{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.942675159235669,
  "eval_steps": 500,
  "global_step": 390,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.012738853503184714,
      "grad_norm": 11.927668571472168,
      "learning_rate": 0.0,
      "loss": 3.94,
      "mean_token_accuracy": 0.4029204398393631,
      "num_tokens": 11165.0,
      "step": 1
    },
    {
      "epoch": 0.025477707006369428,
      "grad_norm": 11.546525001525879,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 3.9572,
      "mean_token_accuracy": 0.39499808847904205,
      "num_tokens": 22924.0,
      "step": 2
    },
    {
      "epoch": 0.03821656050955414,
      "grad_norm": 8.797052383422852,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 3.857,
      "mean_token_accuracy": 0.39673660695552826,
      "num_tokens": 34380.0,
      "step": 3
    },
    {
      "epoch": 0.050955414012738856,
      "grad_norm": 7.517773628234863,
      "learning_rate": 5e-05,
      "loss": 3.4045,
      "mean_token_accuracy": 0.4453563541173935,
      "num_tokens": 46499.0,
      "step": 4
    },
    {
      "epoch": 0.06369426751592357,
      "grad_norm": 8.536116600036621,
      "learning_rate": 6.666666666666667e-05,
      "loss": 3.022,
      "mean_token_accuracy": 0.4896288365125656,
      "num_tokens": 58004.0,
      "step": 5
    },
    {
      "epoch": 0.07643312101910828,
      "grad_norm": 3.957106828689575,
      "learning_rate": 8.333333333333334e-05,
      "loss": 2.5071,
      "mean_token_accuracy": 0.5356468260288239,
      "num_tokens": 69676.0,
      "step": 6
    },
    {
      "epoch": 0.08917197452229299,
      "grad_norm": 2.6128907203674316,
      "learning_rate": 0.0001,
      "loss": 2.2479,
      "mean_token_accuracy": 0.6039679944515228,
      "num_tokens": 80510.0,
      "step": 7
    },
    {
      "epoch": 0.10191082802547771,
      "grad_norm": 1.6600596904754639,
      "learning_rate": 0.00011666666666666668,
      "loss": 1.9613,
      "mean_token_accuracy": 0.6389846205711365,
      "num_tokens": 92115.0,
      "step": 8
    },
    {
      "epoch": 0.11464968152866242,
      "grad_norm": 1.1175001859664917,
      "learning_rate": 0.00013333333333333334,
      "loss": 1.828,
      "mean_token_accuracy": 0.6838500797748566,
      "num_tokens": 103256.0,
      "step": 9
    },
    {
      "epoch": 0.12738853503184713,
      "grad_norm": 2.3084917068481445,
      "learning_rate": 0.00015000000000000001,
      "loss": 1.9173,
      "mean_token_accuracy": 0.6635328233242035,
      "num_tokens": 115085.0,
      "step": 10
    },
    {
      "epoch": 0.14012738853503184,
      "grad_norm": 5.995814323425293,
      "learning_rate": 0.0001666666666666667,
      "loss": 1.9407,
      "mean_token_accuracy": 0.6618090569972992,
      "num_tokens": 126597.0,
      "step": 11
    },
    {
      "epoch": 0.15286624203821655,
      "grad_norm": 2.0090432167053223,
      "learning_rate": 0.00018333333333333334,
      "loss": 1.815,
      "mean_token_accuracy": 0.6675754189491272,
      "num_tokens": 138225.0,
      "step": 12
    },
    {
      "epoch": 0.16560509554140126,
      "grad_norm": 4.294322490692139,
      "learning_rate": 0.0002,
      "loss": 1.7763,
      "mean_token_accuracy": 0.6821876168251038,
      "num_tokens": 149400.0,
      "step": 13
    },
    {
      "epoch": 0.17834394904458598,
      "grad_norm": 1.290095329284668,
      "learning_rate": 0.0001999965463076377,
      "loss": 1.7811,
      "mean_token_accuracy": 0.6758189797401428,
      "num_tokens": 161043.0,
      "step": 14
    },
    {
      "epoch": 0.1910828025477707,
      "grad_norm": 1.1799252033233643,
      "learning_rate": 0.00019998618546911056,
      "loss": 1.6578,
      "mean_token_accuracy": 0.6912119388580322,
      "num_tokens": 172432.0,
      "step": 15
    },
    {
      "epoch": 0.20382165605095542,
      "grad_norm": 1.018792748451233,
      "learning_rate": 0.00019996891820008164,
      "loss": 1.7968,
      "mean_token_accuracy": 0.6723423898220062,
      "num_tokens": 184203.0,
      "step": 16
    },
    {
      "epoch": 0.21656050955414013,
      "grad_norm": 1.8456875085830688,
      "learning_rate": 0.00019994474569326757,
      "loss": 1.7761,
      "mean_token_accuracy": 0.6694794595241547,
      "num_tokens": 196265.0,
      "step": 17
    },
    {
      "epoch": 0.22929936305732485,
      "grad_norm": 1.2281064987182617,
      "learning_rate": 0.00019991366961835642,
      "loss": 1.6356,
      "mean_token_accuracy": 0.699277937412262,
      "num_tokens": 207637.0,
      "step": 18
    },
    {
      "epoch": 0.24203821656050956,
      "grad_norm": 1.1970819234848022,
      "learning_rate": 0.00019987569212189224,
      "loss": 1.7103,
      "mean_token_accuracy": 0.6887260973453522,
      "num_tokens": 218753.0,
      "step": 19
    },
    {
      "epoch": 0.25477707006369427,
      "grad_norm": 0.4748772084712982,
      "learning_rate": 0.00019983081582712685,
      "loss": 1.7107,
      "mean_token_accuracy": 0.6865908205509186,
      "num_tokens": 230421.0,
      "step": 20
    },
    {
      "epoch": 0.267515923566879,
      "grad_norm": 0.6084349751472473,
      "learning_rate": 0.0001997790438338385,
      "loss": 1.685,
      "mean_token_accuracy": 0.6917995810508728,
      "num_tokens": 241772.0,
      "step": 21
    },
    {
      "epoch": 0.2802547770700637,
      "grad_norm": 0.6356724500656128,
      "learning_rate": 0.00019972037971811802,
      "loss": 1.6516,
      "mean_token_accuracy": 0.6938049793243408,
      "num_tokens": 252948.0,
      "step": 22
    },
    {
      "epoch": 0.2929936305732484,
      "grad_norm": 0.5958734750747681,
      "learning_rate": 0.00019965482753212156,
      "loss": 1.604,
      "mean_token_accuracy": 0.7011500597000122,
      "num_tokens": 264329.0,
      "step": 23
    },
    {
      "epoch": 0.3057324840764331,
      "grad_norm": 0.556065022945404,
      "learning_rate": 0.0001995823918037908,
      "loss": 1.6563,
      "mean_token_accuracy": 0.6891041994094849,
      "num_tokens": 275501.0,
      "step": 24
    },
    {
      "epoch": 0.3184713375796178,
      "grad_norm": 0.4661395251750946,
      "learning_rate": 0.00019950307753654017,
      "loss": 1.6723,
      "mean_token_accuracy": 0.691864550113678,
      "num_tokens": 286871.0,
      "step": 25
    },
    {
      "epoch": 0.33121019108280253,
      "grad_norm": 0.4898029863834381,
      "learning_rate": 0.0001994168902089112,
      "loss": 1.7023,
      "mean_token_accuracy": 0.6883004903793335,
      "num_tokens": 298281.0,
      "step": 26
    },
    {
      "epoch": 0.34394904458598724,
      "grad_norm": 0.627682626247406,
      "learning_rate": 0.00019932383577419432,
      "loss": 1.6222,
      "mean_token_accuracy": 0.6978872120380402,
      "num_tokens": 309610.0,
      "step": 27
    },
    {
      "epoch": 0.35668789808917195,
      "grad_norm": 0.5940748453140259,
      "learning_rate": 0.00019922392066001722,
      "loss": 1.5325,
      "mean_token_accuracy": 0.7118729054927826,
      "num_tokens": 320475.0,
      "step": 28
    },
    {
      "epoch": 0.36942675159235666,
      "grad_norm": 0.5121145248413086,
      "learning_rate": 0.0001991171517679013,
      "loss": 1.6756,
      "mean_token_accuracy": 0.6806293725967407,
      "num_tokens": 332574.0,
      "step": 29
    },
    {
      "epoch": 0.3821656050955414,
      "grad_norm": 0.4396442472934723,
      "learning_rate": 0.00019900353647278466,
      "loss": 1.6497,
      "mean_token_accuracy": 0.693237692117691,
      "num_tokens": 344055.0,
      "step": 30
    },
    {
      "epoch": 0.39490445859872614,
      "grad_norm": 0.4567514657974243,
      "learning_rate": 0.00019888308262251285,
      "loss": 1.7773,
      "mean_token_accuracy": 0.6751766800880432,
      "num_tokens": 356104.0,
      "step": 31
    },
    {
      "epoch": 0.40764331210191085,
      "grad_norm": 0.5211770534515381,
      "learning_rate": 0.00019875579853729676,
      "loss": 1.663,
      "mean_token_accuracy": 0.6896547377109528,
      "num_tokens": 367582.0,
      "step": 32
    },
    {
      "epoch": 0.42038216560509556,
      "grad_norm": 0.5280752182006836,
      "learning_rate": 0.00019862169300913785,
      "loss": 1.7045,
      "mean_token_accuracy": 0.6897095143795013,
      "num_tokens": 378963.0,
      "step": 33
    },
    {
      "epoch": 0.43312101910828027,
      "grad_norm": 0.4466302990913391,
      "learning_rate": 0.00019848077530122083,
      "loss": 1.6252,
      "mean_token_accuracy": 0.6934365034103394,
      "num_tokens": 390443.0,
      "step": 34
    },
    {
      "epoch": 0.445859872611465,
      "grad_norm": 0.42075785994529724,
      "learning_rate": 0.00019833305514727395,
      "loss": 1.7402,
      "mean_token_accuracy": 0.6816126108169556,
      "num_tokens": 402077.0,
      "step": 35
    },
    {
      "epoch": 0.4585987261146497,
      "grad_norm": 0.44698265194892883,
      "learning_rate": 0.0001981785427508966,
      "loss": 1.6314,
      "mean_token_accuracy": 0.7031489312648773,
      "num_tokens": 413215.0,
      "step": 36
    },
    {
      "epoch": 0.4713375796178344,
      "grad_norm": 0.5105100274085999,
      "learning_rate": 0.00019801724878485438,
      "loss": 1.6837,
      "mean_token_accuracy": 0.6941170394420624,
      "num_tokens": 424372.0,
      "step": 37
    },
    {
      "epoch": 0.4840764331210191,
      "grad_norm": 0.49002698063850403,
      "learning_rate": 0.00019784918439034216,
      "loss": 1.678,
      "mean_token_accuracy": 0.6886717677116394,
      "num_tokens": 436262.0,
      "step": 38
    },
    {
      "epoch": 0.4968152866242038,
      "grad_norm": 0.4677882790565491,
      "learning_rate": 0.00019767436117621413,
      "loss": 1.6349,
      "mean_token_accuracy": 0.6940007507801056,
      "num_tokens": 447409.0,
      "step": 39
    },
    {
      "epoch": 0.5095541401273885,
      "grad_norm": 0.41676995158195496,
      "learning_rate": 0.00019749279121818235,
      "loss": 1.6196,
      "mean_token_accuracy": 0.6940004527568817,
      "num_tokens": 458986.0,
      "step": 40
    },
    {
      "epoch": 0.5222929936305732,
      "grad_norm": 0.4395844340324402,
      "learning_rate": 0.00019730448705798239,
      "loss": 1.5676,
      "mean_token_accuracy": 0.708065003156662,
      "num_tokens": 470110.0,
      "step": 41
    },
    {
      "epoch": 0.535031847133758,
      "grad_norm": 0.4528355598449707,
      "learning_rate": 0.000197109461702507,
      "loss": 1.62,
      "mean_token_accuracy": 0.6976854205131531,
      "num_tokens": 481475.0,
      "step": 42
    },
    {
      "epoch": 0.5477707006369427,
      "grad_norm": 0.45771703124046326,
      "learning_rate": 0.0001969077286229078,
      "loss": 1.6287,
      "mean_token_accuracy": 0.6974550783634186,
      "num_tokens": 492753.0,
      "step": 43
    },
    {
      "epoch": 0.5605095541401274,
      "grad_norm": 0.4307943880558014,
      "learning_rate": 0.00019669930175366472,
      "loss": 1.6269,
      "mean_token_accuracy": 0.7005655765533447,
      "num_tokens": 504069.0,
      "step": 44
    },
    {
      "epoch": 0.5732484076433121,
      "grad_norm": 0.43306100368499756,
      "learning_rate": 0.00019648419549162348,
      "loss": 1.6797,
      "mean_token_accuracy": 0.6910405457019806,
      "num_tokens": 515873.0,
      "step": 45
    },
    {
      "epoch": 0.5859872611464968,
      "grad_norm": 0.40998607873916626,
      "learning_rate": 0.0001962624246950012,
      "loss": 1.6704,
      "mean_token_accuracy": 0.6908841133117676,
      "num_tokens": 527285.0,
      "step": 46
    },
    {
      "epoch": 0.5987261146496815,
      "grad_norm": 0.423794150352478,
      "learning_rate": 0.00019603400468235998,
      "loss": 1.6225,
      "mean_token_accuracy": 0.6972988545894623,
      "num_tokens": 539008.0,
      "step": 47
    },
    {
      "epoch": 0.6114649681528662,
      "grad_norm": 0.4383174479007721,
      "learning_rate": 0.0001957989512315489,
      "loss": 1.6959,
      "mean_token_accuracy": 0.6844666004180908,
      "num_tokens": 550975.0,
      "step": 48
    },
    {
      "epoch": 0.6242038216560509,
      "grad_norm": 0.426257461309433,
      "learning_rate": 0.0001955572805786141,
      "loss": 1.5511,
      "mean_token_accuracy": 0.7047916054725647,
      "num_tokens": 562174.0,
      "step": 49
    },
    {
      "epoch": 0.6369426751592356,
      "grad_norm": 0.42514124512672424,
      "learning_rate": 0.0001953090094166773,
      "loss": 1.6383,
      "mean_token_accuracy": 0.6950445175170898,
      "num_tokens": 573565.0,
      "step": 50
    },
    {
      "epoch": 0.6496815286624203,
      "grad_norm": 0.4026625156402588,
      "learning_rate": 0.0001950541548947829,
      "loss": 1.5672,
      "mean_token_accuracy": 0.706762433052063,
      "num_tokens": 584655.0,
      "step": 51
    },
    {
      "epoch": 0.6624203821656051,
      "grad_norm": 0.43084436655044556,
      "learning_rate": 0.0001947927346167132,
      "loss": 1.5972,
      "mean_token_accuracy": 0.7003117203712463,
      "num_tokens": 595966.0,
      "step": 52
    },
    {
      "epoch": 0.6751592356687898,
      "grad_norm": 0.4430312514305115,
      "learning_rate": 0.00019452476663977248,
      "loss": 1.654,
      "mean_token_accuracy": 0.687600314617157,
      "num_tokens": 607211.0,
      "step": 53
    },
    {
      "epoch": 0.6878980891719745,
      "grad_norm": 0.41114339232444763,
      "learning_rate": 0.00019425026947353992,
      "loss": 1.5984,
      "mean_token_accuracy": 0.7026942670345306,
      "num_tokens": 618727.0,
      "step": 54
    },
    {
      "epoch": 0.7006369426751592,
      "grad_norm": 0.4260961413383484,
      "learning_rate": 0.00019396926207859084,
      "loss": 1.6138,
      "mean_token_accuracy": 0.6967755258083344,
      "num_tokens": 630397.0,
      "step": 55
    },
    {
      "epoch": 0.7133757961783439,
      "grad_norm": 0.47998639941215515,
      "learning_rate": 0.0001936817638651871,
      "loss": 1.5905,
      "mean_token_accuracy": 0.7036975026130676,
      "num_tokens": 641671.0,
      "step": 56
    },
    {
      "epoch": 0.7261146496815286,
      "grad_norm": 0.4328579604625702,
      "learning_rate": 0.00019338779469193639,
      "loss": 1.4959,
      "mean_token_accuracy": 0.7174738645553589,
      "num_tokens": 652634.0,
      "step": 57
    },
    {
      "epoch": 0.7388535031847133,
      "grad_norm": 0.42225387692451477,
      "learning_rate": 0.00019308737486442045,
      "loss": 1.6193,
      "mean_token_accuracy": 0.6919732391834259,
      "num_tokens": 664179.0,
      "step": 58
    },
    {
      "epoch": 0.7515923566878981,
      "grad_norm": 0.41233500838279724,
      "learning_rate": 0.00019278052513379255,
      "loss": 1.5598,
      "mean_token_accuracy": 0.7107926309108734,
      "num_tokens": 675385.0,
      "step": 59
    },
    {
      "epoch": 0.7643312101910829,
      "grad_norm": 0.4194766879081726,
      "learning_rate": 0.00019246726669534415,
      "loss": 1.5903,
      "mean_token_accuracy": 0.7004102170467377,
      "num_tokens": 686817.0,
      "step": 60
    },
    {
      "epoch": 0.7770700636942676,
      "grad_norm": 0.4025111794471741,
      "learning_rate": 0.00019214762118704076,
      "loss": 1.5104,
      "mean_token_accuracy": 0.7114810347557068,
      "num_tokens": 697815.0,
      "step": 61
    },
    {
      "epoch": 0.7898089171974523,
      "grad_norm": 0.41542673110961914,
      "learning_rate": 0.00019182161068802741,
      "loss": 1.6112,
      "mean_token_accuracy": 0.6955437958240509,
      "num_tokens": 709534.0,
      "step": 62
    },
    {
      "epoch": 0.802547770700637,
      "grad_norm": 0.41509905457496643,
      "learning_rate": 0.00019148925771710347,
      "loss": 1.5978,
      "mean_token_accuracy": 0.7051181793212891,
      "num_tokens": 720860.0,
      "step": 63
    },
    {
      "epoch": 0.8152866242038217,
      "grad_norm": 0.40592867136001587,
      "learning_rate": 0.00019115058523116733,
      "loss": 1.566,
      "mean_token_accuracy": 0.7047462165355682,
      "num_tokens": 732087.0,
      "step": 64
    },
    {
      "epoch": 0.8280254777070064,
      "grad_norm": 0.4121105372905731,
      "learning_rate": 0.0001908056166236305,
      "loss": 1.602,
      "mean_token_accuracy": 0.698013037443161,
      "num_tokens": 743945.0,
      "step": 65
    },
    {
      "epoch": 0.8407643312101911,
      "grad_norm": 0.43075302243232727,
      "learning_rate": 0.00019045437572280194,
      "loss": 1.5842,
      "mean_token_accuracy": 0.704198956489563,
      "num_tokens": 755005.0,
      "step": 66
    },
    {
      "epoch": 0.8535031847133758,
      "grad_norm": 0.42977190017700195,
      "learning_rate": 0.0001900968867902419,
      "loss": 1.5603,
      "mean_token_accuracy": 0.7032743096351624,
      "num_tokens": 766318.0,
      "step": 67
    },
    {
      "epoch": 0.8662420382165605,
      "grad_norm": 0.43815624713897705,
      "learning_rate": 0.00018973317451908642,
      "loss": 1.5929,
      "mean_token_accuracy": 0.7039108872413635,
      "num_tokens": 777726.0,
      "step": 68
    },
    {
      "epoch": 0.8789808917197452,
      "grad_norm": 0.46719834208488464,
      "learning_rate": 0.00018936326403234125,
      "loss": 1.5957,
      "mean_token_accuracy": 0.7035625278949738,
      "num_tokens": 788880.0,
      "step": 69
    },
    {
      "epoch": 0.89171974522293,
      "grad_norm": 0.4173584282398224,
      "learning_rate": 0.0001889871808811469,
      "loss": 1.5941,
      "mean_token_accuracy": 0.7020141184329987,
      "num_tokens": 800546.0,
      "step": 70
    },
    {
      "epoch": 0.9044585987261147,
      "grad_norm": 0.42461779713630676,
      "learning_rate": 0.00018860495104301345,
      "loss": 1.5814,
      "mean_token_accuracy": 0.7005370259284973,
      "num_tokens": 812189.0,
      "step": 71
    },
    {
      "epoch": 0.9171974522292994,
      "grad_norm": 0.4130838215351105,
      "learning_rate": 0.00018821660092002641,
      "loss": 1.6225,
      "mean_token_accuracy": 0.6930609941482544,
      "num_tokens": 823885.0,
      "step": 72
    },
    {
      "epoch": 0.9299363057324841,
      "grad_norm": 0.40940752625465393,
      "learning_rate": 0.00018782215733702286,
      "loss": 1.5639,
      "mean_token_accuracy": 0.7046359479427338,
      "num_tokens": 835303.0,
      "step": 73
    },
    {
      "epoch": 0.9426751592356688,
      "grad_norm": 0.4241730868816376,
      "learning_rate": 0.00018742164753973855,
      "loss": 1.5594,
      "mean_token_accuracy": 0.7004008889198303,
      "num_tokens": 847069.0,
      "step": 74
    },
    {
      "epoch": 0.9554140127388535,
      "grad_norm": 0.4133743941783905,
      "learning_rate": 0.00018701509919292613,
      "loss": 1.5119,
      "mean_token_accuracy": 0.712060958147049,
      "num_tokens": 858488.0,
      "step": 75
    },
    {
      "epoch": 0.9681528662420382,
      "grad_norm": 0.4232250452041626,
      "learning_rate": 0.00018660254037844388,
      "loss": 1.5155,
      "mean_token_accuracy": 0.7139077484607697,
      "num_tokens": 869684.0,
      "step": 76
    },
    {
      "epoch": 0.9808917197452229,
      "grad_norm": 0.4363568127155304,
      "learning_rate": 0.0001861839995933164,
      "loss": 1.5544,
      "mean_token_accuracy": 0.7076265513896942,
      "num_tokens": 880729.0,
      "step": 77
    },
    {
      "epoch": 0.9936305732484076,
      "grad_norm": 0.4357876181602478,
      "learning_rate": 0.00018575950574776595,
      "loss": 1.5327,
      "mean_token_accuracy": 0.7089995443820953,
      "num_tokens": 891695.0,
      "step": 78
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.4357876181602478,
      "learning_rate": 0.00018532908816321558,
      "loss": 1.511,
      "mean_token_accuracy": 0.7218481302261353,
      "num_tokens": 897073.0,
      "step": 79
    },
    {
      "epoch": 1.0127388535031847,
      "grad_norm": 0.7058189511299133,
      "learning_rate": 0.00018489277657026375,
      "loss": 1.3819,
      "mean_token_accuracy": 0.7328154742717743,
      "num_tokens": 908448.0,
      "step": 80
    },
    {
      "epoch": 1.0254777070063694,
      "grad_norm": 0.43703678250312805,
      "learning_rate": 0.0001844506011066308,
      "loss": 1.3035,
      "mean_token_accuracy": 0.7432056665420532,
      "num_tokens": 919804.0,
      "step": 81
    },
    {
      "epoch": 1.0382165605095541,
      "grad_norm": 0.3936041593551636,
      "learning_rate": 0.00018400259231507717,
      "loss": 1.344,
      "mean_token_accuracy": 0.7374230027198792,
      "num_tokens": 931397.0,
      "step": 82
    },
    {
      "epoch": 1.0509554140127388,
      "grad_norm": 0.5080980658531189,
      "learning_rate": 0.00018354878114129367,
      "loss": 1.4085,
      "mean_token_accuracy": 0.72797891497612,
      "num_tokens": 942878.0,
      "step": 83
    },
    {
      "epoch": 1.0636942675159236,
      "grad_norm": 0.5713034272193909,
      "learning_rate": 0.00018308919893176396,
      "loss": 1.4078,
      "mean_token_accuracy": 0.7352543771266937,
      "num_tokens": 954219.0,
      "step": 84
    },
    {
      "epoch": 1.0764331210191083,
      "grad_norm": 0.4511411488056183,
      "learning_rate": 0.0001826238774315995,
      "loss": 1.2804,
      "mean_token_accuracy": 0.7498015761375427,
      "num_tokens": 965410.0,
      "step": 85
    },
    {
      "epoch": 1.089171974522293,
      "grad_norm": 0.4187650680541992,
      "learning_rate": 0.00018215284878234642,
      "loss": 1.29,
      "mean_token_accuracy": 0.7490876317024231,
      "num_tokens": 976694.0,
      "step": 86
    },
    {
      "epoch": 1.1019108280254777,
      "grad_norm": 0.4384659230709076,
      "learning_rate": 0.00018167614551976567,
      "loss": 1.3881,
      "mean_token_accuracy": 0.7267170548439026,
      "num_tokens": 988106.0,
      "step": 87
    },
    {
      "epoch": 1.1146496815286624,
      "grad_norm": 0.4444275498390198,
      "learning_rate": 0.00018119380057158568,
      "loss": 1.4328,
      "mean_token_accuracy": 0.7284035384654999,
      "num_tokens": 999816.0,
      "step": 88
    },
    {
      "epoch": 1.127388535031847,
      "grad_norm": 0.45921292901039124,
      "learning_rate": 0.00018070584725522762,
      "loss": 1.2876,
      "mean_token_accuracy": 0.7462576627731323,
      "num_tokens": 1011401.0,
      "step": 89
    },
    {
      "epoch": 1.1401273885350318,
      "grad_norm": 0.4768648147583008,
      "learning_rate": 0.0001802123192755044,
      "loss": 1.3651,
      "mean_token_accuracy": 0.7335284352302551,
      "num_tokens": 1022628.0,
      "step": 90
    },
    {
      "epoch": 1.1528662420382165,
      "grad_norm": 0.5426104068756104,
      "learning_rate": 0.00017971325072229226,
      "loss": 1.3414,
      "mean_token_accuracy": 0.7375869750976562,
      "num_tokens": 1034112.0,
      "step": 91
    },
    {
      "epoch": 1.1656050955414012,
      "grad_norm": 0.4837786853313446,
      "learning_rate": 0.00017920867606817625,
      "loss": 1.3065,
      "mean_token_accuracy": 0.7411251068115234,
      "num_tokens": 1045699.0,
      "step": 92
    },
    {
      "epoch": 1.178343949044586,
      "grad_norm": 0.5525464415550232,
      "learning_rate": 0.0001786986301660689,
      "loss": 1.2697,
      "mean_token_accuracy": 0.7516984343528748,
      "num_tokens": 1057016.0,
      "step": 93
    },
    {
      "epoch": 1.1910828025477707,
      "grad_norm": 0.5046163201332092,
      "learning_rate": 0.000178183148246803,
      "loss": 1.3359,
      "mean_token_accuracy": 0.7414124608039856,
      "num_tokens": 1068767.0,
      "step": 94
    },
    {
      "epoch": 1.2038216560509554,
      "grad_norm": 0.49643421173095703,
      "learning_rate": 0.00017766226591669785,
      "loss": 1.3022,
      "mean_token_accuracy": 0.7424289584159851,
      "num_tokens": 1080338.0,
      "step": 95
    },
    {
      "epoch": 1.21656050955414,
      "grad_norm": 0.4751627743244171,
      "learning_rate": 0.0001771360191551,
      "loss": 1.2796,
      "mean_token_accuracy": 0.7520297765731812,
      "num_tokens": 1091596.0,
      "step": 96
    },
    {
      "epoch": 1.2292993630573248,
      "grad_norm": 0.48828235268592834,
      "learning_rate": 0.0001766044443118978,
      "loss": 1.3714,
      "mean_token_accuracy": 0.7391779124736786,
      "num_tokens": 1103173.0,
      "step": 97
    },
    {
      "epoch": 1.2420382165605095,
      "grad_norm": 0.5226894617080688,
      "learning_rate": 0.00017606757810501088,
      "loss": 1.349,
      "mean_token_accuracy": 0.7393156886100769,
      "num_tokens": 1114890.0,
      "step": 98
    },
    {
      "epoch": 1.2547770700636942,
      "grad_norm": 0.5849791169166565,
      "learning_rate": 0.0001755254576178535,
      "loss": 1.2692,
      "mean_token_accuracy": 0.7468427419662476,
      "num_tokens": 1126503.0,
      "step": 99
    },
    {
      "epoch": 1.267515923566879,
      "grad_norm": 0.5286082625389099,
      "learning_rate": 0.00017497812029677344,
      "loss": 1.3354,
      "mean_token_accuracy": 0.7440425157546997,
      "num_tokens": 1138075.0,
      "step": 100
    },
    {
      "epoch": 1.2802547770700636,
      "grad_norm": 0.5340179800987244,
      "learning_rate": 0.00017442560394846516,
      "loss": 1.3362,
      "mean_token_accuracy": 0.7400276064872742,
      "num_tokens": 1149604.0,
      "step": 101
    },
    {
      "epoch": 1.2929936305732483,
      "grad_norm": 0.5466039180755615,
      "learning_rate": 0.0001738679467373586,
      "loss": 1.3977,
      "mean_token_accuracy": 0.7267497181892395,
      "num_tokens": 1161198.0,
      "step": 102
    },
    {
      "epoch": 1.305732484076433,
      "grad_norm": 0.49215564131736755,
      "learning_rate": 0.00017330518718298264,
      "loss": 1.3782,
      "mean_token_accuracy": 0.7327314019203186,
      "num_tokens": 1172445.0,
      "step": 103
    },
    {
      "epoch": 1.3184713375796178,
      "grad_norm": 0.49711376428604126,
      "learning_rate": 0.00017273736415730488,
      "loss": 1.1644,
      "mean_token_accuracy": 0.7657121419906616,
      "num_tokens": 1183328.0,
      "step": 104
    },
    {
      "epoch": 1.3312101910828025,
      "grad_norm": 0.5252436995506287,
      "learning_rate": 0.0001721645168820462,
      "loss": 1.3197,
      "mean_token_accuracy": 0.74147829413414,
      "num_tokens": 1194762.0,
      "step": 105
    },
    {
      "epoch": 1.3439490445859872,
      "grad_norm": 0.4682621359825134,
      "learning_rate": 0.00017158668492597186,
      "loss": 1.2746,
      "mean_token_accuracy": 0.7538455724716187,
      "num_tokens": 1206185.0,
      "step": 106
    },
    {
      "epoch": 1.356687898089172,
      "grad_norm": 0.5762368440628052,
      "learning_rate": 0.00017100390820215804,
      "loss": 1.343,
      "mean_token_accuracy": 0.738945722579956,
      "num_tokens": 1217884.0,
      "step": 107
    },
    {
      "epoch": 1.3694267515923566,
      "grad_norm": 0.5941218137741089,
      "learning_rate": 0.00017041622696523518,
      "loss": 1.3016,
      "mean_token_accuracy": 0.7432995140552521,
      "num_tokens": 1229627.0,
      "step": 108
    },
    {
      "epoch": 1.3821656050955413,
      "grad_norm": 0.560783326625824,
      "learning_rate": 0.00016982368180860728,
      "loss": 1.276,
      "mean_token_accuracy": 0.7532786428928375,
      "num_tokens": 1240446.0,
      "step": 109
    },
    {
      "epoch": 1.394904458598726,
      "grad_norm": 0.5108466744422913,
      "learning_rate": 0.00016922631366164797,
      "loss": 1.2346,
      "mean_token_accuracy": 0.7562114298343658,
      "num_tokens": 1251736.0,
      "step": 110
    },
    {
      "epoch": 1.4076433121019107,
      "grad_norm": 0.5147649049758911,
      "learning_rate": 0.0001686241637868734,
      "loss": 1.3012,
      "mean_token_accuracy": 0.7439581155776978,
      "num_tokens": 1263064.0,
      "step": 111
    },
    {
      "epoch": 1.4203821656050954,
      "grad_norm": 0.5320042967796326,
      "learning_rate": 0.00016801727377709194,
      "loss": 1.3648,
      "mean_token_accuracy": 0.7372605800628662,
      "num_tokens": 1274626.0,
      "step": 112
    },
    {
      "epoch": 1.4331210191082802,
      "grad_norm": 0.5413700342178345,
      "learning_rate": 0.00016740568555253155,
      "loss": 1.3992,
      "mean_token_accuracy": 0.732212245464325,
      "num_tokens": 1286160.0,
      "step": 113
    },
    {
      "epoch": 1.4458598726114649,
      "grad_norm": 0.557081401348114,
      "learning_rate": 0.00016678944135794374,
      "loss": 1.2657,
      "mean_token_accuracy": 0.7479846477508545,
      "num_tokens": 1297753.0,
      "step": 114
    },
    {
      "epoch": 1.4585987261146496,
      "grad_norm": 0.5675191283226013,
      "learning_rate": 0.00016616858375968595,
      "loss": 1.3236,
      "mean_token_accuracy": 0.742518424987793,
      "num_tokens": 1309224.0,
      "step": 115
    },
    {
      "epoch": 1.4713375796178343,
      "grad_norm": 0.5020204782485962,
      "learning_rate": 0.000165543155642781,
      "loss": 1.2995,
      "mean_token_accuracy": 0.7413056790828705,
      "num_tokens": 1320675.0,
      "step": 116
    },
    {
      "epoch": 1.484076433121019,
      "grad_norm": 0.5299273133277893,
      "learning_rate": 0.0001649132002079552,
      "loss": 1.2335,
      "mean_token_accuracy": 0.7618575990200043,
      "num_tokens": 1332170.0,
      "step": 117
    },
    {
      "epoch": 1.4968152866242037,
      "grad_norm": 0.5424625873565674,
      "learning_rate": 0.00016427876096865394,
      "loss": 1.3476,
      "mean_token_accuracy": 0.7395524382591248,
      "num_tokens": 1343771.0,
      "step": 118
    },
    {
      "epoch": 1.5095541401273884,
      "grad_norm": 0.5220779180526733,
      "learning_rate": 0.00016363988174803638,
      "loss": 1.3493,
      "mean_token_accuracy": 0.7352395057678223,
      "num_tokens": 1355513.0,
      "step": 119
    },
    {
      "epoch": 1.5222929936305731,
      "grad_norm": 0.5388994812965393,
      "learning_rate": 0.00016299660667594814,
      "loss": 1.3686,
      "mean_token_accuracy": 0.7347927093505859,
      "num_tokens": 1367049.0,
      "step": 120
    },
    {
      "epoch": 1.5350318471337578,
      "grad_norm": 0.5220484137535095,
      "learning_rate": 0.00016234898018587337,
      "loss": 1.27,
      "mean_token_accuracy": 0.7542202472686768,
      "num_tokens": 1378847.0,
      "step": 121
    },
    {
      "epoch": 1.5477707006369426,
      "grad_norm": 0.553205132484436,
      "learning_rate": 0.00016169704701186527,
      "loss": 1.2828,
      "mean_token_accuracy": 0.75074902176857,
      "num_tokens": 1389763.0,
      "step": 122
    },
    {
      "epoch": 1.5605095541401273,
      "grad_norm": 0.5726351141929626,
      "learning_rate": 0.00016104085218545633,
      "loss": 1.2323,
      "mean_token_accuracy": 0.7563920617103577,
      "num_tokens": 1401158.0,
      "step": 123
    },
    {
      "epoch": 1.573248407643312,
      "grad_norm": 0.5743599534034729,
      "learning_rate": 0.00016038044103254775,
      "loss": 1.248,
      "mean_token_accuracy": 0.7553873062133789,
      "num_tokens": 1412660.0,
      "step": 124
    },
    {
      "epoch": 1.5859872611464967,
      "grad_norm": 0.6019836664199829,
      "learning_rate": 0.00015971585917027862,
      "loss": 1.2215,
      "mean_token_accuracy": 0.7611487507820129,
      "num_tokens": 1424264.0,
      "step": 125
    },
    {
      "epoch": 1.5987261146496814,
      "grad_norm": 0.5834748148918152,
      "learning_rate": 0.00015904715250387498,
      "loss": 1.3446,
      "mean_token_accuracy": 0.7405554950237274,
      "num_tokens": 1435491.0,
      "step": 126
    },
    {
      "epoch": 1.611464968152866,
      "grad_norm": 0.5433663725852966,
      "learning_rate": 0.000158374367223479,
      "loss": 1.261,
      "mean_token_accuracy": 0.7504373490810394,
      "num_tokens": 1446732.0,
      "step": 127
    },
    {
      "epoch": 1.6242038216560508,
      "grad_norm": 0.5282524824142456,
      "learning_rate": 0.0001576975498009583,
      "loss": 1.3273,
      "mean_token_accuracy": 0.7386387884616852,
      "num_tokens": 1458906.0,
      "step": 128
    },
    {
      "epoch": 1.6369426751592355,
      "grad_norm": 0.6508763432502747,
      "learning_rate": 0.0001570167469866962,
      "loss": 1.2123,
      "mean_token_accuracy": 0.7605562210083008,
      "num_tokens": 1469994.0,
      "step": 129
    },
    {
      "epoch": 1.6496815286624202,
      "grad_norm": 0.5391885638237,
      "learning_rate": 0.0001563320058063622,
      "loss": 1.1396,
      "mean_token_accuracy": 0.7745244204998016,
      "num_tokens": 1480937.0,
      "step": 130
    },
    {
      "epoch": 1.662420382165605,
      "grad_norm": 0.5844919085502625,
      "learning_rate": 0.00015564337355766412,
      "loss": 1.2875,
      "mean_token_accuracy": 0.7473530173301697,
      "num_tokens": 1492634.0,
      "step": 131
    },
    {
      "epoch": 1.6751592356687897,
      "grad_norm": 0.7106832265853882,
      "learning_rate": 0.0001549508978070806,
      "loss": 1.2713,
      "mean_token_accuracy": 0.7540490329265594,
      "num_tokens": 1503864.0,
      "step": 132
    },
    {
      "epoch": 1.6878980891719744,
      "grad_norm": 0.6007422804832458,
      "learning_rate": 0.00015425462638657595,
      "loss": 1.234,
      "mean_token_accuracy": 0.7583559155464172,
      "num_tokens": 1515227.0,
      "step": 133
    },
    {
      "epoch": 1.700636942675159,
      "grad_norm": 0.5394821763038635,
      "learning_rate": 0.00015355460739029586,
      "loss": 1.2763,
      "mean_token_accuracy": 0.7525887191295624,
      "num_tokens": 1526325.0,
      "step": 134
    },
    {
      "epoch": 1.7133757961783438,
      "grad_norm": 0.5472612380981445,
      "learning_rate": 0.00015285088917124556,
      "loss": 1.2394,
      "mean_token_accuracy": 0.7530220746994019,
      "num_tokens": 1538100.0,
      "step": 135
    },
    {
      "epoch": 1.7261146496815285,
      "grad_norm": 0.5598124265670776,
      "learning_rate": 0.0001521435203379498,
      "loss": 1.2431,
      "mean_token_accuracy": 0.7546306550502777,
      "num_tokens": 1549283.0,
      "step": 136
    },
    {
      "epoch": 1.7388535031847132,
      "grad_norm": 0.5618515610694885,
      "learning_rate": 0.00015143254975109538,
      "loss": 1.2521,
      "mean_token_accuracy": 0.7504317760467529,
      "num_tokens": 1560870.0,
      "step": 137
    },
    {
      "epoch": 1.7515923566878981,
      "grad_norm": 0.628064751625061,
      "learning_rate": 0.0001507180265201559,
      "loss": 1.235,
      "mean_token_accuracy": 0.7568688988685608,
      "num_tokens": 1572135.0,
      "step": 138
    },
    {
      "epoch": 1.7643312101910829,
      "grad_norm": 0.6592928171157837,
      "learning_rate": 0.00015000000000000001,
      "loss": 1.241,
      "mean_token_accuracy": 0.757856547832489,
      "num_tokens": 1583354.0,
      "step": 139
    },
    {
      "epoch": 1.7770700636942676,
      "grad_norm": 0.590002179145813,
      "learning_rate": 0.00014927851978748178,
      "loss": 1.1664,
      "mean_token_accuracy": 0.7673115134239197,
      "num_tokens": 1594573.0,
      "step": 140
    },
    {
      "epoch": 1.7898089171974523,
      "grad_norm": 0.5977404117584229,
      "learning_rate": 0.00014855363571801523,
      "loss": 1.2367,
      "mean_token_accuracy": 0.7607390582561493,
      "num_tokens": 1606088.0,
      "step": 141
    },
    {
      "epoch": 1.802547770700637,
      "grad_norm": 0.5965246558189392,
      "learning_rate": 0.00014782539786213183,
      "loss": 1.2311,
      "mean_token_accuracy": 0.7584077715873718,
      "num_tokens": 1617494.0,
      "step": 142
    },
    {
      "epoch": 1.8152866242038217,
      "grad_norm": 0.6000455021858215,
      "learning_rate": 0.00014709385652202203,
      "loss": 1.2735,
      "mean_token_accuracy": 0.7472036778926849,
      "num_tokens": 1629090.0,
      "step": 143
    },
    {
      "epoch": 1.8280254777070064,
      "grad_norm": 0.5743783116340637,
      "learning_rate": 0.00014635906222806058,
      "loss": 1.3982,
      "mean_token_accuracy": 0.7301271557807922,
      "num_tokens": 1641022.0,
      "step": 144
    },
    {
      "epoch": 1.8407643312101911,
      "grad_norm": 0.5715049505233765,
      "learning_rate": 0.0001456210657353163,
      "loss": 1.2934,
      "mean_token_accuracy": 0.7475639283657074,
      "num_tokens": 1652480.0,
      "step": 145
    },
    {
      "epoch": 1.8535031847133758,
      "grad_norm": 0.5563356280326843,
      "learning_rate": 0.00014487991802004623,
      "loss": 1.366,
      "mean_token_accuracy": 0.7382784187793732,
      "num_tokens": 1664089.0,
      "step": 146
    },
    {
      "epoch": 1.8662420382165605,
      "grad_norm": 0.559946596622467,
      "learning_rate": 0.0001441356702761744,
      "loss": 1.2647,
      "mean_token_accuracy": 0.7521970272064209,
      "num_tokens": 1675743.0,
      "step": 147
    },
    {
      "epoch": 1.8789808917197452,
      "grad_norm": 0.577804446220398,
      "learning_rate": 0.00014338837391175582,
      "loss": 1.2639,
      "mean_token_accuracy": 0.7574202716350555,
      "num_tokens": 1686909.0,
      "step": 148
    },
    {
      "epoch": 1.89171974522293,
      "grad_norm": 0.5892558097839355,
      "learning_rate": 0.0001426380805454254,
      "loss": 1.2246,
      "mean_token_accuracy": 0.7640581727027893,
      "num_tokens": 1698422.0,
      "step": 149
    },
    {
      "epoch": 1.9044585987261147,
      "grad_norm": 0.6493586897850037,
      "learning_rate": 0.0001418848420028325,
      "loss": 1.2171,
      "mean_token_accuracy": 0.7571829855442047,
      "num_tokens": 1709941.0,
      "step": 150
    },
    {
      "epoch": 1.9171974522292994,
      "grad_norm": 0.5959882736206055,
      "learning_rate": 0.00014112871031306119,
      "loss": 1.206,
      "mean_token_accuracy": 0.7620928287506104,
      "num_tokens": 1721542.0,
      "step": 151
    },
    {
      "epoch": 1.929936305732484,
      "grad_norm": 0.5942161083221436,
      "learning_rate": 0.00014036973770503624,
      "loss": 1.1675,
      "mean_token_accuracy": 0.7734838724136353,
      "num_tokens": 1732791.0,
      "step": 152
    },
    {
      "epoch": 1.9426751592356688,
      "grad_norm": 0.5867611169815063,
      "learning_rate": 0.0001396079766039157,
      "loss": 1.2261,
      "mean_token_accuracy": 0.7544828951358795,
      "num_tokens": 1743896.0,
      "step": 153
    },
    {
      "epoch": 1.9554140127388535,
      "grad_norm": 0.5893647074699402,
      "learning_rate": 0.00013884347962746948,
      "loss": 1.305,
      "mean_token_accuracy": 0.7446670830249786,
      "num_tokens": 1755279.0,
      "step": 154
    },
    {
      "epoch": 1.9681528662420382,
      "grad_norm": 0.6163603067398071,
      "learning_rate": 0.00013807629958244498,
      "loss": 1.2738,
      "mean_token_accuracy": 0.7518047988414764,
      "num_tokens": 1766674.0,
      "step": 155
    },
    {
      "epoch": 1.980891719745223,
      "grad_norm": 0.5853977799415588,
      "learning_rate": 0.0001373064894609194,
      "loss": 1.1963,
      "mean_token_accuracy": 0.7708700001239777,
      "num_tokens": 1777748.0,
      "step": 156
    },
    {
      "epoch": 1.9936305732484076,
      "grad_norm": 0.5478814244270325,
      "learning_rate": 0.00013653410243663952,
      "loss": 1.1939,
      "mean_token_accuracy": 0.7640825510025024,
      "num_tokens": 1788830.0,
      "step": 157
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8762192130088806,
      "learning_rate": 0.0001357591918613486,
      "loss": 1.008,
      "mean_token_accuracy": 0.795324981212616,
      "num_tokens": 1794124.0,
      "step": 158
    },
    {
      "epoch": 2.0127388535031847,
      "grad_norm": 0.6148726940155029,
      "learning_rate": 0.0001349818112611015,
      "loss": 0.8962,
      "mean_token_accuracy": 0.8148521482944489,
      "num_tokens": 1805636.0,
      "step": 159
    },
    {
      "epoch": 2.0254777070063694,
      "grad_norm": 1.515809416770935,
      "learning_rate": 0.00013420201433256689,
      "loss": 0.9681,
      "mean_token_accuracy": 0.7990589141845703,
      "num_tokens": 1817299.0,
      "step": 160
    },
    {
      "epoch": 2.038216560509554,
      "grad_norm": 0.7491560578346252,
      "learning_rate": 0.00013341985493931877,
      "loss": 0.8922,
      "mean_token_accuracy": 0.815070241689682,
      "num_tokens": 1828936.0,
      "step": 161
    },
    {
      "epoch": 2.050955414012739,
      "grad_norm": 0.6307762861251831,
      "learning_rate": 0.0001326353871081156,
      "loss": 0.8704,
      "mean_token_accuracy": 0.818681925535202,
      "num_tokens": 1840132.0,
      "step": 162
    },
    {
      "epoch": 2.0636942675159236,
      "grad_norm": 0.627448320388794,
      "learning_rate": 0.00013184866502516845,
      "loss": 0.8727,
      "mean_token_accuracy": 0.8210729658603668,
      "num_tokens": 1851489.0,
      "step": 163
    },
    {
      "epoch": 2.0764331210191083,
      "grad_norm": 0.6649007797241211,
      "learning_rate": 0.00013105974303239838,
      "loss": 0.8367,
      "mean_token_accuracy": 0.8308760523796082,
      "num_tokens": 1862876.0,
      "step": 164
    },
    {
      "epoch": 2.089171974522293,
      "grad_norm": 0.7197876572608948,
      "learning_rate": 0.0001302686756236826,
      "loss": 0.7737,
      "mean_token_accuracy": 0.8400779962539673,
      "num_tokens": 1874177.0,
      "step": 165
    },
    {
      "epoch": 2.1019108280254777,
      "grad_norm": 0.7220533490180969,
      "learning_rate": 0.00012947551744109043,
      "loss": 0.8546,
      "mean_token_accuracy": 0.822653740644455,
      "num_tokens": 1885524.0,
      "step": 166
    },
    {
      "epoch": 2.1146496815286624,
      "grad_norm": 0.7428185343742371,
      "learning_rate": 0.00012868032327110904,
      "loss": 0.8556,
      "mean_token_accuracy": 0.8203161954879761,
      "num_tokens": 1897169.0,
      "step": 167
    },
    {
      "epoch": 2.127388535031847,
      "grad_norm": 0.6323121190071106,
      "learning_rate": 0.00012788314804085903,
      "loss": 0.7231,
      "mean_token_accuracy": 0.8477356731891632,
      "num_tokens": 1908703.0,
      "step": 168
    },
    {
      "epoch": 2.140127388535032,
      "grad_norm": 0.7001054286956787,
      "learning_rate": 0.00012708404681430053,
      "loss": 0.8976,
      "mean_token_accuracy": 0.8160674273967743,
      "num_tokens": 1919695.0,
      "step": 169
    },
    {
      "epoch": 2.1528662420382165,
      "grad_norm": 0.6420326232910156,
      "learning_rate": 0.00012628307478842953,
      "loss": 0.8451,
      "mean_token_accuracy": 0.8224026262760162,
      "num_tokens": 1930961.0,
      "step": 170
    },
    {
      "epoch": 2.1656050955414012,
      "grad_norm": 0.7011420130729675,
      "learning_rate": 0.0001254802872894655,
      "loss": 0.8627,
      "mean_token_accuracy": 0.8202202320098877,
      "num_tokens": 1942462.0,
      "step": 171
    },
    {
      "epoch": 2.178343949044586,
      "grad_norm": 0.6822834610939026,
      "learning_rate": 0.00012467573976902935,
      "loss": 0.7987,
      "mean_token_accuracy": 0.8363339900970459,
      "num_tokens": 1953848.0,
      "step": 172
    },
    {
      "epoch": 2.1910828025477707,
      "grad_norm": 0.7396888732910156,
      "learning_rate": 0.0001238694878003138,
      "loss": 0.8632,
      "mean_token_accuracy": 0.8247162401676178,
      "num_tokens": 1965634.0,
      "step": 173
    },
    {
      "epoch": 2.2038216560509554,
      "grad_norm": 0.6941153407096863,
      "learning_rate": 0.00012306158707424403,
      "loss": 0.8203,
      "mean_token_accuracy": 0.8318359553813934,
      "num_tokens": 1977633.0,
      "step": 174
    },
    {
      "epoch": 2.21656050955414,
      "grad_norm": 0.7376344203948975,
      "learning_rate": 0.00012225209339563145,
      "loss": 0.8813,
      "mean_token_accuracy": 0.816738098859787,
      "num_tokens": 1989375.0,
      "step": 175
    },
    {
      "epoch": 2.229299363057325,
      "grad_norm": 0.6955387592315674,
      "learning_rate": 0.00012144106267931876,
      "loss": 0.8062,
      "mean_token_accuracy": 0.8336440622806549,
      "num_tokens": 2000350.0,
      "step": 176
    },
    {
      "epoch": 2.2420382165605095,
      "grad_norm": 0.6509573459625244,
      "learning_rate": 0.00012062855094631778,
      "loss": 0.7839,
      "mean_token_accuracy": 0.8355735242366791,
      "num_tokens": 2011821.0,
      "step": 177
    },
    {
      "epoch": 2.254777070063694,
      "grad_norm": 0.7110678553581238,
      "learning_rate": 0.00011981461431993977,
      "loss": 0.8643,
      "mean_token_accuracy": 0.8200065493583679,
      "num_tokens": 2023398.0,
      "step": 178
    },
    {
      "epoch": 2.267515923566879,
      "grad_norm": 0.7298818230628967,
      "learning_rate": 0.00011899930902191902,
      "loss": 0.8742,
      "mean_token_accuracy": 0.821577250957489,
      "num_tokens": 2035102.0,
      "step": 179
    },
    {
      "epoch": 2.2802547770700636,
      "grad_norm": 0.6927684545516968,
      "learning_rate": 0.00011818269136852909,
      "loss": 0.8681,
      "mean_token_accuracy": 0.8210790157318115,
      "num_tokens": 2046897.0,
      "step": 180
    },
    {
      "epoch": 2.2929936305732483,
      "grad_norm": 0.6942738890647888,
      "learning_rate": 0.00011736481776669306,
      "loss": 0.7765,
      "mean_token_accuracy": 0.8394728899002075,
      "num_tokens": 2057980.0,
      "step": 181
    },
    {
      "epoch": 2.305732484076433,
      "grad_norm": 0.6812620162963867,
      "learning_rate": 0.00011654574471008713,
      "loss": 0.8486,
      "mean_token_accuracy": 0.8229053914546967,
      "num_tokens": 2069683.0,
      "step": 182
    },
    {
      "epoch": 2.3184713375796178,
      "grad_norm": 0.7021245360374451,
      "learning_rate": 0.00011572552877523854,
      "loss": 0.8407,
      "mean_token_accuracy": 0.823371946811676,
      "num_tokens": 2081025.0,
      "step": 183
    },
    {
      "epoch": 2.3312101910828025,
      "grad_norm": 0.6945245265960693,
      "learning_rate": 0.00011490422661761744,
      "loss": 0.7698,
      "mean_token_accuracy": 0.8421520590782166,
      "num_tokens": 2092607.0,
      "step": 184
    },
    {
      "epoch": 2.343949044585987,
      "grad_norm": 0.7140945196151733,
      "learning_rate": 0.00011408189496772368,
      "loss": 0.7344,
      "mean_token_accuracy": 0.8482842147350311,
      "num_tokens": 2103628.0,
      "step": 185
    },
    {
      "epoch": 2.356687898089172,
      "grad_norm": 0.8129433989524841,
      "learning_rate": 0.00011325859062716795,
      "loss": 0.8266,
      "mean_token_accuracy": 0.8301917314529419,
      "num_tokens": 2114651.0,
      "step": 186
    },
    {
      "epoch": 2.3694267515923566,
      "grad_norm": 0.790624737739563,
      "learning_rate": 0.00011243437046474853,
      "loss": 0.8995,
      "mean_token_accuracy": 0.8124870359897614,
      "num_tokens": 2126430.0,
      "step": 187
    },
    {
      "epoch": 2.3821656050955413,
      "grad_norm": 0.6942800283432007,
      "learning_rate": 0.00011160929141252303,
      "loss": 0.8086,
      "mean_token_accuracy": 0.8364989161491394,
      "num_tokens": 2137274.0,
      "step": 188
    },
    {
      "epoch": 2.394904458598726,
      "grad_norm": 0.7138471603393555,
      "learning_rate": 0.00011078341046187589,
      "loss": 0.8914,
      "mean_token_accuracy": 0.8123456239700317,
      "num_tokens": 2148697.0,
      "step": 189
    },
    {
      "epoch": 2.4076433121019107,
      "grad_norm": 0.684965193271637,
      "learning_rate": 0.00010995678465958168,
      "loss": 0.8619,
      "mean_token_accuracy": 0.8225953578948975,
      "num_tokens": 2160613.0,
      "step": 190
    },
    {
      "epoch": 2.4203821656050954,
      "grad_norm": 0.7347207069396973,
      "learning_rate": 0.00010912947110386484,
      "loss": 0.8156,
      "mean_token_accuracy": 0.8292295336723328,
      "num_tokens": 2172303.0,
      "step": 191
    },
    {
      "epoch": 2.43312101910828,
      "grad_norm": 0.7004370093345642,
      "learning_rate": 0.00010830152694045552,
      "loss": 0.8175,
      "mean_token_accuracy": 0.8329605758190155,
      "num_tokens": 2183712.0,
      "step": 192
    },
    {
      "epoch": 2.445859872611465,
      "grad_norm": 0.842786967754364,
      "learning_rate": 0.00010747300935864243,
      "loss": 0.8424,
      "mean_token_accuracy": 0.82681804895401,
      "num_tokens": 2195201.0,
      "step": 193
    },
    {
      "epoch": 2.4585987261146496,
      "grad_norm": 0.7165960073471069,
      "learning_rate": 0.00010664397558732244,
      "loss": 0.7305,
      "mean_token_accuracy": 0.8466060161590576,
      "num_tokens": 2206348.0,
      "step": 194
    },
    {
      "epoch": 2.4713375796178343,
      "grad_norm": 0.7170045375823975,
      "learning_rate": 0.00010581448289104758,
      "loss": 0.8841,
      "mean_token_accuracy": 0.817249596118927,
      "num_tokens": 2218307.0,
      "step": 195
    },
    {
      "epoch": 2.484076433121019,
      "grad_norm": 0.6497987508773804,
      "learning_rate": 0.00010498458856606972,
      "loss": 0.7677,
      "mean_token_accuracy": 0.8430719673633575,
      "num_tokens": 2229766.0,
      "step": 196
    },
    {
      "epoch": 2.4968152866242037,
      "grad_norm": 0.6531058549880981,
      "learning_rate": 0.00010415434993638269,
      "loss": 0.7116,
      "mean_token_accuracy": 0.8531897962093353,
      "num_tokens": 2240599.0,
      "step": 197
    },
    {
      "epoch": 2.5095541401273884,
      "grad_norm": 0.6574857831001282,
      "learning_rate": 0.00010332382434976266,
      "loss": 0.7306,
      "mean_token_accuracy": 0.8475680351257324,
      "num_tokens": 2252398.0,
      "step": 198
    },
    {
      "epoch": 2.522292993630573,
      "grad_norm": 0.7400979995727539,
      "learning_rate": 0.0001024930691738073,
      "loss": 0.7793,
      "mean_token_accuracy": 0.840068131685257,
      "num_tokens": 2264039.0,
      "step": 199
    },
    {
      "epoch": 2.535031847133758,
      "grad_norm": 0.7397539019584656,
      "learning_rate": 0.00010166214179197264,
      "loss": 0.8102,
      "mean_token_accuracy": 0.8325523436069489,
      "num_tokens": 2275605.0,
      "step": 200
    },
    {
      "epoch": 2.5477707006369426,
      "grad_norm": 0.8763604164123535,
      "learning_rate": 0.00010083109959960973,
      "loss": 0.8838,
      "mean_token_accuracy": 0.8195656836032867,
      "num_tokens": 2287149.0,
      "step": 201
    },
    {
      "epoch": 2.5605095541401273,
      "grad_norm": 0.7435045838356018,
      "learning_rate": 0.0001,
      "loss": 0.8099,
      "mean_token_accuracy": 0.834762305021286,
      "num_tokens": 2298149.0,
      "step": 202
    },
    {
      "epoch": 2.573248407643312,
      "grad_norm": 0.7856142520904541,
      "learning_rate": 9.916890040039031e-05,
      "loss": 0.8285,
      "mean_token_accuracy": 0.8306257724761963,
      "num_tokens": 2309730.0,
      "step": 203
    },
    {
      "epoch": 2.5859872611464967,
      "grad_norm": 0.6982733607292175,
      "learning_rate": 9.833785820802739e-05,
      "loss": 0.7267,
      "mean_token_accuracy": 0.846727728843689,
      "num_tokens": 2320654.0,
      "step": 204
    },
    {
      "epoch": 2.5987261146496814,
      "grad_norm": 0.6996532082557678,
      "learning_rate": 9.750693082619273e-05,
      "loss": 0.755,
      "mean_token_accuracy": 0.843133270740509,
      "num_tokens": 2332220.0,
      "step": 205
    },
    {
      "epoch": 2.611464968152866,
      "grad_norm": 0.7252086997032166,
      "learning_rate": 9.667617565023735e-05,
      "loss": 0.7711,
      "mean_token_accuracy": 0.8427746593952179,
      "num_tokens": 2343588.0,
      "step": 206
    },
    {
      "epoch": 2.624203821656051,
      "grad_norm": 0.8143007159233093,
      "learning_rate": 9.584565006361734e-05,
      "loss": 0.8082,
      "mean_token_accuracy": 0.8367211222648621,
      "num_tokens": 2355041.0,
      "step": 207
    },
    {
      "epoch": 2.6369426751592355,
      "grad_norm": 0.8434510231018066,
      "learning_rate": 9.501541143393028e-05,
      "loss": 0.8303,
      "mean_token_accuracy": 0.824932187795639,
      "num_tokens": 2366049.0,
      "step": 208
    },
    {
      "epoch": 2.6496815286624202,
      "grad_norm": 0.7303879857063293,
      "learning_rate": 9.418551710895243e-05,
      "loss": 0.802,
      "mean_token_accuracy": 0.8343527615070343,
      "num_tokens": 2377894.0,
      "step": 209
    },
    {
      "epoch": 2.662420382165605,
      "grad_norm": 0.685028076171875,
      "learning_rate": 9.335602441267759e-05,
      "loss": 0.7587,
      "mean_token_accuracy": 0.8418575525283813,
      "num_tokens": 2389701.0,
      "step": 210
    },
    {
      "epoch": 2.6751592356687897,
      "grad_norm": 0.7157097458839417,
      "learning_rate": 9.252699064135758e-05,
      "loss": 0.7537,
      "mean_token_accuracy": 0.8456466197967529,
      "num_tokens": 2400854.0,
      "step": 211
    },
    {
      "epoch": 2.6878980891719744,
      "grad_norm": 0.7124730944633484,
      "learning_rate": 9.169847305954447e-05,
      "loss": 0.8539,
      "mean_token_accuracy": 0.8244422972202301,
      "num_tokens": 2412586.0,
      "step": 212
    },
    {
      "epoch": 2.700636942675159,
      "grad_norm": 0.6919587254524231,
      "learning_rate": 9.087052889613518e-05,
      "loss": 0.7594,
      "mean_token_accuracy": 0.8418782651424408,
      "num_tokens": 2423882.0,
      "step": 213
    },
    {
      "epoch": 2.713375796178344,
      "grad_norm": 0.740878164768219,
      "learning_rate": 9.004321534041835e-05,
      "loss": 0.6822,
      "mean_token_accuracy": 0.8537989258766174,
      "num_tokens": 2434645.0,
      "step": 214
    },
    {
      "epoch": 2.7261146496815285,
      "grad_norm": 0.7817885875701904,
      "learning_rate": 8.921658953812415e-05,
      "loss": 0.7664,
      "mean_token_accuracy": 0.8422251641750336,
      "num_tokens": 2445879.0,
      "step": 215
    },
    {
      "epoch": 2.738853503184713,
      "grad_norm": 0.802693247795105,
      "learning_rate": 8.839070858747697e-05,
      "loss": 0.7031,
      "mean_token_accuracy": 0.8553369045257568,
      "num_tokens": 2457114.0,
      "step": 216
    },
    {
      "epoch": 2.7515923566878984,
      "grad_norm": 0.7690209150314331,
      "learning_rate": 8.756562953525152e-05,
      "loss": 0.7587,
      "mean_token_accuracy": 0.8424797356128693,
      "num_tokens": 2468624.0,
      "step": 217
    },
    {
      "epoch": 2.7643312101910826,
      "grad_norm": 0.7805998921394348,
      "learning_rate": 8.674140937283208e-05,
      "loss": 0.7921,
      "mean_token_accuracy": 0.8381544649600983,
      "num_tokens": 2479981.0,
      "step": 218
    },
    {
      "epoch": 2.777070063694268,
      "grad_norm": 0.6912605166435242,
      "learning_rate": 8.591810503227635e-05,
      "loss": 0.6652,
      "mean_token_accuracy": 0.861592561006546,
      "num_tokens": 2491405.0,
      "step": 219
    },
    {
      "epoch": 2.789808917197452,
      "grad_norm": 0.7387929558753967,
      "learning_rate": 8.509577338238255e-05,
      "loss": 0.8042,
      "mean_token_accuracy": 0.8338701725006104,
      "num_tokens": 2502813.0,
      "step": 220
    },
    {
      "epoch": 2.802547770700637,
      "grad_norm": 0.7098203301429749,
      "learning_rate": 8.427447122476148e-05,
      "loss": 0.7327,
      "mean_token_accuracy": 0.8484758734703064,
      "num_tokens": 2514190.0,
      "step": 221
    },
    {
      "epoch": 2.8152866242038215,
      "grad_norm": 0.7439514994621277,
      "learning_rate": 8.345425528991288e-05,
      "loss": 0.6598,
      "mean_token_accuracy": 0.860679030418396,
      "num_tokens": 2525108.0,
      "step": 222
    },
    {
      "epoch": 2.8280254777070066,
      "grad_norm": 0.7285000681877136,
      "learning_rate": 8.263518223330697e-05,
      "loss": 0.7639,
      "mean_token_accuracy": 0.842808336019516,
      "num_tokens": 2536505.0,
      "step": 223
    },
    {
      "epoch": 2.840764331210191,
      "grad_norm": 0.7520717978477478,
      "learning_rate": 8.181730863147093e-05,
      "loss": 0.796,
      "mean_token_accuracy": 0.8394459187984467,
      "num_tokens": 2548300.0,
      "step": 224
    },
    {
      "epoch": 2.853503184713376,
      "grad_norm": 0.7541860342025757,
      "learning_rate": 8.100069097808103e-05,
      "loss": 0.788,
      "mean_token_accuracy": 0.8363643884658813,
      "num_tokens": 2559823.0,
      "step": 225
    },
    {
      "epoch": 2.8662420382165603,
      "grad_norm": 0.7908371090888977,
      "learning_rate": 8.018538568006027e-05,
      "loss": 0.7885,
      "mean_token_accuracy": 0.8349603712558746,
      "num_tokens": 2571347.0,
      "step": 226
    },
    {
      "epoch": 2.8789808917197455,
      "grad_norm": 0.7263608574867249,
      "learning_rate": 7.937144905368226e-05,
      "loss": 0.7057,
      "mean_token_accuracy": 0.8521650433540344,
      "num_tokens": 2582905.0,
      "step": 227
    },
    {
      "epoch": 2.8917197452229297,
      "grad_norm": 0.7562076449394226,
      "learning_rate": 7.855893732068125e-05,
      "loss": 0.8208,
      "mean_token_accuracy": 0.8311346769332886,
      "num_tokens": 2594285.0,
      "step": 228
    },
    {
      "epoch": 2.904458598726115,
      "grad_norm": 0.7348179221153259,
      "learning_rate": 7.774790660436858e-05,
      "loss": 0.7344,
      "mean_token_accuracy": 0.8490732312202454,
      "num_tokens": 2605746.0,
      "step": 229
    },
    {
      "epoch": 2.917197452229299,
      "grad_norm": 0.7636176943778992,
      "learning_rate": 7.693841292575598e-05,
      "loss": 0.8584,
      "mean_token_accuracy": 0.8263067901134491,
      "num_tokens": 2617280.0,
      "step": 230
    },
    {
      "epoch": 2.9299363057324843,
      "grad_norm": 0.7639762759208679,
      "learning_rate": 7.613051219968623e-05,
      "loss": 0.7856,
      "mean_token_accuracy": 0.8378620445728302,
      "num_tokens": 2629030.0,
      "step": 231
    },
    {
      "epoch": 2.9426751592356686,
      "grad_norm": 0.7438153028488159,
      "learning_rate": 7.532426023097063e-05,
      "loss": 0.7406,
      "mean_token_accuracy": 0.8455101847648621,
      "num_tokens": 2640001.0,
      "step": 232
    },
    {
      "epoch": 2.9554140127388537,
      "grad_norm": 0.7770811915397644,
      "learning_rate": 7.451971271053455e-05,
      "loss": 0.8496,
      "mean_token_accuracy": 0.8213350176811218,
      "num_tokens": 2651498.0,
      "step": 233
    },
    {
      "epoch": 2.968152866242038,
      "grad_norm": 0.7544724345207214,
      "learning_rate": 7.371692521157048e-05,
      "loss": 0.7548,
      "mean_token_accuracy": 0.8470020294189453,
      "num_tokens": 2662618.0,
      "step": 234
    },
    {
      "epoch": 2.980891719745223,
      "grad_norm": 0.7707628011703491,
      "learning_rate": 7.291595318569951e-05,
      "loss": 0.794,
      "mean_token_accuracy": 0.8322769701480865,
      "num_tokens": 2674269.0,
      "step": 235
    },
    {
      "epoch": 2.9936305732484074,
      "grad_norm": 0.8508571982383728,
      "learning_rate": 7.211685195914097e-05,
      "loss": 0.7893,
      "mean_token_accuracy": 0.8407465815544128,
      "num_tokens": 2685723.0,
      "step": 236
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8508571982383728,
      "learning_rate": 7.131967672889101e-05,
      "loss": 0.5708,
      "mean_token_accuracy": 0.8794448971748352,
      "num_tokens": 2691520.0,
      "step": 237
    },
    {
      "epoch": 3.0127388535031847,
      "grad_norm": 1.1057546138763428,
      "learning_rate": 7.052448255890957e-05,
      "loss": 0.4499,
      "mean_token_accuracy": 0.9071509540081024,
      "num_tokens": 2702735.0,
      "step": 238
    },
    {
      "epoch": 3.0254777070063694,
      "grad_norm": 0.6588014960289001,
      "learning_rate": 6.973132437631742e-05,
      "loss": 0.4431,
      "mean_token_accuracy": 0.9075900018215179,
      "num_tokens": 2714184.0,
      "step": 239
    },
    {
      "epoch": 3.038216560509554,
      "grad_norm": 0.6847231388092041,
      "learning_rate": 6.894025696760163e-05,
      "loss": 0.414,
      "mean_token_accuracy": 0.911798506975174,
      "num_tokens": 2726186.0,
      "step": 240
    },
    {
      "epoch": 3.050955414012739,
      "grad_norm": 0.7595067620277405,
      "learning_rate": 6.815133497483157e-05,
      "loss": 0.4353,
      "mean_token_accuracy": 0.9072951972484589,
      "num_tokens": 2737492.0,
      "step": 241
    },
    {
      "epoch": 3.0636942675159236,
      "grad_norm": 0.9101142883300781,
      "learning_rate": 6.736461289188445e-05,
      "loss": 0.486,
      "mean_token_accuracy": 0.8956897854804993,
      "num_tokens": 2749284.0,
      "step": 242
    },
    {
      "epoch": 3.0764331210191083,
      "grad_norm": 1.0063393115997314,
      "learning_rate": 6.658014506068126e-05,
      "loss": 0.4754,
      "mean_token_accuracy": 0.9023579955101013,
      "num_tokens": 2760625.0,
      "step": 243
    },
    {
      "epoch": 3.089171974522293,
      "grad_norm": 0.7680675387382507,
      "learning_rate": 6.579798566743314e-05,
      "loss": 0.491,
      "mean_token_accuracy": 0.8949213922023773,
      "num_tokens": 2772261.0,
      "step": 244
    },
    {
      "epoch": 3.1019108280254777,
      "grad_norm": 0.6828211545944214,
      "learning_rate": 6.501818873889855e-05,
      "loss": 0.4538,
      "mean_token_accuracy": 0.9027508199214935,
      "num_tokens": 2783770.0,
      "step": 245
    },
    {
      "epoch": 3.1146496815286624,
      "grad_norm": 0.7493994832038879,
      "learning_rate": 6.424080813865138e-05,
      "loss": 0.4336,
      "mean_token_accuracy": 0.9114145934581757,
      "num_tokens": 2794903.0,
      "step": 246
    },
    {
      "epoch": 3.127388535031847,
      "grad_norm": 0.6365121603012085,
      "learning_rate": 6.34658975633605e-05,
      "loss": 0.4023,
      "mean_token_accuracy": 0.9125161468982697,
      "num_tokens": 2806179.0,
      "step": 247
    },
    {
      "epoch": 3.140127388535032,
      "grad_norm": 0.6832964420318604,
      "learning_rate": 6.269351053908061e-05,
      "loss": 0.4003,
      "mean_token_accuracy": 0.9133158922195435,
      "num_tokens": 2817415.0,
      "step": 248
    },
    {
      "epoch": 3.1528662420382165,
      "grad_norm": 0.6966664791107178,
      "learning_rate": 6.192370041755505e-05,
      "loss": 0.3681,
      "mean_token_accuracy": 0.9190676212310791,
      "num_tokens": 2828929.0,
      "step": 249
    },
    {
      "epoch": 3.1656050955414012,
      "grad_norm": 0.7440553903579712,
      "learning_rate": 6.115652037253053e-05,
      "loss": 0.4212,
      "mean_token_accuracy": 0.9078826904296875,
      "num_tokens": 2840532.0,
      "step": 250
    },
    {
      "epoch": 3.178343949044586,
      "grad_norm": 0.805282473564148,
      "learning_rate": 6.039202339608432e-05,
      "loss": 0.393,
      "mean_token_accuracy": 0.9137612283229828,
      "num_tokens": 2852314.0,
      "step": 251
    },
    {
      "epoch": 3.1910828025477707,
      "grad_norm": 0.8093756437301636,
      "learning_rate": 5.963026229496378e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8994007706642151,
      "num_tokens": 2863849.0,
      "step": 252
    },
    {
      "epoch": 3.2038216560509554,
      "grad_norm": 0.8103389143943787,
      "learning_rate": 5.887128968693887e-05,
      "loss": 0.3956,
      "mean_token_accuracy": 0.9146359264850616,
      "num_tokens": 2874972.0,
      "step": 253
    },
    {
      "epoch": 3.21656050955414,
      "grad_norm": 0.7377609014511108,
      "learning_rate": 5.8115157997167536e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.9041499197483063,
      "num_tokens": 2886716.0,
      "step": 254
    },
    {
      "epoch": 3.229299363057325,
      "grad_norm": 0.7186458110809326,
      "learning_rate": 5.736191945457463e-05,
      "loss": 0.4116,
      "mean_token_accuracy": 0.9132950901985168,
      "num_tokens": 2897985.0,
      "step": 255
    },
    {
      "epoch": 3.2420382165605095,
      "grad_norm": 0.6944336295127869,
      "learning_rate": 5.6611626088244194e-05,
      "loss": 0.4971,
      "mean_token_accuracy": 0.8925972282886505,
      "num_tokens": 2909541.0,
      "step": 256
    },
    {
      "epoch": 3.254777070063694,
      "grad_norm": 0.773350715637207,
      "learning_rate": 5.58643297238256e-05,
      "loss": 0.4191,
      "mean_token_accuracy": 0.909559965133667,
      "num_tokens": 2920862.0,
      "step": 257
    },
    {
      "epoch": 3.267515923566879,
      "grad_norm": 0.6871331930160522,
      "learning_rate": 5.5120081979953785e-05,
      "loss": 0.41,
      "mean_token_accuracy": 0.9117248952388763,
      "num_tokens": 2931948.0,
      "step": 258
    },
    {
      "epoch": 3.2802547770700636,
      "grad_norm": 0.7305521965026855,
      "learning_rate": 5.43789342646837e-05,
      "loss": 0.4134,
      "mean_token_accuracy": 0.9124027788639069,
      "num_tokens": 2943315.0,
      "step": 259
    },
    {
      "epoch": 3.2929936305732483,
      "grad_norm": 0.738551676273346,
      "learning_rate": 5.3640937771939436e-05,
      "loss": 0.4455,
      "mean_token_accuracy": 0.9042632281780243,
      "num_tokens": 2954251.0,
      "step": 260
    },
    {
      "epoch": 3.305732484076433,
      "grad_norm": 0.7478283047676086,
      "learning_rate": 5.290614347797802e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.8965472280979156,
      "num_tokens": 2966119.0,
      "step": 261
    },
    {
      "epoch": 3.3184713375796178,
      "grad_norm": 0.7674943208694458,
      "learning_rate": 5.217460213786821e-05,
      "loss": 0.3871,
      "mean_token_accuracy": 0.9164425730705261,
      "num_tokens": 2977756.0,
      "step": 262
    },
    {
      "epoch": 3.3312101910828025,
      "grad_norm": 0.7106198668479919,
      "learning_rate": 5.1446364281984774e-05,
      "loss": 0.3952,
      "mean_token_accuracy": 0.9167567789554596,
      "num_tokens": 2989020.0,
      "step": 263
    },
    {
      "epoch": 3.343949044585987,
      "grad_norm": 0.776825487613678,
      "learning_rate": 5.072148021251821e-05,
      "loss": 0.3951,
      "mean_token_accuracy": 0.9141735136508942,
      "num_tokens": 3000498.0,
      "step": 264
    },
    {
      "epoch": 3.356687898089172,
      "grad_norm": 0.7228910326957703,
      "learning_rate": 5.000000000000002e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8982515037059784,
      "num_tokens": 3012283.0,
      "step": 265
    },
    {
      "epoch": 3.3694267515923566,
      "grad_norm": 0.8001947999000549,
      "learning_rate": 4.92819734798441e-05,
      "loss": 0.3931,
      "mean_token_accuracy": 0.9135704636573792,
      "num_tokens": 3023734.0,
      "step": 266
    },
    {
      "epoch": 3.3821656050955413,
      "grad_norm": 0.7749711275100708,
      "learning_rate": 4.856745024890466e-05,
      "loss": 0.4205,
      "mean_token_accuracy": 0.9102110266685486,
      "num_tokens": 3035422.0,
      "step": 267
    },
    {
      "epoch": 3.394904458598726,
      "grad_norm": 0.6763160228729248,
      "learning_rate": 4.78564796620502e-05,
      "loss": 0.3824,
      "mean_token_accuracy": 0.9176193177700043,
      "num_tokens": 3047227.0,
      "step": 268
    },
    {
      "epoch": 3.4076433121019107,
      "grad_norm": 0.733139157295227,
      "learning_rate": 4.7149110828754464e-05,
      "loss": 0.4641,
      "mean_token_accuracy": 0.9008893370628357,
      "num_tokens": 3058515.0,
      "step": 269
    },
    {
      "epoch": 3.4203821656050954,
      "grad_norm": 0.7419499754905701,
      "learning_rate": 4.644539260970416e-05,
      "loss": 0.4134,
      "mean_token_accuracy": 0.9126076698303223,
      "num_tokens": 3070154.0,
      "step": 270
    },
    {
      "epoch": 3.43312101910828,
      "grad_norm": 0.7160090804100037,
      "learning_rate": 4.574537361342407e-05,
      "loss": 0.4642,
      "mean_token_accuracy": 0.9061474502086639,
      "num_tokens": 3081238.0,
      "step": 271
    },
    {
      "epoch": 3.445859872611465,
      "grad_norm": 0.7175310254096985,
      "learning_rate": 4.50491021929194e-05,
      "loss": 0.3756,
      "mean_token_accuracy": 0.918995589017868,
      "num_tokens": 3092931.0,
      "step": 272
    },
    {
      "epoch": 3.4585987261146496,
      "grad_norm": 0.7008662223815918,
      "learning_rate": 4.435662644233594e-05,
      "loss": 0.392,
      "mean_token_accuracy": 0.9195187389850616,
      "num_tokens": 3104621.0,
      "step": 273
    },
    {
      "epoch": 3.4713375796178343,
      "grad_norm": 0.7543473839759827,
      "learning_rate": 4.3667994193637796e-05,
      "loss": 0.3809,
      "mean_token_accuracy": 0.9197768867015839,
      "num_tokens": 3116470.0,
      "step": 274
    },
    {
      "epoch": 3.484076433121019,
      "grad_norm": 0.6894095540046692,
      "learning_rate": 4.298325301330383e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.9104044735431671,
      "num_tokens": 3127792.0,
      "step": 275
    },
    {
      "epoch": 3.4968152866242037,
      "grad_norm": 0.7606199979782104,
      "learning_rate": 4.23024501990417e-05,
      "loss": 0.3707,
      "mean_token_accuracy": 0.9214757680892944,
      "num_tokens": 3139330.0,
      "step": 276
    },
    {
      "epoch": 3.5095541401273884,
      "grad_norm": 0.7206736207008362,
      "learning_rate": 4.1625632776521037e-05,
      "loss": 0.3894,
      "mean_token_accuracy": 0.9143988788127899,
      "num_tokens": 3150599.0,
      "step": 277
    },
    {
      "epoch": 3.522292993630573,
      "grad_norm": 0.7822532057762146,
      "learning_rate": 4.095284749612503e-05,
      "loss": 0.4631,
      "mean_token_accuracy": 0.9017830193042755,
      "num_tokens": 3161908.0,
      "step": 278
    },
    {
      "epoch": 3.535031847133758,
      "grad_norm": 0.7029759287834167,
      "learning_rate": 4.028414082972141e-05,
      "loss": 0.3864,
      "mean_token_accuracy": 0.9187280237674713,
      "num_tokens": 3173218.0,
      "step": 279
    },
    {
      "epoch": 3.5477707006369426,
      "grad_norm": 0.7151777148246765,
      "learning_rate": 3.961955896745224e-05,
      "loss": 0.4131,
      "mean_token_accuracy": 0.9083968102931976,
      "num_tokens": 3184570.0,
      "step": 280
    },
    {
      "epoch": 3.5605095541401273,
      "grad_norm": 0.759780764579773,
      "learning_rate": 3.89591478145437e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.9122490584850311,
      "num_tokens": 3196499.0,
      "step": 281
    },
    {
      "epoch": 3.573248407643312,
      "grad_norm": 0.6918774843215942,
      "learning_rate": 3.8302952988134756e-05,
      "loss": 0.4089,
      "mean_token_accuracy": 0.9108126163482666,
      "num_tokens": 3207925.0,
      "step": 282
    },
    {
      "epoch": 3.5859872611464967,
      "grad_norm": 0.6682057976722717,
      "learning_rate": 3.7651019814126654e-05,
      "loss": 0.364,
      "mean_token_accuracy": 0.9240628480911255,
      "num_tokens": 3218971.0,
      "step": 283
    },
    {
      "epoch": 3.5987261146496814,
      "grad_norm": 0.7228002548217773,
      "learning_rate": 3.7003393324051874e-05,
      "loss": 0.4119,
      "mean_token_accuracy": 0.9100562930107117,
      "num_tokens": 3230456.0,
      "step": 284
    },
    {
      "epoch": 3.611464968152866,
      "grad_norm": 0.7470158338546753,
      "learning_rate": 3.6360118251963645e-05,
      "loss": 0.403,
      "mean_token_accuracy": 0.914577841758728,
      "num_tokens": 3241801.0,
      "step": 285
    },
    {
      "epoch": 3.624203821656051,
      "grad_norm": 0.7659094333648682,
      "learning_rate": 3.5721239031346066e-05,
      "loss": 0.4182,
      "mean_token_accuracy": 0.910229355096817,
      "num_tokens": 3253004.0,
      "step": 286
    },
    {
      "epoch": 3.6369426751592355,
      "grad_norm": 0.7678000926971436,
      "learning_rate": 3.508679979204481e-05,
      "loss": 0.4304,
      "mean_token_accuracy": 0.9080959856510162,
      "num_tokens": 3264403.0,
      "step": 287
    },
    {
      "epoch": 3.6496815286624202,
      "grad_norm": 0.7375931739807129,
      "learning_rate": 3.445684435721897e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.9104042947292328,
      "num_tokens": 3276391.0,
      "step": 288
    },
    {
      "epoch": 3.662420382165605,
      "grad_norm": 0.7385655641555786,
      "learning_rate": 3.383141624031408e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.9077424705028534,
      "num_tokens": 3287738.0,
      "step": 289
    },
    {
      "epoch": 3.6751592356687897,
      "grad_norm": 0.7327329516410828,
      "learning_rate": 3.3210558642056275e-05,
      "loss": 0.3339,
      "mean_token_accuracy": 0.9267513453960419,
      "num_tokens": 3299143.0,
      "step": 290
    },
    {
      "epoch": 3.6878980891719744,
      "grad_norm": 0.6492987275123596,
      "learning_rate": 3.259431444746846e-05,
      "loss": 0.3775,
      "mean_token_accuracy": 0.9218702614307404,
      "num_tokens": 3310487.0,
      "step": 291
    },
    {
      "epoch": 3.700636942675159,
      "grad_norm": 0.759364902973175,
      "learning_rate": 3.198272622290804e-05,
      "loss": 0.3809,
      "mean_token_accuracy": 0.9182988703250885,
      "num_tokens": 3321796.0,
      "step": 292
    },
    {
      "epoch": 3.713375796178344,
      "grad_norm": 0.7147376537322998,
      "learning_rate": 3.137583621312665e-05,
      "loss": 0.4111,
      "mean_token_accuracy": 0.9116301238536835,
      "num_tokens": 3332604.0,
      "step": 293
    },
    {
      "epoch": 3.7261146496815285,
      "grad_norm": 0.7770649194717407,
      "learning_rate": 3.077368633835205e-05,
      "loss": 0.3964,
      "mean_token_accuracy": 0.9151590764522552,
      "num_tokens": 3344007.0,
      "step": 294
    },
    {
      "epoch": 3.738853503184713,
      "grad_norm": 0.7108721733093262,
      "learning_rate": 3.0176318191392726e-05,
      "loss": 0.4136,
      "mean_token_accuracy": 0.9083453416824341,
      "num_tokens": 3355090.0,
      "step": 295
    },
    {
      "epoch": 3.7515923566878984,
      "grad_norm": 0.7094624638557434,
      "learning_rate": 2.9583773034764826e-05,
      "loss": 0.3454,
      "mean_token_accuracy": 0.9238831102848053,
      "num_tokens": 3365904.0,
      "step": 296
    },
    {
      "epoch": 3.7643312101910826,
      "grad_norm": 0.6758113503456116,
      "learning_rate": 2.8996091797841973e-05,
      "loss": 0.3816,
      "mean_token_accuracy": 0.9179931282997131,
      "num_tokens": 3377964.0,
      "step": 297
    },
    {
      "epoch": 3.777070063694268,
      "grad_norm": 0.7704566121101379,
      "learning_rate": 2.8413315074028158e-05,
      "loss": 0.431,
      "mean_token_accuracy": 0.9047513604164124,
      "num_tokens": 3389347.0,
      "step": 298
    },
    {
      "epoch": 3.789808917197452,
      "grad_norm": 0.6735308766365051,
      "learning_rate": 2.7835483117953788e-05,
      "loss": 0.3715,
      "mean_token_accuracy": 0.9192857146263123,
      "num_tokens": 3400439.0,
      "step": 299
    },
    {
      "epoch": 3.802547770700637,
      "grad_norm": 0.8025267124176025,
      "learning_rate": 2.7262635842695127e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.9085029065608978,
      "num_tokens": 3411639.0,
      "step": 300
    },
    {
      "epoch": 3.8152866242038215,
      "grad_norm": 0.6782514452934265,
      "learning_rate": 2.669481281701739e-05,
      "loss": 0.353,
      "mean_token_accuracy": 0.9254265427589417,
      "num_tokens": 3423172.0,
      "step": 301
    },
    {
      "epoch": 3.8280254777070066,
      "grad_norm": 0.7359219789505005,
      "learning_rate": 2.6132053262641466e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.9082939028739929,
      "num_tokens": 3434947.0,
      "step": 302
    },
    {
      "epoch": 3.840764331210191,
      "grad_norm": 0.7089890241622925,
      "learning_rate": 2.5574396051534832e-05,
      "loss": 0.3922,
      "mean_token_accuracy": 0.9144908785820007,
      "num_tokens": 3446512.0,
      "step": 303
    },
    {
      "epoch": 3.853503184713376,
      "grad_norm": 0.6748204827308655,
      "learning_rate": 2.502187970322657e-05,
      "loss": 0.3985,
      "mean_token_accuracy": 0.9171700477600098,
      "num_tokens": 3458505.0,
      "step": 304
    },
    {
      "epoch": 3.8662420382165603,
      "grad_norm": 0.7042027711868286,
      "learning_rate": 2.4474542382146537e-05,
      "loss": 0.3913,
      "mean_token_accuracy": 0.9139814674854279,
      "num_tokens": 3470163.0,
      "step": 305
    },
    {
      "epoch": 3.8789808917197455,
      "grad_norm": 0.6868739128112793,
      "learning_rate": 2.3932421894989167e-05,
      "loss": 0.3378,
      "mean_token_accuracy": 0.9283902049064636,
      "num_tokens": 3481090.0,
      "step": 306
    },
    {
      "epoch": 3.8917197452229297,
      "grad_norm": 0.714598536491394,
      "learning_rate": 2.339555568810221e-05,
      "loss": 0.4005,
      "mean_token_accuracy": 0.9129302501678467,
      "num_tokens": 3491946.0,
      "step": 307
    },
    {
      "epoch": 3.904458598726115,
      "grad_norm": 0.8096787929534912,
      "learning_rate": 2.2863980844900036e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.9040016531944275,
      "num_tokens": 3502961.0,
      "step": 308
    },
    {
      "epoch": 3.917197452229299,
      "grad_norm": 0.7268587946891785,
      "learning_rate": 2.2337734083302164e-05,
      "loss": 0.3623,
      "mean_token_accuracy": 0.9210439920425415,
      "num_tokens": 3514116.0,
      "step": 309
    },
    {
      "epoch": 3.9299363057324843,
      "grad_norm": 0.7080064415931702,
      "learning_rate": 2.181685175319702e-05,
      "loss": 0.3944,
      "mean_token_accuracy": 0.9152760803699493,
      "num_tokens": 3525855.0,
      "step": 310
    },
    {
      "epoch": 3.9426751592356686,
      "grad_norm": 0.6933088898658752,
      "learning_rate": 2.1301369833931117e-05,
      "loss": 0.3955,
      "mean_token_accuracy": 0.9157300591468811,
      "num_tokens": 3536972.0,
      "step": 311
    },
    {
      "epoch": 3.9554140127388537,
      "grad_norm": 0.7148348093032837,
      "learning_rate": 2.079132393182378e-05,
      "loss": 0.3739,
      "mean_token_accuracy": 0.917779803276062,
      "num_tokens": 3548281.0,
      "step": 312
    },
    {
      "epoch": 3.968152866242038,
      "grad_norm": 0.7383513450622559,
      "learning_rate": 2.0286749277707782e-05,
      "loss": 0.3799,
      "mean_token_accuracy": 0.9188788533210754,
      "num_tokens": 3559528.0,
      "step": 313
    },
    {
      "epoch": 3.980891719745223,
      "grad_norm": 0.7167903184890747,
      "learning_rate": 1.9787680724495617e-05,
      "loss": 0.4203,
      "mean_token_accuracy": 0.9090096354484558,
      "num_tokens": 3571410.0,
      "step": 314
    },
    {
      "epoch": 3.9936305732484074,
      "grad_norm": 0.7360700964927673,
      "learning_rate": 1.929415274477239e-05,
      "loss": 0.3843,
      "mean_token_accuracy": 0.917841374874115,
      "num_tokens": 3582989.0,
      "step": 315
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9755165576934814,
      "learning_rate": 1.880619942841435e-05,
      "loss": 0.2811,
      "mean_token_accuracy": 0.9407498836517334,
      "num_tokens": 3588675.0,
      "step": 316
    },
    {
      "epoch": 4.012738853503185,
      "grad_norm": 0.6373947858810425,
      "learning_rate": 1.832385448023435e-05,
      "loss": 0.305,
      "mean_token_accuracy": 0.937866747379303,
      "num_tokens": 3600582.0,
      "step": 317
    },
    {
      "epoch": 4.025477707006369,
      "grad_norm": 0.513500452041626,
      "learning_rate": 1.7847151217653624e-05,
      "loss": 0.1996,
      "mean_token_accuracy": 0.9598060250282288,
      "num_tokens": 3612395.0,
      "step": 318
    },
    {
      "epoch": 4.038216560509555,
      "grad_norm": 0.5696059465408325,
      "learning_rate": 1.7376122568400532e-05,
      "loss": 0.2323,
      "mean_token_accuracy": 0.9527164399623871,
      "num_tokens": 3623934.0,
      "step": 319
    },
    {
      "epoch": 4.050955414012739,
      "grad_norm": 0.6302847266197205,
      "learning_rate": 1.6910801068236016e-05,
      "loss": 0.2627,
      "mean_token_accuracy": 0.9468440115451813,
      "num_tokens": 3635141.0,
      "step": 320
    },
    {
      "epoch": 4.063694267515924,
      "grad_norm": 0.5786097049713135,
      "learning_rate": 1.6451218858706374e-05,
      "loss": 0.2515,
      "mean_token_accuracy": 0.9479193389415741,
      "num_tokens": 3646412.0,
      "step": 321
    },
    {
      "epoch": 4.076433121019108,
      "grad_norm": 0.5529419779777527,
      "learning_rate": 1.5997407684922862e-05,
      "loss": 0.2329,
      "mean_token_accuracy": 0.9493387043476105,
      "num_tokens": 3658369.0,
      "step": 322
    },
    {
      "epoch": 4.089171974522293,
      "grad_norm": 0.5960792303085327,
      "learning_rate": 1.5549398893369216e-05,
      "loss": 0.2179,
      "mean_token_accuracy": 0.9559203088283539,
      "num_tokens": 3669448.0,
      "step": 323
    },
    {
      "epoch": 4.101910828025478,
      "grad_norm": 0.6120526790618896,
      "learning_rate": 1.5107223429736272e-05,
      "loss": 0.2356,
      "mean_token_accuracy": 0.950799435377121,
      "num_tokens": 3681023.0,
      "step": 324
    },
    {
      "epoch": 4.114649681528663,
      "grad_norm": 0.6094869375228882,
      "learning_rate": 1.467091183678444e-05,
      "loss": 0.221,
      "mean_token_accuracy": 0.9547978937625885,
      "num_tokens": 3692657.0,
      "step": 325
    },
    {
      "epoch": 4.127388535031847,
      "grad_norm": 0.7375300526618958,
      "learning_rate": 1.4240494252234049e-05,
      "loss": 0.2434,
      "mean_token_accuracy": 0.9470761716365814,
      "num_tokens": 3704395.0,
      "step": 326
    },
    {
      "epoch": 4.140127388535032,
      "grad_norm": 0.7640740871429443,
      "learning_rate": 1.3816000406683604e-05,
      "loss": 0.2591,
      "mean_token_accuracy": 0.9460343718528748,
      "num_tokens": 3716489.0,
      "step": 327
    },
    {
      "epoch": 4.1528662420382165,
      "grad_norm": 0.750254213809967,
      "learning_rate": 1.339745962155613e-05,
      "loss": 0.2566,
      "mean_token_accuracy": 0.9457094967365265,
      "num_tokens": 3727830.0,
      "step": 328
    },
    {
      "epoch": 4.165605095541402,
      "grad_norm": 0.6968578100204468,
      "learning_rate": 1.2984900807073919e-05,
      "loss": 0.2296,
      "mean_token_accuracy": 0.9500870108604431,
      "num_tokens": 3739056.0,
      "step": 329
    },
    {
      "epoch": 4.178343949044586,
      "grad_norm": 0.7135255932807922,
      "learning_rate": 1.2578352460261456e-05,
      "loss": 0.2714,
      "mean_token_accuracy": 0.943615198135376,
      "num_tokens": 3750358.0,
      "step": 330
    },
    {
      "epoch": 4.191082802547771,
      "grad_norm": 0.721027135848999,
      "learning_rate": 1.2177842662977135e-05,
      "loss": 0.2812,
      "mean_token_accuracy": 0.9410761296749115,
      "num_tokens": 3761678.0,
      "step": 331
    },
    {
      "epoch": 4.203821656050955,
      "grad_norm": 0.7138113975524902,
      "learning_rate": 1.1783399079973578e-05,
      "loss": 0.2824,
      "mean_token_accuracy": 0.9425844848155975,
      "num_tokens": 3772919.0,
      "step": 332
    },
    {
      "epoch": 4.2165605095541405,
      "grad_norm": 0.6014682054519653,
      "learning_rate": 1.1395048956986575e-05,
      "loss": 0.2404,
      "mean_token_accuracy": 0.9489006102085114,
      "num_tokens": 3784964.0,
      "step": 333
    },
    {
      "epoch": 4.229299363057325,
      "grad_norm": 0.6314707398414612,
      "learning_rate": 1.1012819118853147e-05,
      "loss": 0.2496,
      "mean_token_accuracy": 0.9463862776756287,
      "num_tokens": 3796759.0,
      "step": 334
    },
    {
      "epoch": 4.24203821656051,
      "grad_norm": 0.5552114248275757,
      "learning_rate": 1.0636735967658784e-05,
      "loss": 0.2108,
      "mean_token_accuracy": 0.9555350244045258,
      "num_tokens": 3807780.0,
      "step": 335
    },
    {
      "epoch": 4.254777070063694,
      "grad_norm": 0.6171155571937561,
      "learning_rate": 1.0266825480913611e-05,
      "loss": 0.2588,
      "mean_token_accuracy": 0.9461368918418884,
      "num_tokens": 3818794.0,
      "step": 336
    },
    {
      "epoch": 4.267515923566879,
      "grad_norm": 0.5676184892654419,
      "learning_rate": 9.903113209758096e-06,
      "loss": 0.2269,
      "mean_token_accuracy": 0.952668160200119,
      "num_tokens": 3830160.0,
      "step": 337
    },
    {
      "epoch": 4.280254777070064,
      "grad_norm": 0.5797393321990967,
      "learning_rate": 9.545624277198084e-06,
      "loss": 0.2436,
      "mean_token_accuracy": 0.9488925039768219,
      "num_tokens": 3841632.0,
      "step": 338
    },
    {
      "epoch": 4.292993630573249,
      "grad_norm": 0.6091304421424866,
      "learning_rate": 9.194383376369508e-06,
      "loss": 0.2399,
      "mean_token_accuracy": 0.9505036473274231,
      "num_tokens": 3852969.0,
      "step": 339
    },
    {
      "epoch": 4.305732484076433,
      "grad_norm": 0.5556879639625549,
      "learning_rate": 8.849414768832687e-06,
      "loss": 0.2102,
      "mean_token_accuracy": 0.9555756449699402,
      "num_tokens": 3864175.0,
      "step": 340
    },
    {
      "epoch": 4.318471337579618,
      "grad_norm": 0.6041027307510376,
      "learning_rate": 8.510742282896544e-06,
      "loss": 0.2442,
      "mean_token_accuracy": 0.9477723836898804,
      "num_tokens": 3875709.0,
      "step": 341
    },
    {
      "epoch": 4.3312101910828025,
      "grad_norm": 0.541427731513977,
      "learning_rate": 8.178389311972612e-06,
      "loss": 0.2127,
      "mean_token_accuracy": 0.955585241317749,
      "num_tokens": 3887256.0,
      "step": 342
    },
    {
      "epoch": 4.343949044585988,
      "grad_norm": 0.6119688749313354,
      "learning_rate": 7.852378812959227e-06,
      "loss": 0.2542,
      "mean_token_accuracy": 0.945166677236557,
      "num_tokens": 3898876.0,
      "step": 343
    },
    {
      "epoch": 4.356687898089172,
      "grad_norm": 0.5536457896232605,
      "learning_rate": 7.532733304655848e-06,
      "loss": 0.2285,
      "mean_token_accuracy": 0.9501866400241852,
      "num_tokens": 3910522.0,
      "step": 344
    },
    {
      "epoch": 4.369426751592357,
      "grad_norm": 0.6129220128059387,
      "learning_rate": 7.219474866207465e-06,
      "loss": 0.232,
      "mean_token_accuracy": 0.950534850358963,
      "num_tokens": 3921730.0,
      "step": 345
    },
    {
      "epoch": 4.382165605095541,
      "grad_norm": 0.5599251985549927,
      "learning_rate": 6.9126251355795864e-06,
      "loss": 0.2213,
      "mean_token_accuracy": 0.9533770978450775,
      "num_tokens": 3933613.0,
      "step": 346
    },
    {
      "epoch": 4.3949044585987265,
      "grad_norm": 0.6107044816017151,
      "learning_rate": 6.612205308063646e-06,
      "loss": 0.2474,
      "mean_token_accuracy": 0.9481737017631531,
      "num_tokens": 3945095.0,
      "step": 347
    },
    {
      "epoch": 4.407643312101911,
      "grad_norm": 0.6116899847984314,
      "learning_rate": 6.318236134812916e-06,
      "loss": 0.2405,
      "mean_token_accuracy": 0.9522660076618195,
      "num_tokens": 3956565.0,
      "step": 348
    },
    {
      "epoch": 4.420382165605096,
      "grad_norm": 0.649278998374939,
      "learning_rate": 6.030737921409169e-06,
      "loss": 0.2637,
      "mean_token_accuracy": 0.9446616172790527,
      "num_tokens": 3967950.0,
      "step": 349
    },
    {
      "epoch": 4.43312101910828,
      "grad_norm": 0.6206309199333191,
      "learning_rate": 5.749730526460073e-06,
      "loss": 0.2551,
      "mean_token_accuracy": 0.9476028084754944,
      "num_tokens": 3979520.0,
      "step": 350
    },
    {
      "epoch": 4.445859872611465,
      "grad_norm": 0.6442816853523254,
      "learning_rate": 5.475233360227516e-06,
      "loss": 0.2491,
      "mean_token_accuracy": 0.9484961330890656,
      "num_tokens": 3990605.0,
      "step": 351
    },
    {
      "epoch": 4.45859872611465,
      "grad_norm": 0.615239679813385,
      "learning_rate": 5.20726538328683e-06,
      "loss": 0.248,
      "mean_token_accuracy": 0.9499914944171906,
      "num_tokens": 4001897.0,
      "step": 352
    },
    {
      "epoch": 4.471337579617835,
      "grad_norm": 0.5654546618461609,
      "learning_rate": 4.945845105217117e-06,
      "loss": 0.208,
      "mean_token_accuracy": 0.9568901658058167,
      "num_tokens": 4013030.0,
      "step": 353
    },
    {
      "epoch": 4.484076433121019,
      "grad_norm": 0.593754768371582,
      "learning_rate": 4.6909905833226966e-06,
      "loss": 0.2368,
      "mean_token_accuracy": 0.9518877267837524,
      "num_tokens": 4024717.0,
      "step": 354
    },
    {
      "epoch": 4.496815286624204,
      "grad_norm": 0.6317068934440613,
      "learning_rate": 4.442719421385922e-06,
      "loss": 0.2602,
      "mean_token_accuracy": 0.9456787109375,
      "num_tokens": 4036521.0,
      "step": 355
    },
    {
      "epoch": 4.509554140127388,
      "grad_norm": 0.5658568739891052,
      "learning_rate": 4.20104876845111e-06,
      "loss": 0.2149,
      "mean_token_accuracy": 0.9562082588672638,
      "num_tokens": 4048246.0,
      "step": 356
    },
    {
      "epoch": 4.522292993630574,
      "grad_norm": 0.6423606872558594,
      "learning_rate": 3.965995317640025e-06,
      "loss": 0.2246,
      "mean_token_accuracy": 0.9531770050525665,
      "num_tokens": 4059468.0,
      "step": 357
    },
    {
      "epoch": 4.535031847133758,
      "grad_norm": 0.6500980257987976,
      "learning_rate": 3.7375753049987973e-06,
      "loss": 0.2625,
      "mean_token_accuracy": 0.9455468654632568,
      "num_tokens": 4071007.0,
      "step": 358
    },
    {
      "epoch": 4.547770700636943,
      "grad_norm": 0.6197830438613892,
      "learning_rate": 3.515804508376508e-06,
      "loss": 0.2299,
      "mean_token_accuracy": 0.9506585597991943,
      "num_tokens": 4081959.0,
      "step": 359
    },
    {
      "epoch": 4.560509554140127,
      "grad_norm": 0.6473744511604309,
      "learning_rate": 3.3006982463352766e-06,
      "loss": 0.2615,
      "mean_token_accuracy": 0.9437356889247894,
      "num_tokens": 4093354.0,
      "step": 360
    },
    {
      "epoch": 4.573248407643312,
      "grad_norm": 0.6327031254768372,
      "learning_rate": 3.092271377092215e-06,
      "loss": 0.267,
      "mean_token_accuracy": 0.9425212442874908,
      "num_tokens": 4105174.0,
      "step": 361
    },
    {
      "epoch": 4.585987261146497,
      "grad_norm": 0.5570584535598755,
      "learning_rate": 2.8905382974930172e-06,
      "loss": 0.2064,
      "mean_token_accuracy": 0.954309344291687,
      "num_tokens": 4116576.0,
      "step": 362
    },
    {
      "epoch": 4.598726114649682,
      "grad_norm": 0.623688280582428,
      "learning_rate": 2.6955129420176196e-06,
      "loss": 0.2206,
      "mean_token_accuracy": 0.9519834220409393,
      "num_tokens": 4127873.0,
      "step": 363
    },
    {
      "epoch": 4.611464968152866,
      "grad_norm": 0.6340871453285217,
      "learning_rate": 2.5072087818176382e-06,
      "loss": 0.2358,
      "mean_token_accuracy": 0.9510488510131836,
      "num_tokens": 4138888.0,
      "step": 364
    },
    {
      "epoch": 4.624203821656051,
      "grad_norm": 0.657813549041748,
      "learning_rate": 2.3256388237858807e-06,
      "loss": 0.2499,
      "mean_token_accuracy": 0.9460207223892212,
      "num_tokens": 4150232.0,
      "step": 365
    },
    {
      "epoch": 4.6369426751592355,
      "grad_norm": 0.7132663726806641,
      "learning_rate": 2.150815609657875e-06,
      "loss": 0.2893,
      "mean_token_accuracy": 0.9409700334072113,
      "num_tokens": 4161733.0,
      "step": 366
    },
    {
      "epoch": 4.649681528662421,
      "grad_norm": 0.6530131697654724,
      "learning_rate": 1.9827512151456173e-06,
      "loss": 0.2769,
      "mean_token_accuracy": 0.9446337521076202,
      "num_tokens": 4173910.0,
      "step": 367
    },
    {
      "epoch": 4.662420382165605,
      "grad_norm": 0.5950355529785156,
      "learning_rate": 1.8214572491034198e-06,
      "loss": 0.2145,
      "mean_token_accuracy": 0.9520914554595947,
      "num_tokens": 4185377.0,
      "step": 368
    },
    {
      "epoch": 4.67515923566879,
      "grad_norm": 0.584748387336731,
      "learning_rate": 1.66694485272606e-06,
      "loss": 0.2142,
      "mean_token_accuracy": 0.9542392492294312,
      "num_tokens": 4196586.0,
      "step": 369
    },
    {
      "epoch": 4.687898089171974,
      "grad_norm": 0.6257981657981873,
      "learning_rate": 1.5192246987791981e-06,
      "loss": 0.2138,
      "mean_token_accuracy": 0.9550438523292542,
      "num_tokens": 4207564.0,
      "step": 370
    },
    {
      "epoch": 4.7006369426751595,
      "grad_norm": 0.663212776184082,
      "learning_rate": 1.378306990862177e-06,
      "loss": 0.2523,
      "mean_token_accuracy": 0.9431774616241455,
      "num_tokens": 4218790.0,
      "step": 371
    },
    {
      "epoch": 4.713375796178344,
      "grad_norm": 0.6407923698425293,
      "learning_rate": 1.2442014627032316e-06,
      "loss": 0.2385,
      "mean_token_accuracy": 0.9502703845500946,
      "num_tokens": 4230117.0,
      "step": 372
    },
    {
      "epoch": 4.726114649681529,
      "grad_norm": 0.5782377123832703,
      "learning_rate": 1.1169173774871478e-06,
      "loss": 0.2311,
      "mean_token_accuracy": 0.9525089859962463,
      "num_tokens": 4241547.0,
      "step": 373
    },
    {
      "epoch": 4.738853503184713,
      "grad_norm": 0.5933149456977844,
      "learning_rate": 9.964635272153633e-07,
      "loss": 0.2354,
      "mean_token_accuracy": 0.950641393661499,
      "num_tokens": 4253164.0,
      "step": 374
    },
    {
      "epoch": 4.751592356687898,
      "grad_norm": 0.667209804058075,
      "learning_rate": 8.828482320987319e-07,
      "loss": 0.2699,
      "mean_token_accuracy": 0.9431827366352081,
      "num_tokens": 4264634.0,
      "step": 375
    },
    {
      "epoch": 4.764331210191083,
      "grad_norm": 0.6330809593200684,
      "learning_rate": 7.760793399827937e-07,
      "loss": 0.2219,
      "mean_token_accuracy": 0.950275719165802,
      "num_tokens": 4275920.0,
      "step": 376
    },
    {
      "epoch": 4.777070063694268,
      "grad_norm": 0.5830065011978149,
      "learning_rate": 6.761642258056978e-07,
      "loss": 0.2112,
      "mean_token_accuracy": 0.9556083977222443,
      "num_tokens": 4287045.0,
      "step": 377
    },
    {
      "epoch": 4.789808917197452,
      "grad_norm": 0.5773307681083679,
      "learning_rate": 5.831097910887873e-07,
      "loss": 0.2303,
      "mean_token_accuracy": 0.951410561800003,
      "num_tokens": 4298056.0,
      "step": 378
    },
    {
      "epoch": 4.802547770700637,
      "grad_norm": 0.6692339181900024,
      "learning_rate": 4.969224634598591e-07,
      "loss": 0.2468,
      "mean_token_accuracy": 0.9466256201267242,
      "num_tokens": 4309058.0,
      "step": 379
    },
    {
      "epoch": 4.8152866242038215,
      "grad_norm": 0.592280924320221,
      "learning_rate": 4.176081962092182e-07,
      "loss": 0.2243,
      "mean_token_accuracy": 0.9543558657169342,
      "num_tokens": 4320492.0,
      "step": 380
    },
    {
      "epoch": 4.828025477707007,
      "grad_norm": 0.5458922386169434,
      "learning_rate": 3.451724678784518e-07,
      "loss": 0.1969,
      "mean_token_accuracy": 0.9583150148391724,
      "num_tokens": 4331497.0,
      "step": 381
    },
    {
      "epoch": 4.840764331210191,
      "grad_norm": 0.6117383241653442,
      "learning_rate": 2.7962028188198706e-07,
      "loss": 0.2284,
      "mean_token_accuracy": 0.9501660764217377,
      "num_tokens": 4343022.0,
      "step": 382
    },
    {
      "epoch": 4.853503184713376,
      "grad_norm": 0.6400023698806763,
      "learning_rate": 2.2095616616150115e-07,
      "loss": 0.2443,
      "mean_token_accuracy": 0.9481694400310516,
      "num_tokens": 4354393.0,
      "step": 383
    },
    {
      "epoch": 4.86624203821656,
      "grad_norm": 0.5784524083137512,
      "learning_rate": 1.6918417287318245e-07,
      "loss": 0.2129,
      "mean_token_accuracy": 0.9560071527957916,
      "num_tokens": 4365580.0,
      "step": 384
    },
    {
      "epoch": 4.8789808917197455,
      "grad_norm": 0.5761446952819824,
      "learning_rate": 1.2430787810776555e-07,
      "loss": 0.2261,
      "mean_token_accuracy": 0.9526739120483398,
      "num_tokens": 4376649.0,
      "step": 385
    },
    {
      "epoch": 4.89171974522293,
      "grad_norm": 0.6780246496200562,
      "learning_rate": 8.633038164358454e-08,
      "loss": 0.2584,
      "mean_token_accuracy": 0.9444167017936707,
      "num_tokens": 4388351.0,
      "step": 386
    },
    {
      "epoch": 4.904458598726115,
      "grad_norm": 0.5891669392585754,
      "learning_rate": 5.5254306732444025e-08,
      "loss": 0.2168,
      "mean_token_accuracy": 0.9538559019565582,
      "num_tokens": 4400015.0,
      "step": 387
    },
    {
      "epoch": 4.917197452229299,
      "grad_norm": 0.6162629127502441,
      "learning_rate": 3.1081799918375454e-08,
      "loss": 0.2271,
      "mean_token_accuracy": 0.9495961964130402,
      "num_tokens": 4411248.0,
      "step": 388
    },
    {
      "epoch": 4.929936305732484,
      "grad_norm": 0.5894432663917542,
      "learning_rate": 1.3814530889433296e-08,
      "loss": 0.2199,
      "mean_token_accuracy": 0.954120010137558,
      "num_tokens": 4422081.0,
      "step": 389
    },
    {
      "epoch": 4.942675159235669,
      "grad_norm": 0.6032901406288147,
      "learning_rate": 3.453692362309635e-09,
      "loss": 0.2383,
      "mean_token_accuracy": 0.9507102966308594,
      "num_tokens": 4433867.0,
      "step": 390
    },
    {
      "epoch": 4.942675159235669,
      "step": 390,
      "total_flos": 2.6684669817153126e+17,
      "train_loss": 0.9139279948595243,
      "train_runtime": 757.4985,
      "train_samples_per_second": 33.003,
      "train_steps_per_second": 0.515
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 390,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.6684669817153126e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}