6377058974140138,
      "grad_norm": 0.2525033514112838,
      "learning_rate": 6.991677175713449e-06,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8571298722103126,
      "step": 467
    },
    {
      "epoch": 0.639071434667577,
      "grad_norm": 0.4126261428152561,
      "learning_rate": 6.946178736192053e-06,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8495778270375115,
      "step": 468
    },
    {
      "epoch": 0.6404369719211402,
      "grad_norm": 0.25938255418921724,
      "learning_rate": 6.900749909734406e-06,
      "loss": 0.433,
      "mean_token_accuracy": 0.8528369297456597,
      "step": 469
    },
    {
      "epoch": 0.6418025091747034,
      "grad_norm": 0.2505229138650988,
      "learning_rate": 6.8553917319085676e-06,
      "loss": 0.4148,
      "mean_token_accuracy": 0.8588444157084706,
      "step": 470
    },
    {
      "epoch": 0.6431680464282666,
      "grad_norm": 0.21310658415353784,
      "learning_rate": 6.810105236672155e-06,
      "loss": 0.3849,
      "mean_token_accuracy": 0.8671909034888359,
      "step": 471
    },
    {
      "epoch": 0.6445335836818298,
      "grad_norm": 0.23204325202817497,
      "learning_rate": 6.76489145634873e-06,
      "loss": 0.4074,
      "mean_token_accuracy": 0.8617341122845665,
      "step": 472
    },
    {
      "epoch": 0.645899120935393,
      "grad_norm": 0.27377537704030214,
      "learning_rate": 6.719751421604309e-06,
      "loss": 0.4052,
      "mean_token_accuracy": 0.8605801437822177,
      "step": 473
    },
    {
      "epoch": 0.6472646581889562,
      "grad_norm": 0.26667506542190905,
      "learning_rate": 6.6746861614238425e-06,
      "loss": 0.4172,
      "mean_token_accuracy": 0.8587473642719939,
      "step": 474
    },
    {
      "epoch": 0.6486301954425194,
      "grad_norm": 0.2359219117164129,
      "learning_rate": 6.629696703087755e-06,
      "loss": 0.3979,
      "mean_token_accuracy": 0.8643947766486794,
      "step": 475
    },
    {
      "epoch": 0.6499957326960826,
      "grad_norm": 0.2405700433141458,
      "learning_rate": 6.584784072148554e-06,
      "loss": 0.4117,
      "mean_token_accuracy": 0.8583265490823113,
      "step": 476
    },
    {
      "epoch": 0.6513612699496458,
      "grad_norm": 0.23752686356493705,
      "learning_rate": 6.5399492924074215e-06,
      "loss": 0.4079,
      "mean_token_accuracy": 0.8599794679393602,
      "step": 477
    },
    {
      "epoch": 0.652726807203209,
      "grad_norm": 0.260987271653807,
      "learning_rate": 6.495193385890901e-06,
      "loss": 0.3978,
      "mean_token_accuracy": 0.863796731863899,
      "step": 478
    },
    {
      "epoch": 0.6540923444567722,
      "grad_norm": 0.2389148131609978,
      "learning_rate": 6.450517372827591e-06,
      "loss": 0.4048,
      "mean_token_accuracy": 0.8625655846917668,
      "step": 479
    },
    {
      "epoch": 0.6554578817103354,
      "grad_norm": 0.23754616663334344,
      "learning_rate": 6.405922271624874e-06,
      "loss": 0.3894,
      "mean_token_accuracy": 0.8650371535045855,
      "step": 480
    },
    {
      "epoch": 0.6568234189638986,
      "grad_norm": 0.24392056114704547,
      "learning_rate": 6.3614090988457255e-06,
      "loss": 0.4078,
      "mean_token_accuracy": 0.860617293319343,
      "step": 481
    },
    {
      "epoch": 0.6581889562174618,
      "grad_norm": 0.26180580733734304,
      "learning_rate": 6.3169788691855326e-06,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8568173015759246,
      "step": 482
    },
    {
      "epoch": 0.659554493471025,
      "grad_norm": 0.24631836835717166,
      "learning_rate": 6.2726325954489474e-06,
      "loss": 0.4214,
      "mean_token_accuracy": 0.8552957608827673,
      "step": 483
    },
    {
      "epoch": 0.6609200307245882,
      "grad_norm": 0.2805589427499748,
      "learning_rate": 6.22837128852683e-06,
      "loss": 0.4209,
      "mean_token_accuracy": 0.857800855823934,
      "step": 484
    },
    {
      "epoch": 0.6622855679781514,
      "grad_norm": 0.232371020464702,
      "learning_rate": 6.184195957373176e-06,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8600986670212696,
      "step": 485
    },
    {
      "epoch": 0.6636511052317146,
      "grad_norm": 0.24474802604092138,
      "learning_rate": 6.140107608982137e-06,
      "loss": 0.3952,
      "mean_token_accuracy": 0.8657601107656259,
      "step": 486
    },
    {
      "epoch": 0.6650166424852778,
      "grad_norm": 0.25305102083356706,
      "learning_rate": 6.0961072483650526e-06,
      "loss": 0.4037,
      "mean_token_accuracy": 0.8629620469736944,
      "step": 487
    },
    {
      "epoch": 0.666382179738841,
      "grad_norm": 0.22975337882908747,
      "learning_rate": 6.052195878527551e-06,
      "loss": 0.4038,
      "mean_token_accuracy": 0.8623166696689788,
      "step": 488
    },
    {
      "epoch": 0.6677477169924042,
      "grad_norm": 0.24494072522854304,
      "learning_rate": 6.008374500446676e-06,
      "loss": 0.4203,
      "mean_token_accuracy": 0.8575659022852821,
      "step": 489
    },
    {
      "epoch": 0.6691132542459673,
      "grad_norm": 0.24564454125360483,
      "learning_rate": 5.964644113048079e-06,
      "loss": 0.4108,
      "mean_token_accuracy": 0.8592899012026329,
      "step": 490
    },
    {
      "epoch": 0.6704787914995306,
      "grad_norm": 0.2970193563961585,
      "learning_rate": 5.921005713183236e-06,
      "loss": 0.414,
      "mean_token_accuracy": 0.8578038401374278,
      "step": 491
    },
    {
      "epoch": 0.6718443287530937,
      "grad_norm": 0.22522151812319466,
      "learning_rate": 5.877460295606739e-06,
      "loss": 0.3909,
      "mean_token_accuracy": 0.8663197254697791,
      "step": 492
    },
    {
      "epoch": 0.673209866006657,
      "grad_norm": 0.24777229494666286,
      "learning_rate": 5.834008852953603e-06,
      "loss": 0.3987,
      "mean_token_accuracy": 0.8633314691916142,
      "step": 493
    },
    {
      "epoch": 0.6745754032602201,
      "grad_norm": 0.27546512967034037,
      "learning_rate": 5.790652375716653e-06,
      "loss": 0.3887,
      "mean_token_accuracy": 0.8660309352349229,
      "step": 494
    },
    {
      "epoch": 0.6759409405137834,
      "grad_norm": 0.24981866455290716,
      "learning_rate": 5.74739185222394e-06,
      "loss": 0.4182,
      "mean_token_accuracy": 0.8580747989016917,
      "step": 495
    },
    {
      "epoch": 0.6773064777673465,
      "grad_norm": 0.23497798723811336,
      "learning_rate": 5.704228268616208e-06,
      "loss": 0.4034,
      "mean_token_accuracy": 0.8624885421549002,
      "step": 496
    },
    {
      "epoch": 0.6786720150209098,
      "grad_norm": 0.22626842482811318,
      "learning_rate": 5.66116260882442e-06,
      "loss": 0.3743,
      "mean_token_accuracy": 0.8722626987396589,
      "step": 497
    },
    {
      "epoch": 0.6800375522744729,
      "grad_norm": 0.273018811101812,
      "learning_rate": 5.618195854547333e-06,
      "loss": 0.4118,
      "mean_token_accuracy": 0.8591764611730132,
      "step": 498
    },
    {
      "epoch": 0.6814030895280362,
      "grad_norm": 0.2517060577155918,
      "learning_rate": 5.575328985229098e-06,
      "loss": 0.4307,
      "mean_token_accuracy": 0.8540759587427639,
      "step": 499
    },
    {
      "epoch": 0.6827686267815993,
      "grad_norm": 0.23433321670856827,
      "learning_rate": 5.532562978036964e-06,
      "loss": 0.3938,
      "mean_token_accuracy": 0.8640612399037781,
      "step": 500
    },
    {
      "epoch": 0.6841341640351626,
      "grad_norm": 0.2425154734570016,
      "learning_rate": 5.48989880783898e-06,
      "loss": 0.422,
      "mean_token_accuracy": 0.8574342653414885,
      "step": 501
    },
    {
      "epoch": 0.6854997012887258,
      "grad_norm": 0.23347937555277293,
      "learning_rate": 5.4473374471817906e-06,
      "loss": 0.4191,
      "mean_token_accuracy": 0.8575295591207817,
      "step": 502
    },
    {
      "epoch": 0.686865238542289,
      "grad_norm": 0.2545433711957671,
      "learning_rate": 5.404879866268438e-06,
      "loss": 0.4002,
      "mean_token_accuracy": 0.8639634173808828,
      "step": 503
    },
    {
      "epoch": 0.6882307757958522,
      "grad_norm": 0.24243974966286372,
      "learning_rate": 5.362527032936278e-06,
      "loss": 0.4028,
      "mean_token_accuracy": 0.862244193071149,
      "step": 504
    },
    {
      "epoch": 0.6895963130494154,
      "grad_norm": 0.23059100155723,
      "learning_rate": 5.320279912634907e-06,
      "loss": 0.4035,
      "mean_token_accuracy": 0.8626255651551141,
      "step": 505
    },
    {
      "epoch": 0.6909618503029786,
      "grad_norm": 0.23428705595664845,
      "learning_rate": 5.278139468404133e-06,
      "loss": 0.4083,
      "mean_token_accuracy": 0.8603081045639313,
      "step": 506
    },
    {
      "epoch": 0.6923273875565418,
      "grad_norm": 0.27274705668530197,
      "learning_rate": 5.236106660852058e-06,
      "loss": 0.4212,
      "mean_token_accuracy": 0.8560662849082573,
      "step": 507
    },
    {
      "epoch": 0.693692924810105,
      "grad_norm": 0.2845641889872319,
      "learning_rate": 5.194182448133163e-06,
      "loss": 0.3944,
      "mean_token_accuracy": 0.8640756812761311,
      "step": 508
    },
    {
      "epoch": 0.6950584620636682,
      "grad_norm": 0.22664565522900781,
      "learning_rate": 5.152367785926452e-06,
      "loss": 0.4094,
      "mean_token_accuracy": 0.8593605320038422,
      "step": 509
    },
    {
      "epoch": 0.6964239993172314,
      "grad_norm": 0.22478587370218284,
      "learning_rate": 5.110663627413695e-06,
      "loss": 0.396,
      "mean_token_accuracy": 0.8648808615935969,
      "step": 510
    },
    {
      "epoch": 0.6977895365707946,
      "grad_norm": 0.21591948807401068,
      "learning_rate": 5.069070923257685e-06,
      "loss": 0.399,
      "mean_token_accuracy": 0.8633928472317149,
      "step": 511
    },
    {
      "epoch": 0.6991550738243578,
      "grad_norm": 0.26461022257675626,
      "learning_rate": 5.027590621580563e-06,
      "loss": 0.4394,
      "mean_token_accuracy": 0.8507808502833459,
      "step": 512
    },
    {
      "epoch": 0.700520611077921,
      "grad_norm": 0.21356588082748598,
      "learning_rate": 4.986223667942213e-06,
      "loss": 0.4271,
      "mean_token_accuracy": 0.8533322294161114,
      "step": 513
    },
    {
      "epoch": 0.7018861483314842,
      "grad_norm": 0.21403249722142711,
      "learning_rate": 4.944971005318716e-06,
      "loss": 0.3987,
      "mean_token_accuracy": 0.8636949003001929,
      "step": 514
    },
    {
      "epoch": 0.7032516855850474,
      "grad_norm": 0.23897074188676798,
      "learning_rate": 4.903833574080825e-06,
      "loss": 0.3925,
      "mean_token_accuracy": 0.8641888082172495,
      "step": 515
    },
    {
      "epoch": 0.7046172228386106,
      "grad_norm": 0.22004960179660596,
      "learning_rate": 4.862812311972567e-06,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8554980900000537,
      "step": 516
    },
    {
      "epoch": 0.7059827600921738,
      "grad_norm": 0.21948469336679155,
      "learning_rate": 4.82190815408983e-06,
      "loss": 0.4033,
      "mean_token_accuracy": 0.8607597379326859,
      "step": 517
    },
    {
      "epoch": 0.707348297345737,
      "grad_norm": 0.23185408625161705,
      "learning_rate": 4.781122032859079e-06,
      "loss": 0.4137,
      "mean_token_accuracy": 0.8586254422034981,
      "step": 518
    },
    {
      "epoch": 0.7087138345993002,
      "grad_norm": 0.23176435351640776,
      "learning_rate": 4.740454878016084e-06,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8592659758915503,
      "step": 519
    },
    {
      "epoch": 0.7100793718528634,
      "grad_norm": 0.24913384906741762,
      "learning_rate": 4.6999076165847214e-06,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8513752575363313,
      "step": 520
    },
    {
      "epoch": 0.7114449091064265,
      "grad_norm": 0.20722173613037567,
      "learning_rate": 4.659481172855859e-06,
      "loss": 0.4005,
      "mean_token_accuracy": 0.8633166282089912,
      "step": 521
    },
    {
      "epoch": 0.7128104463599898,
      "grad_norm": 0.2181641440286907,
      "learning_rate": 4.619176468366274e-06,
      "loss": 0.3917,
      "mean_token_accuracy": 0.8660028930864674,
      "step": 522
    },
    {
      "epoch": 0.714175983613553,
      "grad_norm": 0.25367503570810757,
      "learning_rate": 4.578994421877645e-06,
      "loss": 0.4096,
      "mean_token_accuracy": 0.8606353223418006,
      "step": 523
    },
    {
      "epoch": 0.7155415208671162,
      "grad_norm": 0.21964433917606951,
      "learning_rate": 4.538935949355623e-06,
      "loss": 0.4115,
      "mean_token_accuracy": 0.8579058023067518,
      "step": 524
    },
    {
      "epoch": 0.7169070581206793,
      "grad_norm": 0.23194379718009556,
      "learning_rate": 4.499001963948929e-06,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8567179649132983,
      "step": 525
    },
    {
      "epoch": 0.7182725953742426,
      "grad_norm": 0.23794903701170841,
      "learning_rate": 4.45919337596856e-06,
      "loss": 0.4044,
      "mean_token_accuracy": 0.8625243196220369,
      "step": 526
    },
    {
      "epoch": 0.7196381326278057,
      "grad_norm": 0.24368825515761053,
      "learning_rate": 4.41951109286703e-06,
      "loss": 0.3946,
      "mean_token_accuracy": 0.8651203225887669,
      "step": 527
    },
    {
      "epoch": 0.721003669881369,
      "grad_norm": 0.2753308025130241,
      "learning_rate": 4.379956019217675e-06,
      "loss": 0.415,
      "mean_token_accuracy": 0.8604877721998891,
      "step": 528
    },
    {
      "epoch": 0.7223692071349321,
      "grad_norm": 0.21688742696840135,
      "learning_rate": 4.3405290566940475e-06,
      "loss": 0.397,
      "mean_token_accuracy": 0.8641772641669326,
      "step": 529
    },
    {
      "epoch": 0.7237347443884954,
      "grad_norm": 0.21378940665861854,
      "learning_rate": 4.301231104049359e-06,
      "loss": 0.4174,
      "mean_token_accuracy": 0.8589884247304818,
      "step": 530
    },
    {
      "epoch": 0.7251002816420585,
      "grad_norm": 0.2014478430774182,
      "learning_rate": 4.262063057095978e-06,
      "loss": 0.3961,
      "mean_token_accuracy": 0.864822926034779,
      "step": 531
    },
    {
      "epoch": 0.7264658188956218,
      "grad_norm": 0.21961573395687345,
      "learning_rate": 4.2230258086850375e-06,
      "loss": 0.4018,
      "mean_token_accuracy": 0.8609590482386245,
      "step": 532
    },
    {
      "epoch": 0.7278313561491849,
      "grad_norm": 0.22346096554308978,
      "learning_rate": 4.184120248686048e-06,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8551864724213167,
      "step": 533
    },
    {
      "epoch": 0.7291968934027482,
      "grad_norm": 0.2324113039004406,
      "learning_rate": 4.145347263966646e-06,
      "loss": 0.4106,
      "mean_token_accuracy": 0.8610600169630592,
      "step": 534
    },
    {
      "epoch": 0.7305624306563113,
      "grad_norm": 0.21693513782961235,
      "learning_rate": 4.106707738372357e-06,
      "loss": 0.4047,
      "mean_token_accuracy": 0.8620982081151527,
      "step": 535
    },
    {
      "epoch": 0.7319279679098746,
      "grad_norm": 0.24610715663770288,
      "learning_rate": 4.0682025527064486e-06,
      "loss": 0.4142,
      "mean_token_accuracy": 0.8586329074626934,
      "step": 536
    },
    {
      "epoch": 0.7332935051634377,
      "grad_norm": 0.2496537037626066,
      "learning_rate": 4.029832584709864e-06,
      "loss": 0.4133,
      "mean_token_accuracy": 0.8591482955360918,
      "step": 537
    },
    {
      "epoch": 0.734659042417001,
      "grad_norm": 0.25290330143622763,
      "learning_rate": 3.991598709041196e-06,
      "loss": 0.4185,
      "mean_token_accuracy": 0.8575748615139153,
      "step": 538
    },
    {
      "epoch": 0.7360245796705641,
      "grad_norm": 0.23375530633596542,
      "learning_rate": 3.953501797256768e-06,
      "loss": 0.4082,
      "mean_token_accuracy": 0.8608288259179717,
      "step": 539
    },
    {
      "epoch": 0.7373901169241274,
      "grad_norm": 0.23750273636894062,
      "learning_rate": 3.915542717790759e-06,
      "loss": 0.4087,
      "mean_token_accuracy": 0.8604965692097658,
      "step": 540
    },
    {
      "epoch": 0.7387556541776905,
      "grad_norm": 0.22536670381141075,
      "learning_rate": 3.877722335935394e-06,
      "loss": 0.4274,
      "mean_token_accuracy": 0.8535887513090156,
      "step": 541
    },
    {
      "epoch": 0.7401211914312538,
      "grad_norm": 0.2337249133077971,
      "learning_rate": 3.840041513821243e-06,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8543320104591253,
      "step": 542
    },
    {
      "epoch": 0.7414867286848169,
      "grad_norm": 0.23006457934706245,
      "learning_rate": 3.802501110397553e-06,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8595805110377353,
      "step": 543
    },
    {
      "epoch": 0.7428522659383802,
      "grad_norm": 0.2264725327560074,
      "learning_rate": 3.7651019814126656e-06,
      "loss": 0.3978,
      "mean_token_accuracy": 0.8640107757543986,
      "step": 544
    },
    {
      "epoch": 0.7442178031919433,
      "grad_norm": 0.21662115139126908,
      "learning_rate": 3.727844979394526e-06,
      "loss": 0.4171,
      "mean_token_accuracy": 0.8583177721469002,
      "step": 545
    },
    {
      "epoch": 0.7455833404455066,
      "grad_norm": 0.24171854514556812,
      "learning_rate": 3.6907309536312276e-06,
      "loss": 0.4157,
      "mean_token_accuracy": 0.8568488785351885,
      "step": 546
    },
    {
      "epoch": 0.7469488776990697,
      "grad_norm": 0.20105071964248444,
      "learning_rate": 3.6537607501516716e-06,
      "loss": 0.3969,
      "mean_token_accuracy": 0.8654952584767601,
      "step": 547
    },
    {
      "epoch": 0.748314414952633,
      "grad_norm": 0.22792825494280627,
      "learning_rate": 3.616935211706275e-06,
      "loss": 0.3932,
      "mean_token_accuracy": 0.8653115392045422,
      "step": 548
    },
    {
      "epoch": 0.7496799522061961,
      "grad_norm": 0.2314969835671945,
      "learning_rate": 3.5802551777477477e-06,
      "loss": 0.3936,
      "mean_token_accuracy": 0.8653473275464092,
      "step": 549
    },
    {
      "epoch": 0.7510454894597594,
      "grad_norm": 0.21855230450778143,
      "learning_rate": 3.543721484411976e-06,
      "loss": 0.4186,
      "mean_token_accuracy": 0.8572386847718454,
      "step": 550
    },
    {
      "epoch": 0.7524110267133225,
      "grad_norm": 0.2179108335195832,
      "learning_rate": 3.5073349644989563e-06,
      "loss": 0.3948,
      "mean_token_accuracy": 0.8645206681557961,
      "step": 551
    },
    {
      "epoch": 0.7537765639668857,
      "grad_norm": 0.2216215829360909,
      "learning_rate": 3.4710964474537967e-06,
      "loss": 0.3988,
      "mean_token_accuracy": 0.8638077505445968,
      "step": 552
    },
    {
      "epoch": 0.7551421012204489,
      "grad_norm": 0.2693768802921377,
      "learning_rate": 3.435006759347835e-06,
      "loss": 0.4142,
      "mean_token_accuracy": 0.8578324828930424,
      "step": 553
    },
    {
      "epoch": 0.7565076384740121,
      "grad_norm": 0.2226472320558841,
      "learning_rate": 3.3990667228597816e-06,
      "loss": 0.4028,
      "mean_token_accuracy": 0.8616278351864007,
      "step": 554
    },
    {
      "epoch": 0.7578731757275753,
      "grad_norm": 0.23046863738461842,
      "learning_rate": 3.3632771572569878e-06,
      "loss": 0.4096,
      "mean_token_accuracy": 0.858772543145848,
      "step": 555
    },
    {
      "epoch": 0.7592387129811385,
      "grad_norm": 0.21268282367594163,
      "learning_rate": 3.3276388783767644e-06,
      "loss": 0.4126,
      "mean_token_accuracy": 0.8590454775080085,
      "step": 556
    },
    {
      "epoch": 0.7606042502347017,
      "grad_norm": 0.2258472656546357,
      "learning_rate": 3.292152698607768e-06,
      "loss": 0.3868,
      "mean_token_accuracy": 0.8656557706456838,
      "step": 557
    },
    {
      "epoch": 0.7619697874882649,
      "grad_norm": 0.21882486076601618,
      "learning_rate": 3.256819426871507e-06,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8588001715592681,
      "step": 558
    },
    {
      "epoch": 0.7633353247418281,
      "grad_norm": 0.20092513277692103,
      "learning_rate": 3.221639868603893e-06,
      "loss": 0.4017,
      "mean_token_accuracy": 0.8617123529703908,
      "step": 559
    },
    {
      "epoch": 0.7647008619953913,
      "grad_norm": 0.2121347752460375,
      "learning_rate": 3.1866148257368666e-06,
      "loss": 0.3975,
      "mean_token_accuracy": 0.8637410178015515,
      "step": 560
    },
    {
      "epoch": 0.7660663992489545,
      "grad_norm": 0.22278871306746495,
      "learning_rate": 3.15174509668014e-06,
      "loss": 0.4025,
      "mean_token_accuracy": 0.8612379444311294,
      "step": 561
    },
    {
      "epoch": 0.7674319365025177,
      "grad_norm": 0.23213748039939272,
      "learning_rate": 3.117031476302975e-06,
      "loss": 0.4055,
      "mean_token_accuracy": 0.8602274083124468,
      "step": 562
    },
    {
      "epoch": 0.7687974737560809,
      "grad_norm": 0.2254901351758214,
      "learning_rate": 3.082474755916084e-06,
      "loss": 0.4075,
      "mean_token_accuracy": 0.8596426308589397,
      "step": 563
    },
    {
      "epoch": 0.7701630110096441,
      "grad_norm": 0.2005034006194305,
      "learning_rate": 3.0480757232535773e-06,
      "loss": 0.3981,
      "mean_token_accuracy": 0.8630627269518893,
      "step": 564
    },
    {
      "epoch": 0.7715285482632073,
      "grad_norm": 0.19345269846448648,
      "learning_rate": 3.0138351624550165e-06,
      "loss": 0.4049,
      "mean_token_accuracy": 0.8611821074357777,
      "step": 565
    },
    {
      "epoch": 0.7728940855167705,
      "grad_norm": 0.20062995179443605,
      "learning_rate": 2.9797538540475223e-06,
      "loss": 0.3988,
      "mean_token_accuracy": 0.8634990578142336,
      "step": 566
    },
    {
      "epoch": 0.7742596227703337,
      "grad_norm": 0.23164980231281976,
      "learning_rate": 2.945832574928006e-06,
      "loss": 0.4215,
      "mean_token_accuracy": 0.8557358468375269,
      "step": 567
    },
    {
      "epoch": 0.7756251600238969,
      "grad_norm": 0.20342056199797945,
      "learning_rate": 2.9120720983454465e-06,
      "loss": 0.4121,
      "mean_token_accuracy": 0.8570414758839693,
      "step": 568
    },
    {
      "epoch": 0.7769906972774601,
      "grad_norm": 0.22910941122839265,
      "learning_rate": 2.8784731938832556e-06,
      "loss": 0.4087,
      "mean_token_accuracy": 0.8596079874910577,
      "step": 569
    },
    {
      "epoch": 0.7783562345310233,
      "grad_norm": 0.21416157641581723,
      "learning_rate": 2.845036627441755e-06,
      "loss": 0.3833,
      "mean_token_accuracy": 0.8691450521802337,
      "step": 570
    },
    {
      "epoch": 0.7797217717845865,
      "grad_norm": 0.1977219313107081,
      "learning_rate": 2.8117631612207084e-06,
      "loss": 0.4121,
      "mean_token_accuracy": 0.8610763435801025,
      "step": 571
    },
    {
      "epoch": 0.7810873090381497,
      "grad_norm": 0.20321588988831457,
      "learning_rate": 2.778653553701932e-06,
      "loss": 0.4057,
      "mean_token_accuracy": 0.8604487017005163,
      "step": 572
    },
    {
      "epoch": 0.7824528462917129,
      "grad_norm": 0.21685215914126496,
      "learning_rate": 2.745708559632032e-06,
      "loss": 0.3965,
      "mean_token_accuracy": 0.8631075593834338,
      "step": 573
    },
    {
      "epoch": 0.7838183835452761,
      "grad_norm": 0.21247170342394922,
      "learning_rate": 2.7129289300051788e-06,
      "loss": 0.4066,
      "mean_token_accuracy": 0.8607861090034046,
      "step": 574
    },
    {
      "epoch": 0.7851839207988393,
      "grad_norm": 0.22155579546490167,
      "learning_rate": 2.6803154120460007e-06,
      "loss": 0.4072,
      "mean_token_accuracy": 0.8621495738744454,
      "step": 575
    },
    {
      "epoch": 0.7865494580524025,
      "grad_norm": 0.20936523040531713,
      "learning_rate": 2.647868749192536e-06,
      "loss": 0.4187,
      "mean_token_accuracy": 0.8561885820547165,
      "step": 576
    },
    {
      "epoch": 0.7879149953059656,
      "grad_norm": 0.2169635727623937,
      "learning_rate": 2.6155896810793036e-06,
      "loss": 0.4118,
      "mean_token_accuracy": 0.8589148130153295,
      "step": 577
    },
    {
      "epoch": 0.7892805325595289,
      "grad_norm": 0.20681767217525526,
      "learning_rate": 2.5834789435204245e-06,
      "loss": 0.4088,
      "mean_token_accuracy": 0.8606342267092991,
      "step": 578
    },
    {
      "epoch": 0.790646069813092,
      "grad_norm": 0.2213528976805314,
      "learning_rate": 2.5515372684928687e-06,
      "loss": 0.4222,
      "mean_token_accuracy": 0.8550210063940413,
      "step": 579
    },
    {
      "epoch": 0.7920116070666553,
      "grad_norm": 0.22149304881479331,
      "learning_rate": 2.5197653841197546e-06,
      "loss": 0.4114,
      "mean_token_accuracy": 0.8591611937568427,
      "step": 580
    },
    {
      "epoch": 0.7933771443202184,
      "grad_norm": 0.19889209166062613,
      "learning_rate": 2.48816401465375e-06,
      "loss": 0.4057,
      "mean_token_accuracy": 0.861537611573371,
      "step": 581
    },
    {
      "epoch": 0.7947426815737817,
      "grad_norm": 0.24049345793108362,
      "learning_rate": 2.4567338804605756e-06,
      "loss": 0.4,
      "mean_token_accuracy": 0.86237969255477,
      "step": 582
    },
    {
      "epoch": 0.7961082188273448,
      "grad_norm": 0.23885430272558048,
      "learning_rate": 2.425475698002577e-06,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8575069291832286,
      "step": 583
    },
    {
      "epoch": 0.7974737560809081,
      "grad_norm": 0.22824330369698562,
      "learning_rate": 2.394390179822382e-06,
      "loss": 0.4182,
      "mean_token_accuracy": 0.8576008710768586,
      "step": 584
    },
    {
      "epoch": 0.7988392933344712,
      "grad_norm": 0.21974915679821652,
      "learning_rate": 2.3634780345266805e-06,
      "loss": 0.3963,
      "mean_token_accuracy": 0.8647796993722388,
      "step": 585
    },
    {
      "epoch": 0.8002048305880345,
      "grad_norm": 0.2310148673403252,
      "learning_rate": 2.332739966770048e-06,
      "loss": 0.4005,
      "mean_token_accuracy": 0.8606528012118717,
      "step": 586
    },
    {
      "epoch": 0.8015703678415976,
      "grad_norm": 0.21268357727111933,
      "learning_rate": 2.3021766772388986e-06,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8596851429925462,
      "step": 587
    },
    {
      "epoch": 0.8029359050951609,
      "grad_norm": 0.22436026449039995,
      "learning_rate": 2.271788862635513e-06,
      "loss": 0.415,
      "mean_token_accuracy": 0.8579554313395271,
      "step": 588
    },
    {
      "epoch": 0.804301442348724,
      "grad_norm": 0.2401381634931562,
      "learning_rate": 2.2415772156621387e-06,
      "loss": 0.4281,
      "mean_token_accuracy": 0.8525374331069293,
      "step": 589
    },
    {
      "epoch": 0.8056669796022873,
      "grad_norm": 0.23642180241330024,
      "learning_rate": 2.211542425005223e-06,
      "loss": 0.427,
      "mean_token_accuracy": 0.8546586111147729,
      "step": 590
    },
    {
      "epoch": 0.8070325168558504,
      "grad_norm": 0.2115384062221628,
      "learning_rate": 2.1816851753197023e-06,
      "loss": 0.4114,
      "mean_token_accuracy": 0.8595850270091632,
      "step": 591
    },
    {
      "epoch": 0.8083980541094137,
      "grad_norm": 0.21400197301419835,
      "learning_rate": 2.1520061472133903e-06,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8604726451795677,
      "step": 592
    },
    {
      "epoch": 0.8097635913629768,
      "grad_norm": 0.2017117241684196,
      "learning_rate": 2.1225060172314773e-06,
      "loss": 0.3995,
      "mean_token_accuracy": 0.8632549469473196,
      "step": 593
    },
    {
      "epoch": 0.8111291286165401,
      "grad_norm": 0.19894011888129304,
      "learning_rate": 2.0931854578410904e-06,
      "loss": 0.3991,
      "mean_token_accuracy": 0.8642028385916231,
      "step": 594
    },
    {
      "epoch": 0.8124946658701032,
      "grad_norm": 0.2298631393329404,
      "learning_rate": 2.064045137415982e-06,
      "loss": 0.4166,
      "mean_token_accuracy": 0.8576577016298331,
      "step": 595
    },
    {
      "epoch": 0.8138602031236665,
      "grad_norm": 0.19264903302284703,
      "learning_rate": 2.0350857202212883e-06,
      "loss": 0.4031,
      "mean_token_accuracy": 0.8630792968512606,
      "step": 596
    },
    {
      "epoch": 0.8152257403772296,
      "grad_norm": 0.24074756197143313,
      "learning_rate": 2.0063078663983716e-06,
      "loss": 0.3853,
      "mean_token_accuracy": 0.867487619049804,
      "step": 597
    },
    {
      "epoch": 0.8165912776307929,
      "grad_norm": 0.22693604058548678,
      "learning_rate": 1.977712231949799e-06,
      "loss": 0.4025,
      "mean_token_accuracy": 0.8613917192076461,
      "step": 598
    },
    {
      "epoch": 0.817956814884356,
      "grad_norm": 0.2158385478614588,
      "learning_rate": 1.9492994687243715e-06,
      "loss": 0.4148,
      "mean_token_accuracy": 0.8566573413371225,
      "step": 599
    },
    {
      "epoch": 0.8193223521379193,
      "grad_norm": 0.2531283651455049,
      "learning_rate": 1.9210702244022616e-06,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8552704816161152,
      "step": 600
    },
    {
      "epoch": 0.8206878893914825,
      "grad_norm": 0.22124530534828935,
      "learning_rate": 1.8930251424802648e-06,
      "loss": 0.4173,
      "mean_token_accuracy": 0.857422562323455,
      "step": 601
    },
    {
      "epoch": 0.8220534266450457,
      "grad_norm": 0.22916123571012667,
      "learning_rate": 1.8651648622571128e-06,
      "loss": 0.4093,
      "mean_token_accuracy": 0.8592705248451191,
      "step": 602
    },
    {
      "epoch": 0.8234189638986089,
      "grad_norm": 0.23653504780930373,
      "learning_rate": 1.8374900188189172e-06,
      "loss": 0.4115,
      "mean_token_accuracy": 0.859768333670514,
      "step": 603
    },
    {
      "epoch": 0.8247845011521721,
      "grad_norm": 0.2335856356579208,
      "learning_rate": 1.8100012430246838e-06,
      "loss": 0.3975,
      "mean_token_accuracy": 0.8634662846156729,
      "step": 604
    },
    {
      "epoch": 0.8261500384057353,
      "grad_norm": 0.25602917462585195,
      "learning_rate": 1.7826991614919264e-06,
      "loss": 0.399,
      "mean_token_accuracy": 0.8641089744244426,
      "step": 605
    },
    {
      "epoch": 0.8275155756592985,
      "grad_norm": 0.2764089716972941,
      "learning_rate": 1.7555843965823992e-06,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8591620612370429,
      "step": 606
    },
    {
      "epoch": 0.8288811129128617,
      "grad_norm": 0.22584056903219576,
      "learning_rate": 1.728657566387888e-06,
      "loss": 0.3899,
      "mean_token_accuracy": 0.8659440792321365,
      "step": 607
    },
    {
      "epoch": 0.8302466501664248,
      "grad_norm": 0.20560426304959245,
      "learning_rate": 1.7019192847161425e-06,
      "loss": 0.4031,
      "mean_token_accuracy": 0.8641318011003372,
      "step": 608
    },
    {
      "epoch": 0.8316121874199881,
      "grad_norm": 0.2153106481024899,
      "learning_rate": 1.6753701610768724e-06,
      "loss": 0.4017,
      "mean_token_accuracy": 0.8615005117867286,
      "step": 609
    },
    {
      "epoch": 0.8329777246735512,
      "grad_norm": 0.21220319167956136,
      "learning_rate": 1.6490108006678495e-06,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8564356615328043,
      "step": 610
    },
    {
      "epoch": 0.8343432619271145,
      "grad_norm": 0.22552058407128117,
      "learning_rate": 1.6228418043611227e-06,
      "loss": 0.4168,
      "mean_token_accuracy": 0.8588175846725294,
      "step": 611
    },
    {
      "epoch": 0.8357087991806776,
      "grad_norm": 0.24876240121034418,
      "learning_rate": 1.5968637686893186e-06,
      "loss": 0.4094,
      "mean_token_accuracy": 0.8588914171490348,
      "step": 612
    },
    {
      "epoch": 0.8370743364342409,
      "grad_norm": 0.21709497476785303,
      "learning_rate": 1.57107728583203e-06,
      "loss": 0.3742,
      "mean_token_accuracy": 0.8702544465648172,
      "step": 613
    },
    {
      "epoch": 0.838439873687804,
      "grad_norm": 0.2300739934314909,
      "learning_rate": 1.5454829436023411e-06,
      "loss": 0.4011,
      "mean_token_accuracy": 0.8626383013253395,
      "step": 614
    },
    {
      "epoch": 0.8398054109413673,
      "grad_norm": 0.2233661831188188,
      "learning_rate": 1.5200813254334013e-06,
      "loss": 0.4013,
      "mean_token_accuracy": 0.8617884600134901,
      "step": 615
    },
    {
      "epoch": 0.8411709481949304,
      "grad_norm": 0.2352272063568746,
      "learning_rate": 1.4948730103651498e-06,
      "loss": 0.3927,
      "mean_token_accuracy": 0.8648909153839965,
      "step": 616
    },
    {
      "epoch": 0.8425364854484937,
      "grad_norm": 0.2040512264971376,
      "learning_rate": 1.4698585730311e-06,
      "loss": 0.409,
      "mean_token_accuracy": 0.8602341351681331,
      "step": 617
    },
    {
      "epoch": 0.8439020227020568,
      "grad_norm": 0.2318894554427145,
      "learning_rate": 1.445038583645243e-06,
      "loss": 0.4161,
      "mean_token_accuracy": 0.8594872435807032,
      "step": 618
    },
    {
      "epoch": 0.8452675599556201,
      "grad_norm": 0.20445356255956093,
      "learning_rate": 1.4204136079890585e-06,
      "loss": 0.4031,
      "mean_token_accuracy": 0.8638401437613202,
      "step": 619
    },
    {
      "epoch": 0.8466330972091832,
      "grad_norm": 0.21641039619872623,
      "learning_rate": 1.3959842073986085e-06,
      "loss": 0.3941,
      "mean_token_accuracy": 0.8646075956314934,
      "step": 620
    },
    {
      "epoch": 0.8479986344627465,
      "grad_norm": 0.20773725495938888,
      "learning_rate": 1.3717509387517393e-06,
      "loss": 0.4082,
      "mean_token_accuracy": 0.860045570440718,
      "step": 621
    },
    {
      "epoch": 0.8493641717163096,
      "grad_norm": 0.21271505922050996,
      "learning_rate": 1.3477143544553994e-06,
      "loss": 0.4082,
      "mean_token_accuracy": 0.8599915463398872,
      "step": 622
    },
    {
      "epoch": 0.8507297089698729,
      "grad_norm": 0.22401954459630313,
      "learning_rate": 1.3238750024330338e-06,
      "loss": 0.4192,
      "mean_token_accuracy": 0.8563828512549275,
      "step": 623
    },
    {
      "epoch": 0.852095246223436,
      "grad_norm": 0.22260874162883426,
      "learning_rate": 1.300233426112103e-06,
      "loss": 0.4048,
      "mean_token_accuracy": 0.8611301206735221,
      "step": 624
    },
    {
      "epoch": 0.8534607834769993,
      "grad_norm": 0.21152827385871018,
      "learning_rate": 1.2767901644116943e-06,
      "loss": 0.4146,
      "mean_token_accuracy": 0.8595371254035457,
      "step": 625
    },
    {
      "epoch": 0.8548263207305624,
      "grad_norm": 0.22819406153754265,
      "learning_rate": 1.2535457517302262e-06,
      "loss": 0.4084,
      "mean_token_accuracy": 0.860028563771859,
      "step": 626
    },
    {
      "epoch": 0.8561918579841257,
      "grad_norm": 0.20005533487863839,
      "learning_rate": 1.2305007179332851e-06,
      "loss": 0.4017,
      "mean_token_accuracy": 0.8622309813185363,
      "step": 627
    },
    {
      "epoch": 0.8575573952376888,
      "grad_norm": 0.20176762634176584,
      "learning_rate": 1.2076555883415342e-06,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8536327333990061,
      "step": 628
    },
    {
      "epoch": 0.8589229324912521,
      "grad_norm": 0.2011241526651703,
      "learning_rate": 1.1850108837187336e-06,
      "loss": 0.4086,
      "mean_token_accuracy": 0.8592316098999795,
      "step": 629
    },
    {
      "epoch": 0.8602884697448152,
      "grad_norm": 0.22681886387157124,
      "learning_rate": 1.1625671202598875e-06,
      "loss": 0.4146,
      "mean_token_accuracy": 0.8590380670131639,
      "step": 630
    },
    {
      "epoch": 0.8616540069983785,
      "grad_norm": 0.21912007754709245,
      "learning_rate": 1.1403248095794629e-06,
      "loss": 0.3988,
      "mean_token_accuracy": 0.862137042972187,
      "step": 631
    },
    {
      "epoch": 0.8630195442519416,
      "grad_norm": 0.1991013202311376,
      "learning_rate": 1.1182844586997266e-06,
      "loss": 0.4096,
      "mean_token_accuracy": 0.8605257073793695,
      "step": 632
    },
    {
      "epoch": 0.8643850815055049,
      "grad_norm": 0.23467672143416735,
      "learning_rate": 1.0964465700391979e-06,
      "loss": 0.4108,
      "mean_token_accuracy": 0.8596458375918582,
      "step": 633
    },
    {
      "epoch": 0.865750618759068,
      "grad_norm": 0.21912392965364608,
      "learning_rate": 1.074811641401189e-06,
      "loss": 0.4262,
      "mean_token_accuracy": 0.8559964604503356,
      "step": 634
    },
    {
      "epoch": 0.8671161560126313,
      "grad_norm": 0.20615463939848364,
      "learning_rate": 1.0533801659624531e-06,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8560936656792137,
      "step": 635
    },
    {
      "epoch": 0.8684816932661944,
      "grad_norm": 0.19348528692872238,
      "learning_rate": 1.0321526322619536e-06,
      "loss": 0.4037,
      "mean_token_accuracy": 0.8607088903270385,
      "step": 636
    },
    {
      "epoch": 0.8698472305197577,
      "grad_norm": 0.20196629259646406,
      "learning_rate": 1.0111295241897156e-06,
      "loss": 0.4049,
      "mean_token_accuracy": 0.8631067104794972,
      "step": 637
    },
    {
      "epoch": 0.8712127677733208,
      "grad_norm": 0.2106867654585255,
      "learning_rate": 9.903113209758098e-07,
      "loss": 0.4042,
      "mean_token_accuracy": 0.8603632609752478,
      "step": 638
    },
    {
      "epoch": 0.872578305026884,
      "grad_norm": 0.20073052012465886,
      "learning_rate": 9.696984971794066e-07,
      "loss": 0.3981,
      "mean_token_accuracy": 0.8624080819096971,
      "step": 639
    },
    {
      "epoch": 0.8739438422804472,
      "grad_norm": 0.20533835471618186,
      "learning_rate": 9.492915226779809e-07,
      "loss": 0.3915,
      "mean_token_accuracy": 0.8646312124436534,
      "step": 640
    },
    {
      "epoch": 0.8753093795340104,
      "grad_norm": 0.19390504484310184,
      "learning_rate": 9.290908626565931e-07,
      "loss": 0.4064,
      "mean_token_accuracy": 0.8629363178699737,
      "step": 641
    },
    {
      "epoch": 0.8766749167875736,
      "grad_norm": 0.21843007849082088,
      "learning_rate": 9.090969775972736e-07,
      "loss": 0.4028,
      "mean_token_accuracy": 0.8616676869214925,
      "step": 642
    },
    {
      "epoch": 0.8780404540411368,
      "grad_norm": 0.2219811855088307,
      "learning_rate": 8.89310323268544e-07,
      "loss": 0.4001,
      "mean_token_accuracy": 0.8622636044555403,
      "step": 643
    },
    {
      "epoch": 0.8794059912947,
      "grad_norm": 0.24251626425587186,
      "learning_rate": 8.697313507150184e-07,
      "loss": 0.4033,
      "mean_token_accuracy": 0.8601248910938767,
      "step": 644
    },
    {
      "epoch": 0.8807715285482632,
      "grad_norm": 0.19641237184380816,
      "learning_rate": 8.503605062471187e-07,
      "loss": 0.4061,
      "mean_token_accuracy": 0.8596051420978411,
      "step": 645
    },
    {
      "epoch": 0.8821370658018264,
      "grad_norm": 0.20131416744818392,
      "learning_rate": 8.311982314309109e-07,
      "loss": 0.4113,
      "mean_token_accuracy": 0.8598440955132611,
      "step": 646
    },
    {
      "epoch": 0.8835026030553896,
      "grad_norm": 0.21190627535161183,
      "learning_rate": 8.122449630780238e-07,
      "loss": 0.416,
      "mean_token_accuracy": 0.8575635125425897,
      "step": 647
    },
    {
      "epoch": 0.8848681403089528,
      "grad_norm": 0.2167630347166107,
      "learning_rate": 7.935011332357113e-07,
      "loss": 0.4242,
      "mean_token_accuracy": 0.8560313013882472,
      "step": 648
    },
    {
      "epoch": 0.886233677562516,
      "grad_norm": 0.19473548273128163,
      "learning_rate": 7.749671691769911e-07,
      "loss": 0.3845,
      "mean_token_accuracy": 0.8675812913768379,
      "step": 649
    },
    {
      "epoch": 0.8875992148160792,
      "grad_norm": 0.2060396098320868,
      "learning_rate": 7.566434933909006e-07,
      "loss": 0.4075,
      "mean_token_accuracy": 0.8595329796977811,
      "step": 650
    },
    {
      "epoch": 0.8889647520696424,
      "grad_norm": 0.20145474208727362,
      "learning_rate": 7.385305235728801e-07,
      "loss": 0.4082,
      "mean_token_accuracy": 0.8592664341512477,
      "step": 651
    },
    {
      "epoch": 0.8903302893232056,
      "grad_norm": 0.19909231314570291,
      "learning_rate": 7.206286726152434e-07,
      "loss": 0.4054,
      "mean_token_accuracy": 0.8594394158909082,
      "step": 652
    },
    {
      "epoch": 0.8916958265767688,
      "grad_norm": 0.1919793886619024,
      "learning_rate": 7.029383485977625e-07,
      "loss": 0.4035,
      "mean_token_accuracy": 0.8634067122462318,
      "step": 653
    },
    {
      "epoch": 0.893061363830332,
      "grad_norm": 0.20544119012492554,
      "learning_rate": 6.854599547783736e-07,
      "loss": 0.4058,
      "mean_token_accuracy": 0.8596657002546876,
      "step": 654
    },
    {
      "epoch": 0.8944269010838952,
      "grad_norm": 0.2228496860347481,
      "learning_rate": 6.681938895839746e-07,
      "loss": 0.4002,
      "mean_token_accuracy": 0.862631288272368,
      "step": 655
    },
    {
      "epoch": 0.8957924383374584,
      "grad_norm": 0.19018327112122752,
      "learning_rate": 6.511405466013532e-07,
      "loss": 0.418,
      "mean_token_accuracy": 0.8573043089730195,
      "step": 656
    },
    {
      "epoch": 0.8971579755910216,
      "grad_norm": 0.19393215931484176,
      "learning_rate": 6.343003145682114e-07,
      "loss": 0.4026,
      "mean_token_accuracy": 0.8614415069733373,
      "step": 657
    },
    {
      "epoch": 0.8985235128445848,
      "grad_norm": 0.20746626572888896,
      "learning_rate": 6.176735773642962e-07,
      "loss": 0.4043,
      "mean_token_accuracy": 0.8609983269079549,
      "step": 658
    },
    {
      "epoch": 0.899889050098148,
      "grad_norm": 0.23075626629760573,
      "learning_rate": 6.012607140026605e-07,
      "loss": 0.4122,
      "mean_token_accuracy": 0.8590837704293297,
      "step": 659
    },
    {
      "epoch": 0.9012545873517112,
      "grad_norm": 0.21081958088922736,
      "learning_rate": 5.850620986210198e-07,
      "loss": 0.3963,
      "mean_token_accuracy": 0.8651597082259812,
      "step": 660
    },
    {
      "epoch": 0.9026201246052744,
      "grad_norm": 0.21363684414515646,
      "learning_rate": 5.69078100473216e-07,
      "loss": 0.4012,
      "mean_token_accuracy": 0.8625379850354203,
      "step": 661
    },
    {
      "epoch": 0.9039856618588376,
      "grad_norm": 0.19570623175347346,
      "learning_rate": 5.533090839208133e-07,
      "loss": 0.414,
      "mean_token_accuracy": 0.8574432460527732,
      "step": 662
    },
    {
      "epoch": 0.9053511991124008,
      "grad_norm": 0.20136402481480803,
      "learning_rate": 5.377554084247772e-07,
      "loss": 0.4064,
      "mean_token_accuracy": 0.8597524152182047,
      "step": 663
    },
    {
      "epoch": 0.906716736365964,
      "grad_norm": 0.20829638794206976,
      "learning_rate": 5.224174285372973e-07,
      "loss": 0.4066,
      "mean_token_accuracy": 0.8612470283150278,
      "step": 664
    },
    {
      "epoch": 0.9080822736195272,
      "grad_norm": 0.20559896049133375,
      "learning_rate": 5.072954938936925e-07,
      "loss": 0.3977,
      "mean_token_accuracy": 0.8636562342939346,
      "step": 665
    },
    {
      "epoch": 0.9094478108730903,
      "grad_norm": 0.24180013304406187,
      "learning_rate": 4.923899492044437e-07,
      "loss": 0.4071,
      "mean_token_accuracy": 0.8618068633494765,
      "step": 666
    },
    {
      "epoch": 0.9108133481266536,
      "grad_norm": 0.20544052086394768,
      "learning_rate": 4.777011342473392e-07,
      "loss": 0.4069,
      "mean_token_accuracy": 0.8607662403616217,
      "step": 667
    },
    {
      "epoch": 0.9121788853802167,
      "grad_norm": 0.20218844698835448,
      "learning_rate": 4.632293838597246e-07,
      "loss": 0.4086,
      "mean_token_accuracy": 0.8614379669227725,
      "step": 668
    },
    {
      "epoch": 0.91354442263378,
      "grad_norm": 0.19849540081460978,
      "learning_rate": 4.4897502793087576e-07,
      "loss": 0.4041,
      "mean_token_accuracy": 0.8642634187828342,
      "step": 669
    },
    {
      "epoch": 0.9149099598873431,
      "grad_norm": 0.20745329155441244,
      "learning_rate": 4.3493839139447716e-07,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8541953811730532,
      "step": 670
    },
    {
      "epoch": 0.9162754971409064,
      "grad_norm": 0.2287863594624806,
      "learning_rate": 4.2111979422120863e-07,
      "loss": 0.4053,
      "mean_token_accuracy": 0.8608941533244443,
      "step": 671
    },
    {
      "epoch": 0.9176410343944695,
      "grad_norm": 0.1925356796364094,
      "learning_rate": 4.075195514114594e-07,
      "loss": 0.3803,
      "mean_token_accuracy": 0.8682317326543381,
      "step": 672
    },
    {
      "epoch": 0.9190065716480328,
      "grad_norm": 0.18548297023650745,
      "learning_rate": 3.941379729881456e-07,
      "loss": 0.3944,
      "mean_token_accuracy": 0.864907673080918,
      "step": 673
    },
    {
      "epoch": 0.9203721089015959,
      "grad_norm": 0.21719662700717418,
      "learning_rate": 3.8097536398963965e-07,
      "loss": 0.3976,
      "mean_token_accuracy": 0.8632850150504787,
      "step": 674
    },
    {
      "epoch": 0.9217376461551592,
      "grad_norm": 0.21291688480092102,
      "learning_rate": 3.6803202446282217e-07,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8520979656086647,
      "step": 675
    },
    {
      "epoch": 0.9231031834087223,
      "grad_norm": 0.19317640877157852,
      "learning_rate": 3.553082494562354e-07,
      "loss": 0.4044,
      "mean_token_accuracy": 0.860968928653511,
      "step": 676
    },
    {
      "epoch": 0.9244687206622856,
      "grad_norm": 0.20029096615777936,
      "learning_rate": 3.4280432901336423e-07,
      "loss": 0.4327,
      "mean_token_accuracy": 0.8523410994278862,
      "step": 677
    },
    {
      "epoch": 0.9258342579158487,
      "grad_norm": 0.19089996560484057,
      "learning_rate": 3.3052054816602455e-07,
      "loss": 0.4118,
      "mean_token_accuracy": 0.8587402967704104,
      "step": 678
    },
    {
      "epoch": 0.927199795169412,
      "grad_norm": 0.21562854865168446,
      "learning_rate": 3.1845718692785743e-07,
      "loss": 0.3998,
      "mean_token_accuracy": 0.863486058955801,
      "step": 679
    },
    {
      "epoch": 0.9285653324229751,
      "grad_norm": 0.2074181207840677,
      "learning_rate": 3.0661452028795335e-07,
      "loss": 0.418,
      "mean_token_accuracy": 0.8581571716176858,
      "step": 680
    },
    {
      "epoch": 0.9299308696765384,
      "grad_norm": 0.19559493828044613,
      "learning_rate": 2.949928182045869e-07,
      "loss": 0.3859,
      "mean_token_accuracy": 0.8668547803361453,
      "step": 681
    },
    {
      "epoch": 0.9312964069301015,
      "grad_norm": 0.20748047485633067,
      "learning_rate": 2.835923455990508e-07,
      "loss": 0.4038,
      "mean_token_accuracy": 0.862158416990831,
      "step": 682
    },
    {
      "epoch": 0.9326619441836648,
      "grad_norm": 0.20156584121794527,
      "learning_rate": 2.7241336234962943e-07,
      "loss": 0.3962,
      "mean_token_accuracy": 0.864386738221901,
      "step": 683
    },
    {
      "epoch": 0.9340274814372279,
      "grad_norm": 0.18328492764164395,
      "learning_rate": 2.614561232856672e-07,
      "loss": 0.3818,
      "mean_token_accuracy": 0.8677611815631928,
      "step": 684
    },
    {
      "epoch": 0.9353930186907912,
      "grad_norm": 0.20904145120342477,
      "learning_rate": 2.507208781817638e-07,
      "loss": 0.4285,
      "mean_token_accuracy": 0.8544474486773861,
      "step": 685
    },
    {
      "epoch": 0.9367585559443543,
      "grad_norm": 0.20813281413241,
      "learning_rate": 2.402078717520795e-07,
      "loss": 0.413,
      "mean_token_accuracy": 0.8585056450386133,
      "step": 686
    },
    {
      "epoch": 0.9381240931979176,
      "grad_norm": 0.19274067038403314,
      "learning_rate": 2.2991734364475214e-07,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8552607489362258,
      "step": 687
    },
    {
      "epoch": 0.9394896304514807,
      "grad_norm": 0.18657512512018273,
      "learning_rate": 2.1984952843644104e-07,
      "loss": 0.4048,
      "mean_token_accuracy": 0.8605767691236879,
      "step": 688
    },
    {
      "epoch": 0.940855167705044,
      "grad_norm": 0.20415038653399858,
      "learning_rate": 2.1000465562697858e-07,
      "loss": 0.403,
      "mean_token_accuracy": 0.8614550447092127,
      "step": 689
    },
    {
      "epoch": 0.9422207049586071,
      "grad_norm": 0.18802302287270456,
      "learning_rate": 2.0038294963413251e-07,
      "loss": 0.4102,
      "mean_token_accuracy": 0.8589263672995899,
      "step": 690
    },
    {
      "epoch": 0.9435862422121704,
      "grad_norm": 0.20800845771223211,
      "learning_rate": 1.9098462978849875e-07,
      "loss": 0.4061,
      "mean_token_accuracy": 0.8600364168613337,
      "step": 691
    },
    {
      "epoch": 0.9449517794657335,
      "grad_norm": 0.21444000519203277,
      "learning_rate": 1.8180991032849426e-07,
      "loss": 0.4256,
      "mean_token_accuracy": 0.8541757101430708,
      "step": 692
    },
    {
      "epoch": 0.9463173167192968,
      "grad_norm": 0.2049726644452516,
      "learning_rate": 1.7285900039547997e-07,
      "loss": 0.3977,
      "mean_token_accuracy": 0.8644176084772834,
      "step": 693
    },
    {
      "epoch": 0.9476828539728599,
      "grad_norm": 0.19845365768398393,
      "learning_rate": 1.6413210402898895e-07,
      "loss": 0.422,
      "mean_token_accuracy": 0.8564894930291981,
      "step": 694
    },
    {
      "epoch": 0.9490483912264231,
      "grad_norm": 0.20134810058434158,
      "learning_rate": 1.556294201620734e-07,
      "loss": 0.3986,
      "mean_token_accuracy": 0.8628961715481271,
      "step": 695
    },
    {
      "epoch": 0.9504139284799863,
      "grad_norm": 0.20332703044962025,
      "learning_rate": 1.4735114261677842e-07,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8602014785454245,
      "step": 696
    },
    {
      "epoch": 0.9517794657335495,
      "grad_norm": 0.20262245169618026,
      "learning_rate": 1.3929746009971434e-07,
      "loss": 0.4205,
      "mean_token_accuracy": 0.8558728368817288,
      "step": 697
    },
    {
      "epoch": 0.9531450029871127,
      "grad_norm": 0.1986654702159638,
      "learning_rate": 1.3146855619776134e-07,
      "loss": 0.4164,
      "mean_token_accuracy": 0.8583016759076048,
      "step": 698
    },
    {
      "epoch": 0.9545105402406759,
      "grad_norm": 0.20538122761572822,
      "learning_rate": 1.2386460937387824e-07,
      "loss": 0.4122,
      "mean_token_accuracy": 0.8596236188677384,
      "step": 699
    },
    {
      "epoch": 0.9558760774942391,
      "grad_norm": 0.18771743571795288,
      "learning_rate": 1.1648579296304252e-07,
      "loss": 0.389,
      "mean_token_accuracy": 0.8655003903926352,
      "step": 700
    },
    {
      "epoch": 0.9572416147478023,
      "grad_norm": 0.18918786074101446,
      "learning_rate": 1.0933227516829348e-07,
      "loss": 0.3804,
      "mean_token_accuracy": 0.8682751397466016,
      "step": 701
    },
    {
      "epoch": 0.9586071520013656,
      "grad_norm": 0.21947757366560472,
      "learning_rate": 1.0240421905689746e-07,
      "loss": 0.4128,
      "mean_token_accuracy": 0.858528673911104,
      "step": 702
    },
    {
      "epoch": 0.9599726892549287,
      "grad_norm": 0.197515434889879,
      "learning_rate": 9.570178255663532e-08,
      "loss": 0.3926,
      "mean_token_accuracy": 0.8648129525692759,
      "step": 703
    },
    {
      "epoch": 0.961338226508492,
      "grad_norm": 0.1879435454356452,
      "learning_rate": 8.922511845219972e-08,
      "loss": 0.4138,
      "mean_token_accuracy": 0.8581702095308102,
      "step": 704
    },
    {
      "epoch": 0.9627037637620551,
      "grad_norm": 0.18890815403985045,
      "learning_rate": 8.297437438170797e-08,
      "loss": 0.3907,
      "mean_token_accuracy": 0.865506216595041,
      "step": 705
    },
    {
      "epoch": 0.9640693010156184,
      "grad_norm": 0.286341814862515,
      "learning_rate": 7.694969283334575e-08,
      "loss": 0.4113,
      "mean_token_accuracy": 0.8591874470356847,
      "step": 706
    },
    {
      "epoch": 0.9654348382691815,
      "grad_norm": 0.20273945460392515,
      "learning_rate": 7.115121114211198e-08,
      "loss": 0.3943,
      "mean_token_accuracy": 0.8657841466466322,
      "step": 707
    },
    {
      "epoch": 0.9668003755227448,
      "grad_norm": 0.20652336093906176,
      "learning_rate": 6.557906148669025e-08,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8548559094292058,
      "step": 708
    },
    {
      "epoch": 0.9681659127763079,
      "grad_norm": 0.19629348230830565,
      "learning_rate": 6.023337088643665e-08,
      "loss": 0.3905,
      "mean_token_accuracy": 0.866214795452116,
      "step": 709
    },
    {
      "epoch": 0.9695314500298712,
      "grad_norm": 0.18612630221965354,
      "learning_rate": 5.51142611984834e-08,
      "loss": 0.4012,
      "mean_token_accuracy": 0.8630792198659271,
      "step": 710
    },
    {
      "epoch": 0.9708969872834343,
      "grad_norm": 0.20471539063414793,
      "learning_rate": 5.022184911495864e-08,
      "loss": 0.4088,
      "mean_token_accuracy": 0.8615388696106141,
      "step": 711
    },
    {
      "epoch": 0.9722625245369976,
      "grad_norm": 0.20615922685602697,
      "learning_rate": 4.555624616033427e-08,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8509749138520655,
      "step": 712
    },
    {
      "epoch": 0.9736280617905607,
      "grad_norm": 0.1905175337604626,
      "learning_rate": 4.111755868887346e-08,
      "loss": 0.3965,
      "mean_token_accuracy": 0.8641722387632514,
      "step": 713
    },
    {
      "epoch": 0.974993599044124,
      "grad_norm": 0.21850511626165098,
      "learning_rate": 3.690588788221372e-08,
      "loss": 0.4073,
      "mean_token_accuracy": 0.8611413177219913,
      "step": 714
    },
    {
      "epoch": 0.9763591362976871,
      "grad_norm": 0.19471127549408462,
      "learning_rate": 3.2921329747056527e-08,
      "loss": 0.4208,
      "mean_token_accuracy": 0.8565208481356911,
      "step": 715
    },
    {
      "epoch": 0.9777246735512504,
      "grad_norm": 0.20634538207455483,
      "learning_rate": 2.916397511298019e-08,
      "loss": 0.4022,
      "mean_token_accuracy": 0.8626477917792553,
      "step": 716
    },
    {
      "epoch": 0.9790902108048135,
      "grad_norm": 0.17898675224982286,
      "learning_rate": 2.563390963037149e-08,
      "loss": 0.4084,
      "mean_token_accuracy": 0.858742060603683,
      "step": 717
    },
    {
      "epoch": 0.9804557480583768,
      "grad_norm": 0.20186274942480947,
      "learning_rate": 2.2331213768468363e-08,
      "loss": 0.4112,
      "mean_token_accuracy": 0.8604082683839368,
      "step": 718
    },
    {
      "epoch": 0.9818212853119399,
      "grad_norm": 0.20068940005862676,
      "learning_rate": 1.925596281353026e-08,
      "loss": 0.4061,
      "mean_token_accuracy": 0.8611462253446862,
      "step": 719
    },
    {
      "epoch": 0.9831868225655032,
      "grad_norm": 0.1949558079302809,
      "learning_rate": 1.6408226867118404e-08,
      "loss": 0.3991,
      "mean_token_accuracy": 0.8624983217087219,
      "step": 720
    },
    {
      "epoch": 0.9845523598190663,
      "grad_norm": 0.20269545070379796,
      "learning_rate": 1.3788070844501511e-08,
      "loss": 0.3964,
      "mean_token_accuracy": 0.8662190034977201,
      "step": 721
    },
    {
      "epoch": 0.9859178970726296,
      "grad_norm": 0.19803360839942985,
      "learning_rate": 1.1395554473171421e-08,
      "loss": 0.4237,
      "mean_token_accuracy": 0.8558418698658579,
      "step": 722
    },
    {
      "epoch": 0.9872834343261927,
      "grad_norm": 0.20911643943570965,
      "learning_rate": 9.2307322914853e-09,
      "loss": 0.4014,
      "mean_token_accuracy": 0.8616119983012264,
      "step": 723
    },
    {
      "epoch": 0.988648971579756,
      "grad_norm": 0.20281334643579427,
      "learning_rate": 7.293653647421073e-09,
      "loss": 0.3948,
      "mean_token_accuracy": 0.8643835215948916,
      "step": 724
    },
    {
      "epoch": 0.9900145088333191,
      "grad_norm": 0.1901095357986784,
      "learning_rate": 5.584362697453882e-09,
      "loss": 0.4048,
      "mean_token_accuracy": 0.8623485947951701,
      "step": 725
    },
    {
      "epoch": 0.9913800460868823,
      "grad_norm": 0.20379731214220034,
      "learning_rate": 4.1028984055457856e-09,
      "loss": 0.3983,
      "mean_token_accuracy": 0.8660865418772313,
      "step": 726
    },
    {
      "epoch": 0.9927455833404455,
      "grad_norm": 0.21708635868498447,
      "learning_rate": 2.8492945422620157e-09,
      "loss": 0.4147,
      "mean_token_accuracy": 0.8587614489423447,
      "step": 727
    },
    {
      "epoch": 0.9941111205940087,
      "grad_norm": 0.1929220391858935,
      "learning_rate": 1.8235796839982667e-09,
      "loss": 0.435,
      "mean_token_accuracy": 0.8512693561141161,
      "step": 728
    },
    {
      "epoch": 0.9954766578475719,
      "grad_norm": 0.20428589636781275,
      "learning_rate": 1.0257772123312137e-09,
      "loss": 0.3959,
      "mean_token_accuracy": 0.8652902515910991,
      "step": 729
    },
    {
      "epoch": 0.9968421951011351,
      "grad_norm": 0.1904705638210679,
      "learning_rate": 4.5590531348227443e-10,
      "loss": 0.4054,
      "mean_token_accuracy": 0.8607762718652394,
      "step": 730
    },
    {
      "epoch": 0.9982077323546983,
      "grad_norm": 0.19964279194650797,
      "learning_rate": 1.1397697790793693e-10,
      "loss": 0.4149,
      "mean_token_accuracy": 0.8587917368074317,
      "step": 731
    },
    {
      "epoch": 0.9995732696082615,
      "grad_norm": 0.19717881385523706,
      "learning_rate": 0.0,
      "loss": 0.4046,
      "mean_token_accuracy": 0.8610176941050893,
      "step": 732
    },
    {
      "epoch": 0.9995732696082615,
      "step": 732,
      "total_flos": 2.7844372399870968e+19,
      "train_loss": 0.4556342206203221,
      "train_runtime": 169635.2867,
      "train_samples_per_second": 0.553,
      "train_steps_per_second": 0.004
    }
  ],
  "logging_steps": 1,
  "max_steps": 732,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.7844372399870968e+19,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}