{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 5898,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.000508646998982706,
      "grad_norm": 32.115596771240234,
      "learning_rate": 0.0,
      "loss": 1.2961,
      "mean_token_accuracy": 0.7449109554290771,
      "num_tokens": 156200.0,
      "step": 1
    },
    {
      "epoch": 0.001017293997965412,
      "grad_norm": 31.279529571533203,
      "learning_rate": 1.6949152542372882e-08,
      "loss": 1.2669,
      "mean_token_accuracy": 0.7504401206970215,
      "num_tokens": 314183.0,
      "step": 2
    },
    {
      "epoch": 0.001525940996948118,
      "grad_norm": 30.887014389038086,
      "learning_rate": 3.3898305084745764e-08,
      "loss": 1.307,
      "mean_token_accuracy": 0.7392708659172058,
      "num_tokens": 470040.0,
      "step": 3
    },
    {
      "epoch": 0.002034587995930824,
      "grad_norm": 31.73535919189453,
      "learning_rate": 5.0847457627118645e-08,
      "loss": 1.3081,
      "mean_token_accuracy": 0.7396830916404724,
      "num_tokens": 627227.0,
      "step": 4
    },
    {
      "epoch": 0.00254323499491353,
      "grad_norm": 31.332460403442383,
      "learning_rate": 6.779661016949153e-08,
      "loss": 1.3231,
      "mean_token_accuracy": 0.7378069758415222,
      "num_tokens": 785176.0,
      "step": 5
    },
    {
      "epoch": 0.003051881993896236,
      "grad_norm": 32.63821029663086,
      "learning_rate": 8.474576271186442e-08,
      "loss": 1.3232,
      "mean_token_accuracy": 0.7402999401092529,
      "num_tokens": 926658.0,
      "step": 6
    },
    {
      "epoch": 0.003560528992878942,
      "grad_norm": 31.21712303161621,
      "learning_rate": 1.0169491525423729e-07,
      "loss": 1.334,
      "mean_token_accuracy": 0.7354116439819336,
      "num_tokens": 1089803.0,
      "step": 7
    },
    {
      "epoch": 0.004069175991861648,
      "grad_norm": 30.991884231567383,
      "learning_rate": 1.1864406779661018e-07,
      "loss": 1.2785,
      "mean_token_accuracy": 0.7471437454223633,
      "num_tokens": 1253116.0,
      "step": 8
    },
    {
      "epoch": 0.004577822990844354,
      "grad_norm": 29.952585220336914,
      "learning_rate": 1.3559322033898305e-07,
      "loss": 1.2702,
      "mean_token_accuracy": 0.7474691867828369,
      "num_tokens": 1419240.0,
      "step": 9
    },
    {
      "epoch": 0.00508646998982706,
      "grad_norm": 30.35393714904785,
      "learning_rate": 1.5254237288135596e-07,
      "loss": 1.2841,
      "mean_token_accuracy": 0.7456669211387634,
      "num_tokens": 1582136.0,
      "step": 10
    },
    {
      "epoch": 0.005595116988809766,
      "grad_norm": 31.37613868713379,
      "learning_rate": 1.6949152542372883e-07,
      "loss": 1.2995,
      "mean_token_accuracy": 0.7446110844612122,
      "num_tokens": 1738808.0,
      "step": 11
    },
    {
      "epoch": 0.006103763987792472,
      "grad_norm": 30.47648048400879,
      "learning_rate": 1.8644067796610173e-07,
      "loss": 1.2831,
      "mean_token_accuracy": 0.7443124055862427,
      "num_tokens": 1899325.0,
      "step": 12
    },
    {
      "epoch": 0.006612410986775178,
      "grad_norm": 30.103708267211914,
      "learning_rate": 2.0338983050847458e-07,
      "loss": 1.2404,
      "mean_token_accuracy": 0.7517591714859009,
      "num_tokens": 2067149.0,
      "step": 13
    },
    {
      "epoch": 0.007121057985757884,
      "grad_norm": 29.119670867919922,
      "learning_rate": 2.2033898305084748e-07,
      "loss": 1.3089,
      "mean_token_accuracy": 0.7355021834373474,
      "num_tokens": 2225992.0,
      "step": 14
    },
    {
      "epoch": 0.0076297049847405905,
      "grad_norm": 30.91548728942871,
      "learning_rate": 2.3728813559322036e-07,
      "loss": 1.288,
      "mean_token_accuracy": 0.7432087063789368,
      "num_tokens": 2391475.0,
      "step": 15
    },
    {
      "epoch": 0.008138351983723296,
      "grad_norm": 27.432802200317383,
      "learning_rate": 2.5423728813559323e-07,
      "loss": 1.2468,
      "mean_token_accuracy": 0.7456105947494507,
      "num_tokens": 2564876.0,
      "step": 16
    },
    {
      "epoch": 0.008646998982706003,
      "grad_norm": 26.89838218688965,
      "learning_rate": 2.711864406779661e-07,
      "loss": 1.2802,
      "mean_token_accuracy": 0.7338192462921143,
      "num_tokens": 2728912.0,
      "step": 17
    },
    {
      "epoch": 0.009155645981688708,
      "grad_norm": 28.779483795166016,
      "learning_rate": 2.88135593220339e-07,
      "loss": 1.2371,
      "mean_token_accuracy": 0.7503151297569275,
      "num_tokens": 2885094.0,
      "step": 18
    },
    {
      "epoch": 0.009664292980671414,
      "grad_norm": 28.094846725463867,
      "learning_rate": 3.050847457627119e-07,
      "loss": 1.2683,
      "mean_token_accuracy": 0.7387984991073608,
      "num_tokens": 3043505.0,
      "step": 19
    },
    {
      "epoch": 0.01017293997965412,
      "grad_norm": 28.859289169311523,
      "learning_rate": 3.2203389830508473e-07,
      "loss": 1.1843,
      "mean_token_accuracy": 0.7617287635803223,
      "num_tokens": 3200547.0,
      "step": 20
    },
    {
      "epoch": 0.010681586978636826,
      "grad_norm": 27.48524284362793,
      "learning_rate": 3.3898305084745766e-07,
      "loss": 1.2653,
      "mean_token_accuracy": 0.7411770224571228,
      "num_tokens": 3352567.0,
      "step": 21
    },
    {
      "epoch": 0.011190233977619531,
      "grad_norm": 26.313617706298828,
      "learning_rate": 3.5593220338983054e-07,
      "loss": 1.2687,
      "mean_token_accuracy": 0.7347060441970825,
      "num_tokens": 3522299.0,
      "step": 22
    },
    {
      "epoch": 0.011698880976602238,
      "grad_norm": 21.680171966552734,
      "learning_rate": 3.7288135593220347e-07,
      "loss": 1.2141,
      "mean_token_accuracy": 0.7363147735595703,
      "num_tokens": 3678029.0,
      "step": 23
    },
    {
      "epoch": 0.012207527975584944,
      "grad_norm": 19.086170196533203,
      "learning_rate": 3.8983050847457634e-07,
      "loss": 1.1484,
      "mean_token_accuracy": 0.7474343180656433,
      "num_tokens": 3843052.0,
      "step": 24
    },
    {
      "epoch": 0.01271617497456765,
      "grad_norm": 19.647796630859375,
      "learning_rate": 4.0677966101694916e-07,
      "loss": 1.1527,
      "mean_token_accuracy": 0.7474626302719116,
      "num_tokens": 4003590.0,
      "step": 25
    },
    {
      "epoch": 0.013224821973550356,
      "grad_norm": 18.843202590942383,
      "learning_rate": 4.2372881355932204e-07,
      "loss": 1.1387,
      "mean_token_accuracy": 0.748569130897522,
      "num_tokens": 4163923.0,
      "step": 26
    },
    {
      "epoch": 0.013733468972533061,
      "grad_norm": 17.923465728759766,
      "learning_rate": 4.4067796610169497e-07,
      "loss": 1.1809,
      "mean_token_accuracy": 0.7369387149810791,
      "num_tokens": 4329915.0,
      "step": 27
    },
    {
      "epoch": 0.014242115971515769,
      "grad_norm": 16.861265182495117,
      "learning_rate": 4.5762711864406784e-07,
      "loss": 1.1181,
      "mean_token_accuracy": 0.7488915324211121,
      "num_tokens": 4501785.0,
      "step": 28
    },
    {
      "epoch": 0.014750762970498474,
      "grad_norm": 17.190711975097656,
      "learning_rate": 4.745762711864407e-07,
      "loss": 1.1287,
      "mean_token_accuracy": 0.7477043867111206,
      "num_tokens": 4665203.0,
      "step": 29
    },
    {
      "epoch": 0.015259409969481181,
      "grad_norm": 17.02585220336914,
      "learning_rate": 4.915254237288136e-07,
      "loss": 1.1092,
      "mean_token_accuracy": 0.7519813776016235,
      "num_tokens": 4826248.0,
      "step": 30
    },
    {
      "epoch": 0.015768056968463885,
      "grad_norm": 15.061558723449707,
      "learning_rate": 5.084745762711865e-07,
      "loss": 1.0856,
      "mean_token_accuracy": 0.750836193561554,
      "num_tokens": 5002184.0,
      "step": 31
    },
    {
      "epoch": 0.01627670396744659,
      "grad_norm": 8.035051345825195,
      "learning_rate": 5.254237288135593e-07,
      "loss": 1.0415,
      "mean_token_accuracy": 0.7453987002372742,
      "num_tokens": 5159429.0,
      "step": 32
    },
    {
      "epoch": 0.0167853509664293,
      "grad_norm": 5.959025859832764,
      "learning_rate": 5.423728813559322e-07,
      "loss": 1.0122,
      "mean_token_accuracy": 0.752120852470398,
      "num_tokens": 5319981.0,
      "step": 33
    },
    {
      "epoch": 0.017293997965412006,
      "grad_norm": 5.894038677215576,
      "learning_rate": 5.593220338983051e-07,
      "loss": 1.0374,
      "mean_token_accuracy": 0.7448702454566956,
      "num_tokens": 5470920.0,
      "step": 34
    },
    {
      "epoch": 0.01780264496439471,
      "grad_norm": 5.831063270568848,
      "learning_rate": 5.76271186440678e-07,
      "loss": 1.0496,
      "mean_token_accuracy": 0.7414418458938599,
      "num_tokens": 5621279.0,
      "step": 35
    },
    {
      "epoch": 0.018311291963377416,
      "grad_norm": 5.664057731628418,
      "learning_rate": 5.93220338983051e-07,
      "loss": 1.002,
      "mean_token_accuracy": 0.7509836554527283,
      "num_tokens": 5783722.0,
      "step": 36
    },
    {
      "epoch": 0.018819938962360123,
      "grad_norm": 5.4525675773620605,
      "learning_rate": 6.101694915254238e-07,
      "loss": 1.0055,
      "mean_token_accuracy": 0.7488620281219482,
      "num_tokens": 5946247.0,
      "step": 37
    },
    {
      "epoch": 0.019328585961342827,
      "grad_norm": 5.1193952560424805,
      "learning_rate": 6.271186440677966e-07,
      "loss": 1.0142,
      "mean_token_accuracy": 0.7472719550132751,
      "num_tokens": 6105483.0,
      "step": 38
    },
    {
      "epoch": 0.019837232960325534,
      "grad_norm": 4.926507949829102,
      "learning_rate": 6.440677966101695e-07,
      "loss": 1.0268,
      "mean_token_accuracy": 0.7443497776985168,
      "num_tokens": 6267257.0,
      "step": 39
    },
    {
      "epoch": 0.02034587995930824,
      "grad_norm": 4.7316999435424805,
      "learning_rate": 6.610169491525425e-07,
      "loss": 0.9777,
      "mean_token_accuracy": 0.7549794316291809,
      "num_tokens": 6423575.0,
      "step": 40
    },
    {
      "epoch": 0.020854526958290945,
      "grad_norm": 4.418249607086182,
      "learning_rate": 6.779661016949153e-07,
      "loss": 0.9854,
      "mean_token_accuracy": 0.75412917137146,
      "num_tokens": 6580647.0,
      "step": 41
    },
    {
      "epoch": 0.021363173957273652,
      "grad_norm": 3.7720751762390137,
      "learning_rate": 6.949152542372882e-07,
      "loss": 0.956,
      "mean_token_accuracy": 0.7564523816108704,
      "num_tokens": 6740933.0,
      "step": 42
    },
    {
      "epoch": 0.02187182095625636,
      "grad_norm": 4.199706554412842,
      "learning_rate": 7.118644067796611e-07,
      "loss": 0.9435,
      "mean_token_accuracy": 0.7573504447937012,
      "num_tokens": 6888534.0,
      "step": 43
    },
    {
      "epoch": 0.022380467955239063,
      "grad_norm": 5.001425266265869,
      "learning_rate": 7.28813559322034e-07,
      "loss": 0.934,
      "mean_token_accuracy": 0.7598178386688232,
      "num_tokens": 7053795.0,
      "step": 44
    },
    {
      "epoch": 0.02288911495422177,
      "grad_norm": 4.949459552764893,
      "learning_rate": 7.457627118644069e-07,
      "loss": 0.9264,
      "mean_token_accuracy": 0.7590354681015015,
      "num_tokens": 7206610.0,
      "step": 45
    },
    {
      "epoch": 0.023397761953204477,
      "grad_norm": 5.248065948486328,
      "learning_rate": 7.627118644067798e-07,
      "loss": 0.981,
      "mean_token_accuracy": 0.7476550340652466,
      "num_tokens": 7375236.0,
      "step": 46
    },
    {
      "epoch": 0.023906408952187184,
      "grad_norm": 5.145699501037598,
      "learning_rate": 7.796610169491527e-07,
      "loss": 0.9164,
      "mean_token_accuracy": 0.7608495950698853,
      "num_tokens": 7533110.0,
      "step": 47
    },
    {
      "epoch": 0.024415055951169887,
      "grad_norm": 5.29788064956665,
      "learning_rate": 7.966101694915255e-07,
      "loss": 0.943,
      "mean_token_accuracy": 0.7549550533294678,
      "num_tokens": 7693804.0,
      "step": 48
    },
    {
      "epoch": 0.024923702950152594,
      "grad_norm": 4.676819324493408,
      "learning_rate": 8.135593220338983e-07,
      "loss": 0.8807,
      "mean_token_accuracy": 0.7704964280128479,
      "num_tokens": 7873397.0,
      "step": 49
    },
    {
      "epoch": 0.0254323499491353,
      "grad_norm": 4.2016096115112305,
      "learning_rate": 8.305084745762712e-07,
      "loss": 0.9193,
      "mean_token_accuracy": 0.7602928876876831,
      "num_tokens": 8024175.0,
      "step": 50
    },
    {
      "epoch": 0.025940996948118005,
      "grad_norm": 3.9279732704162598,
      "learning_rate": 8.474576271186441e-07,
      "loss": 0.897,
      "mean_token_accuracy": 0.7654056549072266,
      "num_tokens": 8177438.0,
      "step": 51
    },
    {
      "epoch": 0.026449643947100712,
      "grad_norm": 3.5190112590789795,
      "learning_rate": 8.64406779661017e-07,
      "loss": 0.8926,
      "mean_token_accuracy": 0.7674021124839783,
      "num_tokens": 8351946.0,
      "step": 52
    },
    {
      "epoch": 0.02695829094608342,
      "grad_norm": 3.196254253387451,
      "learning_rate": 8.813559322033899e-07,
      "loss": 0.899,
      "mean_token_accuracy": 0.7623940706253052,
      "num_tokens": 8504280.0,
      "step": 53
    },
    {
      "epoch": 0.027466937945066123,
      "grad_norm": 2.7295444011688232,
      "learning_rate": 8.983050847457628e-07,
      "loss": 0.8606,
      "mean_token_accuracy": 0.7741144895553589,
      "num_tokens": 8661017.0,
      "step": 54
    },
    {
      "epoch": 0.02797558494404883,
      "grad_norm": 2.655987024307251,
      "learning_rate": 9.152542372881357e-07,
      "loss": 0.8937,
      "mean_token_accuracy": 0.7659112811088562,
      "num_tokens": 8819796.0,
      "step": 55
    },
    {
      "epoch": 0.028484231943031537,
      "grad_norm": 2.4142653942108154,
      "learning_rate": 9.322033898305086e-07,
      "loss": 0.8583,
      "mean_token_accuracy": 0.7737700343132019,
      "num_tokens": 8975390.0,
      "step": 56
    },
    {
      "epoch": 0.02899287894201424,
      "grad_norm": 2.334871530532837,
      "learning_rate": 9.491525423728814e-07,
      "loss": 0.8747,
      "mean_token_accuracy": 0.7709072828292847,
      "num_tokens": 9128994.0,
      "step": 57
    },
    {
      "epoch": 0.029501525940996948,
      "grad_norm": 2.374896287918091,
      "learning_rate": 9.661016949152544e-07,
      "loss": 0.8484,
      "mean_token_accuracy": 0.7743436098098755,
      "num_tokens": 9280736.0,
      "step": 58
    },
    {
      "epoch": 0.030010172939979655,
      "grad_norm": 2.246040105819702,
      "learning_rate": 9.830508474576272e-07,
      "loss": 0.8892,
      "mean_token_accuracy": 0.7636919021606445,
      "num_tokens": 9431826.0,
      "step": 59
    },
    {
      "epoch": 0.030518819938962362,
      "grad_norm": 2.068854808807373,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.8208,
      "mean_token_accuracy": 0.7815329432487488,
      "num_tokens": 9587259.0,
      "step": 60
    },
    {
      "epoch": 0.031027466937945065,
      "grad_norm": 1.8931329250335693,
      "learning_rate": 1.016949152542373e-06,
      "loss": 0.828,
      "mean_token_accuracy": 0.7761904001235962,
      "num_tokens": 9735856.0,
      "step": 61
    },
    {
      "epoch": 0.03153611393692777,
      "grad_norm": 1.6215320825576782,
      "learning_rate": 1.0338983050847457e-06,
      "loss": 0.8292,
      "mean_token_accuracy": 0.7751116752624512,
      "num_tokens": 9903156.0,
      "step": 62
    },
    {
      "epoch": 0.03204476093591048,
      "grad_norm": 1.7067248821258545,
      "learning_rate": 1.0508474576271187e-06,
      "loss": 0.8219,
      "mean_token_accuracy": 0.7756357192993164,
      "num_tokens": 10054899.0,
      "step": 63
    },
    {
      "epoch": 0.03255340793489318,
      "grad_norm": 1.7187061309814453,
      "learning_rate": 1.0677966101694917e-06,
      "loss": 0.7922,
      "mean_token_accuracy": 0.7845475673675537,
      "num_tokens": 10214147.0,
      "step": 64
    },
    {
      "epoch": 0.03306205493387589,
      "grad_norm": 1.8380779027938843,
      "learning_rate": 1.0847457627118644e-06,
      "loss": 0.7682,
      "mean_token_accuracy": 0.7888185977935791,
      "num_tokens": 10376357.0,
      "step": 65
    },
    {
      "epoch": 0.0335707019328586,
      "grad_norm": 1.7338484525680542,
      "learning_rate": 1.1016949152542374e-06,
      "loss": 0.8376,
      "mean_token_accuracy": 0.7710230350494385,
      "num_tokens": 10528001.0,
      "step": 66
    },
    {
      "epoch": 0.0340793489318413,
      "grad_norm": 1.5959792137145996,
      "learning_rate": 1.1186440677966102e-06,
      "loss": 0.8306,
      "mean_token_accuracy": 0.7736530303955078,
      "num_tokens": 10697458.0,
      "step": 67
    },
    {
      "epoch": 0.03458799593082401,
      "grad_norm": 1.5118943452835083,
      "learning_rate": 1.1355932203389832e-06,
      "loss": 0.8437,
      "mean_token_accuracy": 0.7709776759147644,
      "num_tokens": 10852213.0,
      "step": 68
    },
    {
      "epoch": 0.035096642929806715,
      "grad_norm": 1.425594449043274,
      "learning_rate": 1.152542372881356e-06,
      "loss": 0.803,
      "mean_token_accuracy": 0.7797463536262512,
      "num_tokens": 11018742.0,
      "step": 69
    },
    {
      "epoch": 0.03560528992878942,
      "grad_norm": 1.3931862115859985,
      "learning_rate": 1.169491525423729e-06,
      "loss": 0.8333,
      "mean_token_accuracy": 0.7734776735305786,
      "num_tokens": 11178241.0,
      "step": 70
    },
    {
      "epoch": 0.03611393692777213,
      "grad_norm": 1.4249283075332642,
      "learning_rate": 1.186440677966102e-06,
      "loss": 0.7853,
      "mean_token_accuracy": 0.7858003377914429,
      "num_tokens": 11339213.0,
      "step": 71
    },
    {
      "epoch": 0.03662258392675483,
      "grad_norm": 1.4698734283447266,
      "learning_rate": 1.2033898305084747e-06,
      "loss": 0.7764,
      "mean_token_accuracy": 0.7849376797676086,
      "num_tokens": 11489548.0,
      "step": 72
    },
    {
      "epoch": 0.037131230925737536,
      "grad_norm": 1.3906108140945435,
      "learning_rate": 1.2203389830508477e-06,
      "loss": 0.7814,
      "mean_token_accuracy": 0.785656213760376,
      "num_tokens": 11656772.0,
      "step": 73
    },
    {
      "epoch": 0.03763987792472025,
      "grad_norm": 1.2897934913635254,
      "learning_rate": 1.2372881355932204e-06,
      "loss": 0.7617,
      "mean_token_accuracy": 0.7886397838592529,
      "num_tokens": 11818390.0,
      "step": 74
    },
    {
      "epoch": 0.03814852492370295,
      "grad_norm": 1.343746304512024,
      "learning_rate": 1.2542372881355932e-06,
      "loss": 0.811,
      "mean_token_accuracy": 0.7766581773757935,
      "num_tokens": 11965717.0,
      "step": 75
    },
    {
      "epoch": 0.038657171922685654,
      "grad_norm": 1.2226605415344238,
      "learning_rate": 1.2711864406779662e-06,
      "loss": 0.7907,
      "mean_token_accuracy": 0.7826157808303833,
      "num_tokens": 12131732.0,
      "step": 76
    },
    {
      "epoch": 0.039165818921668365,
      "grad_norm": 1.3067526817321777,
      "learning_rate": 1.288135593220339e-06,
      "loss": 0.8071,
      "mean_token_accuracy": 0.777775764465332,
      "num_tokens": 12283927.0,
      "step": 77
    },
    {
      "epoch": 0.03967446592065107,
      "grad_norm": 1.277056336402893,
      "learning_rate": 1.305084745762712e-06,
      "loss": 0.7693,
      "mean_token_accuracy": 0.7876085042953491,
      "num_tokens": 12436906.0,
      "step": 78
    },
    {
      "epoch": 0.04018311291963377,
      "grad_norm": 1.2849011421203613,
      "learning_rate": 1.322033898305085e-06,
      "loss": 0.7794,
      "mean_token_accuracy": 0.7832804918289185,
      "num_tokens": 12600258.0,
      "step": 79
    },
    {
      "epoch": 0.04069175991861648,
      "grad_norm": 1.2474393844604492,
      "learning_rate": 1.3389830508474577e-06,
      "loss": 0.7745,
      "mean_token_accuracy": 0.7856120467185974,
      "num_tokens": 12747689.0,
      "step": 80
    },
    {
      "epoch": 0.041200406917599186,
      "grad_norm": 1.144174337387085,
      "learning_rate": 1.3559322033898307e-06,
      "loss": 0.7597,
      "mean_token_accuracy": 0.7897878885269165,
      "num_tokens": 12911145.0,
      "step": 81
    },
    {
      "epoch": 0.04170905391658189,
      "grad_norm": 1.2651646137237549,
      "learning_rate": 1.3728813559322034e-06,
      "loss": 0.7638,
      "mean_token_accuracy": 0.7870814800262451,
      "num_tokens": 13081387.0,
      "step": 82
    },
    {
      "epoch": 0.0422177009155646,
      "grad_norm": 1.167580485343933,
      "learning_rate": 1.3898305084745764e-06,
      "loss": 0.7787,
      "mean_token_accuracy": 0.7832891345024109,
      "num_tokens": 13243319.0,
      "step": 83
    },
    {
      "epoch": 0.042726347914547304,
      "grad_norm": 1.133839726448059,
      "learning_rate": 1.4067796610169494e-06,
      "loss": 0.7589,
      "mean_token_accuracy": 0.7909195423126221,
      "num_tokens": 13408622.0,
      "step": 84
    },
    {
      "epoch": 0.04323499491353001,
      "grad_norm": 1.2236430644989014,
      "learning_rate": 1.4237288135593222e-06,
      "loss": 0.7724,
      "mean_token_accuracy": 0.7842013835906982,
      "num_tokens": 13562201.0,
      "step": 85
    },
    {
      "epoch": 0.04374364191251272,
      "grad_norm": 1.2251627445220947,
      "learning_rate": 1.4406779661016951e-06,
      "loss": 0.76,
      "mean_token_accuracy": 0.7882030010223389,
      "num_tokens": 13706016.0,
      "step": 86
    },
    {
      "epoch": 0.04425228891149542,
      "grad_norm": 1.1422505378723145,
      "learning_rate": 1.457627118644068e-06,
      "loss": 0.7532,
      "mean_token_accuracy": 0.7909828424453735,
      "num_tokens": 13872736.0,
      "step": 87
    },
    {
      "epoch": 0.044760935910478125,
      "grad_norm": 1.164376139640808,
      "learning_rate": 1.4745762711864409e-06,
      "loss": 0.7102,
      "mean_token_accuracy": 0.8004215955734253,
      "num_tokens": 14029427.0,
      "step": 88
    },
    {
      "epoch": 0.045269582909460836,
      "grad_norm": 1.1551074981689453,
      "learning_rate": 1.4915254237288139e-06,
      "loss": 0.7426,
      "mean_token_accuracy": 0.791038990020752,
      "num_tokens": 14192123.0,
      "step": 89
    },
    {
      "epoch": 0.04577822990844354,
      "grad_norm": 1.1146512031555176,
      "learning_rate": 1.5084745762711866e-06,
      "loss": 0.7181,
      "mean_token_accuracy": 0.7969399690628052,
      "num_tokens": 14356828.0,
      "step": 90
    },
    {
      "epoch": 0.04628687690742624,
      "grad_norm": 1.0763921737670898,
      "learning_rate": 1.5254237288135596e-06,
      "loss": 0.6807,
      "mean_token_accuracy": 0.8067453503608704,
      "num_tokens": 14522417.0,
      "step": 91
    },
    {
      "epoch": 0.04679552390640895,
      "grad_norm": 1.1564586162567139,
      "learning_rate": 1.5423728813559324e-06,
      "loss": 0.7619,
      "mean_token_accuracy": 0.7866380214691162,
      "num_tokens": 14675485.0,
      "step": 92
    },
    {
      "epoch": 0.04730417090539166,
      "grad_norm": 1.138609528541565,
      "learning_rate": 1.5593220338983054e-06,
      "loss": 0.7253,
      "mean_token_accuracy": 0.7969965934753418,
      "num_tokens": 14831686.0,
      "step": 93
    },
    {
      "epoch": 0.04781281790437437,
      "grad_norm": 1.1271717548370361,
      "learning_rate": 1.5762711864406781e-06,
      "loss": 0.7507,
      "mean_token_accuracy": 0.7900179624557495,
      "num_tokens": 14983974.0,
      "step": 94
    },
    {
      "epoch": 0.04832146490335707,
      "grad_norm": 1.1798120737075806,
      "learning_rate": 1.593220338983051e-06,
      "loss": 0.7766,
      "mean_token_accuracy": 0.7809305191040039,
      "num_tokens": 15135323.0,
      "step": 95
    },
    {
      "epoch": 0.048830111902339775,
      "grad_norm": 1.2915998697280884,
      "learning_rate": 1.6101694915254237e-06,
      "loss": 0.7481,
      "mean_token_accuracy": 0.7903225421905518,
      "num_tokens": 15289033.0,
      "step": 96
    },
    {
      "epoch": 0.049338758901322485,
      "grad_norm": 1.1710273027420044,
      "learning_rate": 1.6271186440677967e-06,
      "loss": 0.7504,
      "mean_token_accuracy": 0.7876750826835632,
      "num_tokens": 15439451.0,
      "step": 97
    },
    {
      "epoch": 0.04984740590030519,
      "grad_norm": 1.1403133869171143,
      "learning_rate": 1.6440677966101694e-06,
      "loss": 0.7274,
      "mean_token_accuracy": 0.7948034405708313,
      "num_tokens": 15594811.0,
      "step": 98
    },
    {
      "epoch": 0.05035605289928789,
      "grad_norm": 1.1216124296188354,
      "learning_rate": 1.6610169491525424e-06,
      "loss": 0.7363,
      "mean_token_accuracy": 0.7924157381057739,
      "num_tokens": 15758944.0,
      "step": 99
    },
    {
      "epoch": 0.0508646998982706,
      "grad_norm": 1.1310335397720337,
      "learning_rate": 1.6779661016949154e-06,
      "loss": 0.6821,
      "mean_token_accuracy": 0.8055119514465332,
      "num_tokens": 15922340.0,
      "step": 100
    },
    {
      "epoch": 0.05137334689725331,
      "grad_norm": 1.1045717000961304,
      "learning_rate": 1.6949152542372882e-06,
      "loss": 0.701,
      "mean_token_accuracy": 0.8013116121292114,
      "num_tokens": 16074162.0,
      "step": 101
    },
    {
      "epoch": 0.05188199389623601,
      "grad_norm": 1.141825556755066,
      "learning_rate": 1.7118644067796611e-06,
      "loss": 0.7233,
      "mean_token_accuracy": 0.7949929237365723,
      "num_tokens": 16238658.0,
      "step": 102
    },
    {
      "epoch": 0.05239064089521872,
      "grad_norm": 1.1329325437545776,
      "learning_rate": 1.728813559322034e-06,
      "loss": 0.6977,
      "mean_token_accuracy": 0.8014352321624756,
      "num_tokens": 16396393.0,
      "step": 103
    },
    {
      "epoch": 0.052899287894201424,
      "grad_norm": 1.1402873992919922,
      "learning_rate": 1.7457627118644069e-06,
      "loss": 0.7219,
      "mean_token_accuracy": 0.7958400249481201,
      "num_tokens": 16546360.0,
      "step": 104
    },
    {
      "epoch": 0.05340793489318413,
      "grad_norm": 1.116026759147644,
      "learning_rate": 1.7627118644067799e-06,
      "loss": 0.7024,
      "mean_token_accuracy": 0.7998898029327393,
      "num_tokens": 16704979.0,
      "step": 105
    },
    {
      "epoch": 0.05391658189216684,
      "grad_norm": 1.0816538333892822,
      "learning_rate": 1.7796610169491526e-06,
      "loss": 0.6952,
      "mean_token_accuracy": 0.8007581233978271,
      "num_tokens": 16860358.0,
      "step": 106
    },
    {
      "epoch": 0.05442522889114954,
      "grad_norm": 1.114751935005188,
      "learning_rate": 1.7966101694915256e-06,
      "loss": 0.7154,
      "mean_token_accuracy": 0.797700047492981,
      "num_tokens": 17034067.0,
      "step": 107
    },
    {
      "epoch": 0.054933875890132246,
      "grad_norm": 1.2079203128814697,
      "learning_rate": 1.8135593220338984e-06,
      "loss": 0.747,
      "mean_token_accuracy": 0.7880150079727173,
      "num_tokens": 17181452.0,
      "step": 108
    },
    {
      "epoch": 0.055442522889114956,
      "grad_norm": 4.387100696563721,
      "learning_rate": 1.8305084745762714e-06,
      "loss": 0.6798,
      "mean_token_accuracy": 0.8064528107643127,
      "num_tokens": 17338963.0,
      "step": 109
    },
    {
      "epoch": 0.05595116988809766,
      "grad_norm": 1.2586617469787598,
      "learning_rate": 1.8474576271186441e-06,
      "loss": 0.7127,
      "mean_token_accuracy": 0.7975611686706543,
      "num_tokens": 17493379.0,
      "step": 110
    },
    {
      "epoch": 0.056459816887080364,
      "grad_norm": 1.1230461597442627,
      "learning_rate": 1.8644067796610171e-06,
      "loss": 0.6778,
      "mean_token_accuracy": 0.8056212663650513,
      "num_tokens": 17651741.0,
      "step": 111
    },
    {
      "epoch": 0.056968463886063074,
      "grad_norm": 1.1480046510696411,
      "learning_rate": 1.88135593220339e-06,
      "loss": 0.6992,
      "mean_token_accuracy": 0.7992340922355652,
      "num_tokens": 17810924.0,
      "step": 112
    },
    {
      "epoch": 0.05747711088504578,
      "grad_norm": 1.0441348552703857,
      "learning_rate": 1.8983050847457629e-06,
      "loss": 0.6995,
      "mean_token_accuracy": 0.8028820753097534,
      "num_tokens": 17976758.0,
      "step": 113
    },
    {
      "epoch": 0.05798575788402848,
      "grad_norm": 1.1486817598342896,
      "learning_rate": 1.9152542372881356e-06,
      "loss": 0.7392,
      "mean_token_accuracy": 0.7913587689399719,
      "num_tokens": 18134320.0,
      "step": 114
    },
    {
      "epoch": 0.05849440488301119,
      "grad_norm": 1.22743821144104,
      "learning_rate": 1.932203389830509e-06,
      "loss": 0.7303,
      "mean_token_accuracy": 0.7928489446640015,
      "num_tokens": 18276678.0,
      "step": 115
    },
    {
      "epoch": 0.059003051881993895,
      "grad_norm": 1.9819073677062988,
      "learning_rate": 1.9491525423728816e-06,
      "loss": 0.6889,
      "mean_token_accuracy": 0.8051313161849976,
      "num_tokens": 18426725.0,
      "step": 116
    },
    {
      "epoch": 0.0595116988809766,
      "grad_norm": 1.3526983261108398,
      "learning_rate": 1.9661016949152544e-06,
      "loss": 0.7195,
      "mean_token_accuracy": 0.7980197668075562,
      "num_tokens": 18587522.0,
      "step": 117
    },
    {
      "epoch": 0.06002034587995931,
      "grad_norm": 1.1162952184677124,
      "learning_rate": 1.9830508474576276e-06,
      "loss": 0.7087,
      "mean_token_accuracy": 0.7977943420410156,
      "num_tokens": 18749574.0,
      "step": 118
    },
    {
      "epoch": 0.06052899287894201,
      "grad_norm": 1.1709731817245483,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.6925,
      "mean_token_accuracy": 0.8042123317718506,
      "num_tokens": 18906800.0,
      "step": 119
    },
    {
      "epoch": 0.061037639877924724,
      "grad_norm": 1.1323966979980469,
      "learning_rate": 2.016949152542373e-06,
      "loss": 0.7452,
      "mean_token_accuracy": 0.7882239818572998,
      "num_tokens": 19068637.0,
      "step": 120
    },
    {
      "epoch": 0.06154628687690743,
      "grad_norm": 1.1877466440200806,
      "learning_rate": 2.033898305084746e-06,
      "loss": 0.7036,
      "mean_token_accuracy": 0.7979512810707092,
      "num_tokens": 19219209.0,
      "step": 121
    },
    {
      "epoch": 0.06205493387589013,
      "grad_norm": 1.1348536014556885,
      "learning_rate": 2.0508474576271186e-06,
      "loss": 0.6824,
      "mean_token_accuracy": 0.8026146292686462,
      "num_tokens": 19367232.0,
      "step": 122
    },
    {
      "epoch": 0.06256358087487283,
      "grad_norm": 1.0699137449264526,
      "learning_rate": 2.0677966101694914e-06,
      "loss": 0.677,
      "mean_token_accuracy": 0.8075043559074402,
      "num_tokens": 19525992.0,
      "step": 123
    },
    {
      "epoch": 0.06307222787385554,
      "grad_norm": 1.1170971393585205,
      "learning_rate": 2.0847457627118646e-06,
      "loss": 0.6963,
      "mean_token_accuracy": 0.8016464710235596,
      "num_tokens": 19693698.0,
      "step": 124
    },
    {
      "epoch": 0.06358087487283826,
      "grad_norm": 1.1918888092041016,
      "learning_rate": 2.1016949152542374e-06,
      "loss": 0.683,
      "mean_token_accuracy": 0.8041555881500244,
      "num_tokens": 19841203.0,
      "step": 125
    },
    {
      "epoch": 0.06408952187182096,
      "grad_norm": 1.1442811489105225,
      "learning_rate": 2.11864406779661e-06,
      "loss": 0.7213,
      "mean_token_accuracy": 0.7939434051513672,
      "num_tokens": 20001680.0,
      "step": 126
    },
    {
      "epoch": 0.06459816887080366,
      "grad_norm": 1.1511716842651367,
      "learning_rate": 2.1355932203389833e-06,
      "loss": 0.6835,
      "mean_token_accuracy": 0.8039097785949707,
      "num_tokens": 20150787.0,
      "step": 127
    },
    {
      "epoch": 0.06510681586978637,
      "grad_norm": 1.0827454328536987,
      "learning_rate": 2.152542372881356e-06,
      "loss": 0.6907,
      "mean_token_accuracy": 0.8021340370178223,
      "num_tokens": 20326096.0,
      "step": 128
    },
    {
      "epoch": 0.06561546286876907,
      "grad_norm": 1.1005094051361084,
      "learning_rate": 2.169491525423729e-06,
      "loss": 0.7339,
      "mean_token_accuracy": 0.7913875579833984,
      "num_tokens": 20490433.0,
      "step": 129
    },
    {
      "epoch": 0.06612410986775177,
      "grad_norm": 1.0656397342681885,
      "learning_rate": 2.1864406779661016e-06,
      "loss": 0.7223,
      "mean_token_accuracy": 0.7938638925552368,
      "num_tokens": 20657596.0,
      "step": 130
    },
    {
      "epoch": 0.06663275686673449,
      "grad_norm": 1.0604357719421387,
      "learning_rate": 2.203389830508475e-06,
      "loss": 0.6711,
      "mean_token_accuracy": 0.8067337870597839,
      "num_tokens": 20827354.0,
      "step": 131
    },
    {
      "epoch": 0.0671414038657172,
      "grad_norm": 1.1440101861953735,
      "learning_rate": 2.2203389830508476e-06,
      "loss": 0.704,
      "mean_token_accuracy": 0.799992024898529,
      "num_tokens": 20978319.0,
      "step": 132
    },
    {
      "epoch": 0.0676500508646999,
      "grad_norm": 1.1271777153015137,
      "learning_rate": 2.2372881355932204e-06,
      "loss": 0.6845,
      "mean_token_accuracy": 0.8048183917999268,
      "num_tokens": 21124807.0,
      "step": 133
    },
    {
      "epoch": 0.0681586978636826,
      "grad_norm": 1.1711620092391968,
      "learning_rate": 2.2542372881355936e-06,
      "loss": 0.6677,
      "mean_token_accuracy": 0.8087946176528931,
      "num_tokens": 21283777.0,
      "step": 134
    },
    {
      "epoch": 0.0686673448626653,
      "grad_norm": 1.1323461532592773,
      "learning_rate": 2.2711864406779663e-06,
      "loss": 0.682,
      "mean_token_accuracy": 0.8036584258079529,
      "num_tokens": 21436297.0,
      "step": 135
    },
    {
      "epoch": 0.06917599186164802,
      "grad_norm": 1.1845351457595825,
      "learning_rate": 2.288135593220339e-06,
      "loss": 0.6991,
      "mean_token_accuracy": 0.7986134886741638,
      "num_tokens": 21590885.0,
      "step": 136
    },
    {
      "epoch": 0.06968463886063073,
      "grad_norm": 1.139819622039795,
      "learning_rate": 2.305084745762712e-06,
      "loss": 0.7272,
      "mean_token_accuracy": 0.7913652062416077,
      "num_tokens": 21746247.0,
      "step": 137
    },
    {
      "epoch": 0.07019328585961343,
      "grad_norm": 1.1518213748931885,
      "learning_rate": 2.322033898305085e-06,
      "loss": 0.6775,
      "mean_token_accuracy": 0.8056344985961914,
      "num_tokens": 21898426.0,
      "step": 138
    },
    {
      "epoch": 0.07070193285859613,
      "grad_norm": 1.1628955602645874,
      "learning_rate": 2.338983050847458e-06,
      "loss": 0.7437,
      "mean_token_accuracy": 0.7898159027099609,
      "num_tokens": 22054155.0,
      "step": 139
    },
    {
      "epoch": 0.07121057985757884,
      "grad_norm": 1.1349354982376099,
      "learning_rate": 2.3559322033898306e-06,
      "loss": 0.7113,
      "mean_token_accuracy": 0.7953236103057861,
      "num_tokens": 22215818.0,
      "step": 140
    },
    {
      "epoch": 0.07171922685656154,
      "grad_norm": 1.136148452758789,
      "learning_rate": 2.372881355932204e-06,
      "loss": 0.6908,
      "mean_token_accuracy": 0.8000525832176208,
      "num_tokens": 22367981.0,
      "step": 141
    },
    {
      "epoch": 0.07222787385554426,
      "grad_norm": 1.0790364742279053,
      "learning_rate": 2.3898305084745766e-06,
      "loss": 0.689,
      "mean_token_accuracy": 0.802665114402771,
      "num_tokens": 22537203.0,
      "step": 142
    },
    {
      "epoch": 0.07273652085452696,
      "grad_norm": 1.1381577253341675,
      "learning_rate": 2.4067796610169493e-06,
      "loss": 0.6895,
      "mean_token_accuracy": 0.801916241645813,
      "num_tokens": 22696776.0,
      "step": 143
    },
    {
      "epoch": 0.07324516785350967,
      "grad_norm": 1.1592439413070679,
      "learning_rate": 2.4237288135593225e-06,
      "loss": 0.6637,
      "mean_token_accuracy": 0.8090922236442566,
      "num_tokens": 22850630.0,
      "step": 144
    },
    {
      "epoch": 0.07375381485249237,
      "grad_norm": 1.1903307437896729,
      "learning_rate": 2.4406779661016953e-06,
      "loss": 0.7183,
      "mean_token_accuracy": 0.7970988750457764,
      "num_tokens": 23002043.0,
      "step": 145
    },
    {
      "epoch": 0.07426246185147507,
      "grad_norm": 1.135934591293335,
      "learning_rate": 2.457627118644068e-06,
      "loss": 0.6473,
      "mean_token_accuracy": 0.8129653930664062,
      "num_tokens": 23159234.0,
      "step": 146
    },
    {
      "epoch": 0.07477110885045778,
      "grad_norm": 1.1974645853042603,
      "learning_rate": 2.474576271186441e-06,
      "loss": 0.6752,
      "mean_token_accuracy": 0.8057656288146973,
      "num_tokens": 23310355.0,
      "step": 147
    },
    {
      "epoch": 0.0752797558494405,
      "grad_norm": 1.0872167348861694,
      "learning_rate": 2.491525423728814e-06,
      "loss": 0.664,
      "mean_token_accuracy": 0.8108680248260498,
      "num_tokens": 23473092.0,
      "step": 148
    },
    {
      "epoch": 0.0757884028484232,
      "grad_norm": 1.1531084775924683,
      "learning_rate": 2.5084745762711864e-06,
      "loss": 0.68,
      "mean_token_accuracy": 0.8035491704940796,
      "num_tokens": 23627766.0,
      "step": 149
    },
    {
      "epoch": 0.0762970498474059,
      "grad_norm": 1.096634864807129,
      "learning_rate": 2.5254237288135596e-06,
      "loss": 0.6537,
      "mean_token_accuracy": 0.812416672706604,
      "num_tokens": 23777307.0,
      "step": 150
    },
    {
      "epoch": 0.0768056968463886,
      "grad_norm": 1.1573667526245117,
      "learning_rate": 2.5423728813559323e-06,
      "loss": 0.6942,
      "mean_token_accuracy": 0.7996103763580322,
      "num_tokens": 23923715.0,
      "step": 151
    },
    {
      "epoch": 0.07731434384537131,
      "grad_norm": 1.1179906129837036,
      "learning_rate": 2.5593220338983055e-06,
      "loss": 0.6794,
      "mean_token_accuracy": 0.8048681020736694,
      "num_tokens": 24087522.0,
      "step": 152
    },
    {
      "epoch": 0.07782299084435401,
      "grad_norm": 1.2234212160110474,
      "learning_rate": 2.576271186440678e-06,
      "loss": 0.6577,
      "mean_token_accuracy": 0.8093551993370056,
      "num_tokens": 24231067.0,
      "step": 153
    },
    {
      "epoch": 0.07833163784333673,
      "grad_norm": 1.1112676858901978,
      "learning_rate": 2.593220338983051e-06,
      "loss": 0.6982,
      "mean_token_accuracy": 0.8004816770553589,
      "num_tokens": 24394212.0,
      "step": 154
    },
    {
      "epoch": 0.07884028484231943,
      "grad_norm": 1.2034029960632324,
      "learning_rate": 2.610169491525424e-06,
      "loss": 0.6691,
      "mean_token_accuracy": 0.8068128824234009,
      "num_tokens": 24539665.0,
      "step": 155
    },
    {
      "epoch": 0.07934893184130214,
      "grad_norm": 1.1976460218429565,
      "learning_rate": 2.627118644067797e-06,
      "loss": 0.7036,
      "mean_token_accuracy": 0.7969769239425659,
      "num_tokens": 24703099.0,
      "step": 156
    },
    {
      "epoch": 0.07985757884028484,
      "grad_norm": 4.0496087074279785,
      "learning_rate": 2.64406779661017e-06,
      "loss": 0.6776,
      "mean_token_accuracy": 0.8049623370170593,
      "num_tokens": 24851707.0,
      "step": 157
    },
    {
      "epoch": 0.08036622583926754,
      "grad_norm": 1.2107317447662354,
      "learning_rate": 2.661016949152543e-06,
      "loss": 0.6605,
      "mean_token_accuracy": 0.8087326288223267,
      "num_tokens": 25018302.0,
      "step": 158
    },
    {
      "epoch": 0.08087487283825025,
      "grad_norm": 1.1594712734222412,
      "learning_rate": 2.6779661016949153e-06,
      "loss": 0.6659,
      "mean_token_accuracy": 0.8069348335266113,
      "num_tokens": 25177030.0,
      "step": 159
    },
    {
      "epoch": 0.08138351983723296,
      "grad_norm": 1.0774612426757812,
      "learning_rate": 2.6949152542372885e-06,
      "loss": 0.6909,
      "mean_token_accuracy": 0.80242520570755,
      "num_tokens": 25345695.0,
      "step": 160
    },
    {
      "epoch": 0.08189216683621567,
      "grad_norm": 1.034545660018921,
      "learning_rate": 2.7118644067796613e-06,
      "loss": 0.6396,
      "mean_token_accuracy": 0.8151862025260925,
      "num_tokens": 25513624.0,
      "step": 161
    },
    {
      "epoch": 0.08240081383519837,
      "grad_norm": 1.1746925115585327,
      "learning_rate": 2.7288135593220336e-06,
      "loss": 0.681,
      "mean_token_accuracy": 0.8041950464248657,
      "num_tokens": 25674463.0,
      "step": 162
    },
    {
      "epoch": 0.08290946083418108,
      "grad_norm": 1.139805555343628,
      "learning_rate": 2.745762711864407e-06,
      "loss": 0.6566,
      "mean_token_accuracy": 0.8090215921401978,
      "num_tokens": 25834424.0,
      "step": 163
    },
    {
      "epoch": 0.08341810783316378,
      "grad_norm": 1.5135971307754517,
      "learning_rate": 2.7627118644067796e-06,
      "loss": 0.6966,
      "mean_token_accuracy": 0.7990592122077942,
      "num_tokens": 25991547.0,
      "step": 164
    },
    {
      "epoch": 0.0839267548321465,
      "grad_norm": 1.1292972564697266,
      "learning_rate": 2.779661016949153e-06,
      "loss": 0.6788,
      "mean_token_accuracy": 0.8033425807952881,
      "num_tokens": 26151034.0,
      "step": 165
    },
    {
      "epoch": 0.0844354018311292,
      "grad_norm": 1.1054205894470215,
      "learning_rate": 2.7966101694915256e-06,
      "loss": 0.6489,
      "mean_token_accuracy": 0.8113046288490295,
      "num_tokens": 26310637.0,
      "step": 166
    },
    {
      "epoch": 0.0849440488301119,
      "grad_norm": 1.0914794206619263,
      "learning_rate": 2.8135593220338988e-06,
      "loss": 0.6457,
      "mean_token_accuracy": 0.8134191632270813,
      "num_tokens": 26465678.0,
      "step": 167
    },
    {
      "epoch": 0.08545269582909461,
      "grad_norm": 1.0801182985305786,
      "learning_rate": 2.830508474576271e-06,
      "loss": 0.676,
      "mean_token_accuracy": 0.8042765855789185,
      "num_tokens": 26632464.0,
      "step": 168
    },
    {
      "epoch": 0.08596134282807731,
      "grad_norm": 1.1082203388214111,
      "learning_rate": 2.8474576271186443e-06,
      "loss": 0.6338,
      "mean_token_accuracy": 0.8155921697616577,
      "num_tokens": 26795346.0,
      "step": 169
    },
    {
      "epoch": 0.08646998982706001,
      "grad_norm": 1.0745552778244019,
      "learning_rate": 2.864406779661017e-06,
      "loss": 0.6937,
      "mean_token_accuracy": 0.7970446944236755,
      "num_tokens": 26951965.0,
      "step": 170
    },
    {
      "epoch": 0.08697863682604273,
      "grad_norm": 1.1002435684204102,
      "learning_rate": 2.8813559322033903e-06,
      "loss": 0.6389,
      "mean_token_accuracy": 0.8136718273162842,
      "num_tokens": 27113922.0,
      "step": 171
    },
    {
      "epoch": 0.08748728382502544,
      "grad_norm": 1.087834358215332,
      "learning_rate": 2.8983050847457626e-06,
      "loss": 0.6479,
      "mean_token_accuracy": 0.8113669157028198,
      "num_tokens": 27279172.0,
      "step": 172
    },
    {
      "epoch": 0.08799593082400814,
      "grad_norm": 1.1026259660720825,
      "learning_rate": 2.915254237288136e-06,
      "loss": 0.6726,
      "mean_token_accuracy": 0.805874764919281,
      "num_tokens": 27449975.0,
      "step": 173
    },
    {
      "epoch": 0.08850457782299084,
      "grad_norm": 1.106402039527893,
      "learning_rate": 2.9322033898305086e-06,
      "loss": 0.684,
      "mean_token_accuracy": 0.8012702465057373,
      "num_tokens": 27595872.0,
      "step": 174
    },
    {
      "epoch": 0.08901322482197355,
      "grad_norm": 1.1090171337127686,
      "learning_rate": 2.9491525423728818e-06,
      "loss": 0.6552,
      "mean_token_accuracy": 0.8108108043670654,
      "num_tokens": 27760835.0,
      "step": 175
    },
    {
      "epoch": 0.08952187182095625,
      "grad_norm": 1.0845763683319092,
      "learning_rate": 2.9661016949152545e-06,
      "loss": 0.6519,
      "mean_token_accuracy": 0.810937225818634,
      "num_tokens": 27923943.0,
      "step": 176
    },
    {
      "epoch": 0.09003051881993897,
      "grad_norm": 1.1143302917480469,
      "learning_rate": 2.9830508474576277e-06,
      "loss": 0.6752,
      "mean_token_accuracy": 0.8048059344291687,
      "num_tokens": 28085891.0,
      "step": 177
    },
    {
      "epoch": 0.09053916581892167,
      "grad_norm": 1.1437448263168335,
      "learning_rate": 3e-06,
      "loss": 0.6196,
      "mean_token_accuracy": 0.818387508392334,
      "num_tokens": 28243518.0,
      "step": 178
    },
    {
      "epoch": 0.09104781281790437,
      "grad_norm": 1.0629551410675049,
      "learning_rate": 3.0169491525423733e-06,
      "loss": 0.6494,
      "mean_token_accuracy": 0.8105939030647278,
      "num_tokens": 28405483.0,
      "step": 179
    },
    {
      "epoch": 0.09155645981688708,
      "grad_norm": 1.1684266328811646,
      "learning_rate": 3.033898305084746e-06,
      "loss": 0.6803,
      "mean_token_accuracy": 0.8017261028289795,
      "num_tokens": 28558441.0,
      "step": 180
    },
    {
      "epoch": 0.09206510681586978,
      "grad_norm": 1.1008362770080566,
      "learning_rate": 3.0508474576271192e-06,
      "loss": 0.6324,
      "mean_token_accuracy": 0.815455436706543,
      "num_tokens": 28707646.0,
      "step": 181
    },
    {
      "epoch": 0.09257375381485249,
      "grad_norm": 1.1428719758987427,
      "learning_rate": 3.0677966101694916e-06,
      "loss": 0.6594,
      "mean_token_accuracy": 0.8092727065086365,
      "num_tokens": 28877502.0,
      "step": 182
    },
    {
      "epoch": 0.0930824008138352,
      "grad_norm": 1.1864022016525269,
      "learning_rate": 3.0847457627118648e-06,
      "loss": 0.7025,
      "mean_token_accuracy": 0.7980859279632568,
      "num_tokens": 29040815.0,
      "step": 183
    },
    {
      "epoch": 0.0935910478128179,
      "grad_norm": 1.1755472421646118,
      "learning_rate": 3.1016949152542375e-06,
      "loss": 0.6417,
      "mean_token_accuracy": 0.8110471963882446,
      "num_tokens": 29189342.0,
      "step": 184
    },
    {
      "epoch": 0.09409969481180061,
      "grad_norm": 1.1948294639587402,
      "learning_rate": 3.1186440677966107e-06,
      "loss": 0.6632,
      "mean_token_accuracy": 0.809227466583252,
      "num_tokens": 29361787.0,
      "step": 185
    },
    {
      "epoch": 0.09460834181078331,
      "grad_norm": 1.1376807689666748,
      "learning_rate": 3.135593220338983e-06,
      "loss": 0.5833,
      "mean_token_accuracy": 0.8278385400772095,
      "num_tokens": 29523929.0,
      "step": 186
    },
    {
      "epoch": 0.09511698880976602,
      "grad_norm": 1.181722640991211,
      "learning_rate": 3.1525423728813563e-06,
      "loss": 0.6812,
      "mean_token_accuracy": 0.8037586212158203,
      "num_tokens": 29682593.0,
      "step": 187
    },
    {
      "epoch": 0.09562563580874874,
      "grad_norm": 1.120546817779541,
      "learning_rate": 3.169491525423729e-06,
      "loss": 0.6301,
      "mean_token_accuracy": 0.8173536062240601,
      "num_tokens": 29839653.0,
      "step": 188
    },
    {
      "epoch": 0.09613428280773144,
      "grad_norm": 1.2152618169784546,
      "learning_rate": 3.186440677966102e-06,
      "loss": 0.6717,
      "mean_token_accuracy": 0.8038546442985535,
      "num_tokens": 30001594.0,
      "step": 189
    },
    {
      "epoch": 0.09664292980671414,
      "grad_norm": 1.130861759185791,
      "learning_rate": 3.203389830508475e-06,
      "loss": 0.6589,
      "mean_token_accuracy": 0.8086852431297302,
      "num_tokens": 30163501.0,
      "step": 190
    },
    {
      "epoch": 0.09715157680569685,
      "grad_norm": 1.117408037185669,
      "learning_rate": 3.2203389830508473e-06,
      "loss": 0.642,
      "mean_token_accuracy": 0.8118177056312561,
      "num_tokens": 30324797.0,
      "step": 191
    },
    {
      "epoch": 0.09766022380467955,
      "grad_norm": 1.0812642574310303,
      "learning_rate": 3.2372881355932205e-06,
      "loss": 0.6553,
      "mean_token_accuracy": 0.809535801410675,
      "num_tokens": 30492907.0,
      "step": 192
    },
    {
      "epoch": 0.09816887080366225,
      "grad_norm": 1.1507676839828491,
      "learning_rate": 3.2542372881355933e-06,
      "loss": 0.6121,
      "mean_token_accuracy": 0.8196245431900024,
      "num_tokens": 30667482.0,
      "step": 193
    },
    {
      "epoch": 0.09867751780264497,
      "grad_norm": 1.0629740953445435,
      "learning_rate": 3.2711864406779665e-06,
      "loss": 0.6429,
      "mean_token_accuracy": 0.8122947216033936,
      "num_tokens": 30830951.0,
      "step": 194
    },
    {
      "epoch": 0.09918616480162767,
      "grad_norm": 1.255733609199524,
      "learning_rate": 3.288135593220339e-06,
      "loss": 0.6417,
      "mean_token_accuracy": 0.8119453191757202,
      "num_tokens": 30992078.0,
      "step": 195
    },
    {
      "epoch": 0.09969481180061038,
      "grad_norm": 1.2162625789642334,
      "learning_rate": 3.305084745762712e-06,
      "loss": 0.6353,
      "mean_token_accuracy": 0.8144551515579224,
      "num_tokens": 31153089.0,
      "step": 196
    },
    {
      "epoch": 0.10020345879959308,
      "grad_norm": 1.1391067504882812,
      "learning_rate": 3.322033898305085e-06,
      "loss": 0.6408,
      "mean_token_accuracy": 0.8137332201004028,
      "num_tokens": 31297334.0,
      "step": 197
    },
    {
      "epoch": 0.10071210579857579,
      "grad_norm": 1.3500770330429077,
      "learning_rate": 3.338983050847458e-06,
      "loss": 0.6841,
      "mean_token_accuracy": 0.803351640701294,
      "num_tokens": 31462790.0,
      "step": 198
    },
    {
      "epoch": 0.10122075279755849,
      "grad_norm": 1.2101603746414185,
      "learning_rate": 3.3559322033898308e-06,
      "loss": 0.6686,
      "mean_token_accuracy": 0.8038467168807983,
      "num_tokens": 31614333.0,
      "step": 199
    },
    {
      "epoch": 0.1017293997965412,
      "grad_norm": 1.2734079360961914,
      "learning_rate": 3.372881355932204e-06,
      "loss": 0.6503,
      "mean_token_accuracy": 0.8107385039329529,
      "num_tokens": 31757983.0,
      "step": 200
    },
    {
      "epoch": 0.10223804679552391,
      "grad_norm": 1.1623011827468872,
      "learning_rate": 3.3898305084745763e-06,
      "loss": 0.6641,
      "mean_token_accuracy": 0.8052271604537964,
      "num_tokens": 31902198.0,
      "step": 201
    },
    {
      "epoch": 0.10274669379450661,
      "grad_norm": 1.1549586057662964,
      "learning_rate": 3.4067796610169495e-06,
      "loss": 0.6403,
      "mean_token_accuracy": 0.8139992952346802,
      "num_tokens": 32072015.0,
      "step": 202
    },
    {
      "epoch": 0.10325534079348932,
      "grad_norm": 1.25985586643219,
      "learning_rate": 3.4237288135593223e-06,
      "loss": 0.6587,
      "mean_token_accuracy": 0.8067578077316284,
      "num_tokens": 32231536.0,
      "step": 203
    },
    {
      "epoch": 0.10376398779247202,
      "grad_norm": 1.0511059761047363,
      "learning_rate": 3.4406779661016955e-06,
      "loss": 0.6077,
      "mean_token_accuracy": 0.8200663924217224,
      "num_tokens": 32401421.0,
      "step": 204
    },
    {
      "epoch": 0.10427263479145472,
      "grad_norm": 1.1063411235809326,
      "learning_rate": 3.457627118644068e-06,
      "loss": 0.6417,
      "mean_token_accuracy": 0.8133328557014465,
      "num_tokens": 32559615.0,
      "step": 205
    },
    {
      "epoch": 0.10478128179043744,
      "grad_norm": 1.2056820392608643,
      "learning_rate": 3.474576271186441e-06,
      "loss": 0.6428,
      "mean_token_accuracy": 0.8104065656661987,
      "num_tokens": 32710194.0,
      "step": 206
    },
    {
      "epoch": 0.10528992878942015,
      "grad_norm": 1.1128931045532227,
      "learning_rate": 3.4915254237288138e-06,
      "loss": 0.6477,
      "mean_token_accuracy": 0.8102855682373047,
      "num_tokens": 32878068.0,
      "step": 207
    },
    {
      "epoch": 0.10579857578840285,
      "grad_norm": 1.1195472478866577,
      "learning_rate": 3.508474576271187e-06,
      "loss": 0.6779,
      "mean_token_accuracy": 0.8018409013748169,
      "num_tokens": 33048273.0,
      "step": 208
    },
    {
      "epoch": 0.10630722278738555,
      "grad_norm": 1.1822260618209839,
      "learning_rate": 3.5254237288135597e-06,
      "loss": 0.6465,
      "mean_token_accuracy": 0.8118051886558533,
      "num_tokens": 33209096.0,
      "step": 209
    },
    {
      "epoch": 0.10681586978636826,
      "grad_norm": 1.11614990234375,
      "learning_rate": 3.5423728813559325e-06,
      "loss": 0.6205,
      "mean_token_accuracy": 0.8186913132667542,
      "num_tokens": 33375803.0,
      "step": 210
    },
    {
      "epoch": 0.10732451678535096,
      "grad_norm": 1.171306848526001,
      "learning_rate": 3.5593220338983053e-06,
      "loss": 0.6644,
      "mean_token_accuracy": 0.8052268028259277,
      "num_tokens": 33519566.0,
      "step": 211
    },
    {
      "epoch": 0.10783316378433368,
      "grad_norm": 1.1052064895629883,
      "learning_rate": 3.5762711864406785e-06,
      "loss": 0.6734,
      "mean_token_accuracy": 0.8042608499526978,
      "num_tokens": 33675958.0,
      "step": 212
    },
    {
      "epoch": 0.10834181078331638,
      "grad_norm": 1.1608377695083618,
      "learning_rate": 3.5932203389830512e-06,
      "loss": 0.6456,
      "mean_token_accuracy": 0.8108775615692139,
      "num_tokens": 33848664.0,
      "step": 213
    },
    {
      "epoch": 0.10885045778229908,
      "grad_norm": 1.1084113121032715,
      "learning_rate": 3.6101694915254244e-06,
      "loss": 0.6494,
      "mean_token_accuracy": 0.8081250190734863,
      "num_tokens": 34013911.0,
      "step": 214
    },
    {
      "epoch": 0.10935910478128179,
      "grad_norm": 1.183755874633789,
      "learning_rate": 3.6271186440677968e-06,
      "loss": 0.6292,
      "mean_token_accuracy": 0.8147969245910645,
      "num_tokens": 34175183.0,
      "step": 215
    },
    {
      "epoch": 0.10986775178026449,
      "grad_norm": 1.4397618770599365,
      "learning_rate": 3.6440677966101695e-06,
      "loss": 0.6242,
      "mean_token_accuracy": 0.8160824179649353,
      "num_tokens": 34337717.0,
      "step": 216
    },
    {
      "epoch": 0.11037639877924721,
      "grad_norm": 1.154085636138916,
      "learning_rate": 3.6610169491525427e-06,
      "loss": 0.657,
      "mean_token_accuracy": 0.807664155960083,
      "num_tokens": 34491024.0,
      "step": 217
    },
    {
      "epoch": 0.11088504577822991,
      "grad_norm": 1.061474323272705,
      "learning_rate": 3.6779661016949155e-06,
      "loss": 0.6206,
      "mean_token_accuracy": 0.8178257942199707,
      "num_tokens": 34668424.0,
      "step": 218
    },
    {
      "epoch": 0.11139369277721262,
      "grad_norm": 1.098557949066162,
      "learning_rate": 3.6949152542372883e-06,
      "loss": 0.6584,
      "mean_token_accuracy": 0.8083437085151672,
      "num_tokens": 34823916.0,
      "step": 219
    },
    {
      "epoch": 0.11190233977619532,
      "grad_norm": 1.1224589347839355,
      "learning_rate": 3.711864406779661e-06,
      "loss": 0.6324,
      "mean_token_accuracy": 0.8144303560256958,
      "num_tokens": 34986133.0,
      "step": 220
    },
    {
      "epoch": 0.11241098677517802,
      "grad_norm": 1.1159178018569946,
      "learning_rate": 3.7288135593220342e-06,
      "loss": 0.6126,
      "mean_token_accuracy": 0.8220010995864868,
      "num_tokens": 35146182.0,
      "step": 221
    },
    {
      "epoch": 0.11291963377416073,
      "grad_norm": 1.103893756866455,
      "learning_rate": 3.745762711864407e-06,
      "loss": 0.6566,
      "mean_token_accuracy": 0.8071621060371399,
      "num_tokens": 35301888.0,
      "step": 222
    },
    {
      "epoch": 0.11342828077314344,
      "grad_norm": 1.0538965463638306,
      "learning_rate": 3.76271186440678e-06,
      "loss": 0.6183,
      "mean_token_accuracy": 0.8180220723152161,
      "num_tokens": 35461669.0,
      "step": 223
    },
    {
      "epoch": 0.11393692777212615,
      "grad_norm": 1.122003436088562,
      "learning_rate": 3.7796610169491525e-06,
      "loss": 0.6365,
      "mean_token_accuracy": 0.8118059635162354,
      "num_tokens": 35621625.0,
      "step": 224
    },
    {
      "epoch": 0.11444557477110885,
      "grad_norm": 1.1433473825454712,
      "learning_rate": 3.7966101694915257e-06,
      "loss": 0.6216,
      "mean_token_accuracy": 0.8183375000953674,
      "num_tokens": 35773479.0,
      "step": 225
    },
    {
      "epoch": 0.11495422177009156,
      "grad_norm": 1.0675594806671143,
      "learning_rate": 3.8135593220338985e-06,
      "loss": 0.6141,
      "mean_token_accuracy": 0.8205496072769165,
      "num_tokens": 35936577.0,
      "step": 226
    },
    {
      "epoch": 0.11546286876907426,
      "grad_norm": 1.119829535484314,
      "learning_rate": 3.830508474576271e-06,
      "loss": 0.661,
      "mean_token_accuracy": 0.8067289590835571,
      "num_tokens": 36094734.0,
      "step": 227
    },
    {
      "epoch": 0.11597151576805696,
      "grad_norm": 1.115922451019287,
      "learning_rate": 3.8474576271186445e-06,
      "loss": 0.616,
      "mean_token_accuracy": 0.8180555105209351,
      "num_tokens": 36258492.0,
      "step": 228
    },
    {
      "epoch": 0.11648016276703968,
      "grad_norm": 1.1563063859939575,
      "learning_rate": 3.864406779661018e-06,
      "loss": 0.6357,
      "mean_token_accuracy": 0.8144519329071045,
      "num_tokens": 36423218.0,
      "step": 229
    },
    {
      "epoch": 0.11698880976602238,
      "grad_norm": 1.1410115957260132,
      "learning_rate": 3.88135593220339e-06,
      "loss": 0.6051,
      "mean_token_accuracy": 0.819241464138031,
      "num_tokens": 36569958.0,
      "step": 230
    },
    {
      "epoch": 0.11749745676500509,
      "grad_norm": 1.147270917892456,
      "learning_rate": 3.898305084745763e-06,
      "loss": 0.6577,
      "mean_token_accuracy": 0.807749330997467,
      "num_tokens": 36735839.0,
      "step": 231
    },
    {
      "epoch": 0.11800610376398779,
      "grad_norm": 1.240912675857544,
      "learning_rate": 3.9152542372881355e-06,
      "loss": 0.6424,
      "mean_token_accuracy": 0.8123055696487427,
      "num_tokens": 36886881.0,
      "step": 232
    },
    {
      "epoch": 0.1185147507629705,
      "grad_norm": 1.1100605726242065,
      "learning_rate": 3.932203389830509e-06,
      "loss": 0.636,
      "mean_token_accuracy": 0.8126474618911743,
      "num_tokens": 37055602.0,
      "step": 233
    },
    {
      "epoch": 0.1190233977619532,
      "grad_norm": 1.1091521978378296,
      "learning_rate": 3.949152542372882e-06,
      "loss": 0.6136,
      "mean_token_accuracy": 0.822191596031189,
      "num_tokens": 37215889.0,
      "step": 234
    },
    {
      "epoch": 0.11953204476093592,
      "grad_norm": 1.1006335020065308,
      "learning_rate": 3.966101694915255e-06,
      "loss": 0.619,
      "mean_token_accuracy": 0.8180478811264038,
      "num_tokens": 37368358.0,
      "step": 235
    },
    {
      "epoch": 0.12004069175991862,
      "grad_norm": 1.2350478172302246,
      "learning_rate": 3.9830508474576275e-06,
      "loss": 0.6275,
      "mean_token_accuracy": 0.8164553642272949,
      "num_tokens": 37528382.0,
      "step": 236
    },
    {
      "epoch": 0.12054933875890132,
      "grad_norm": 1.1838717460632324,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.6435,
      "mean_token_accuracy": 0.811959981918335,
      "num_tokens": 37685103.0,
      "step": 237
    },
    {
      "epoch": 0.12105798575788403,
      "grad_norm": 1.1500831842422485,
      "learning_rate": 4.016949152542373e-06,
      "loss": 0.6558,
      "mean_token_accuracy": 0.8082505464553833,
      "num_tokens": 37848901.0,
      "step": 238
    },
    {
      "epoch": 0.12156663275686673,
      "grad_norm": 1.1872674226760864,
      "learning_rate": 4.033898305084746e-06,
      "loss": 0.6512,
      "mean_token_accuracy": 0.8086193799972534,
      "num_tokens": 38014704.0,
      "step": 239
    },
    {
      "epoch": 0.12207527975584945,
      "grad_norm": 1.246338129043579,
      "learning_rate": 4.0508474576271186e-06,
      "loss": 0.6506,
      "mean_token_accuracy": 0.8080781698226929,
      "num_tokens": 38166048.0,
      "step": 240
    },
    {
      "epoch": 0.12258392675483215,
      "grad_norm": 1.1303752660751343,
      "learning_rate": 4.067796610169492e-06,
      "loss": 0.5975,
      "mean_token_accuracy": 0.8239215016365051,
      "num_tokens": 38326365.0,
      "step": 241
    },
    {
      "epoch": 0.12309257375381485,
      "grad_norm": 1.2110315561294556,
      "learning_rate": 4.084745762711865e-06,
      "loss": 0.6211,
      "mean_token_accuracy": 0.8160551190376282,
      "num_tokens": 38487350.0,
      "step": 242
    },
    {
      "epoch": 0.12360122075279756,
      "grad_norm": 1.161855697631836,
      "learning_rate": 4.101694915254237e-06,
      "loss": 0.6647,
      "mean_token_accuracy": 0.807031512260437,
      "num_tokens": 38653367.0,
      "step": 243
    },
    {
      "epoch": 0.12410986775178026,
      "grad_norm": 1.153646469116211,
      "learning_rate": 4.1186440677966105e-06,
      "loss": 0.6446,
      "mean_token_accuracy": 0.8096780776977539,
      "num_tokens": 38803903.0,
      "step": 244
    },
    {
      "epoch": 0.12461851475076297,
      "grad_norm": 1.1783907413482666,
      "learning_rate": 4.135593220338983e-06,
      "loss": 0.6259,
      "mean_token_accuracy": 0.8163446187973022,
      "num_tokens": 38956439.0,
      "step": 245
    },
    {
      "epoch": 0.12512716174974567,
      "grad_norm": 1.1585876941680908,
      "learning_rate": 4.152542372881356e-06,
      "loss": 0.6092,
      "mean_token_accuracy": 0.819117546081543,
      "num_tokens": 39109262.0,
      "step": 246
    },
    {
      "epoch": 0.12563580874872837,
      "grad_norm": 1.1186983585357666,
      "learning_rate": 4.169491525423729e-06,
      "loss": 0.6135,
      "mean_token_accuracy": 0.8192994594573975,
      "num_tokens": 39271741.0,
      "step": 247
    },
    {
      "epoch": 0.12614445574771108,
      "grad_norm": 1.135287880897522,
      "learning_rate": 4.186440677966102e-06,
      "loss": 0.6397,
      "mean_token_accuracy": 0.8109389543533325,
      "num_tokens": 39438988.0,
      "step": 248
    },
    {
      "epoch": 0.1266531027466938,
      "grad_norm": 1.198437213897705,
      "learning_rate": 4.203389830508475e-06,
      "loss": 0.6526,
      "mean_token_accuracy": 0.809927225112915,
      "num_tokens": 39606007.0,
      "step": 249
    },
    {
      "epoch": 0.1271617497456765,
      "grad_norm": 1.1955232620239258,
      "learning_rate": 4.220338983050848e-06,
      "loss": 0.6858,
      "mean_token_accuracy": 0.7995947003364563,
      "num_tokens": 39780179.0,
      "step": 250
    },
    {
      "epoch": 0.12767039674465921,
      "grad_norm": 1.141065001487732,
      "learning_rate": 4.23728813559322e-06,
      "loss": 0.6405,
      "mean_token_accuracy": 0.8117692470550537,
      "num_tokens": 39939761.0,
      "step": 251
    },
    {
      "epoch": 0.12817904374364192,
      "grad_norm": 1.222945213317871,
      "learning_rate": 4.2542372881355935e-06,
      "loss": 0.6651,
      "mean_token_accuracy": 0.8068735003471375,
      "num_tokens": 40103824.0,
      "step": 252
    },
    {
      "epoch": 0.12868769074262462,
      "grad_norm": 1.299753189086914,
      "learning_rate": 4.271186440677967e-06,
      "loss": 0.6438,
      "mean_token_accuracy": 0.8107660412788391,
      "num_tokens": 40275322.0,
      "step": 253
    },
    {
      "epoch": 0.12919633774160733,
      "grad_norm": 1.1835240125656128,
      "learning_rate": 4.28813559322034e-06,
      "loss": 0.6179,
      "mean_token_accuracy": 0.8164398074150085,
      "num_tokens": 40433373.0,
      "step": 254
    },
    {
      "epoch": 0.12970498474059003,
      "grad_norm": 1.172080636024475,
      "learning_rate": 4.305084745762712e-06,
      "loss": 0.6467,
      "mean_token_accuracy": 0.809874415397644,
      "num_tokens": 40580246.0,
      "step": 255
    },
    {
      "epoch": 0.13021363173957273,
      "grad_norm": 1.3525550365447998,
      "learning_rate": 4.322033898305085e-06,
      "loss": 0.6176,
      "mean_token_accuracy": 0.8184188008308411,
      "num_tokens": 40743349.0,
      "step": 256
    },
    {
      "epoch": 0.13072227873855544,
      "grad_norm": 1.152124285697937,
      "learning_rate": 4.338983050847458e-06,
      "loss": 0.6069,
      "mean_token_accuracy": 0.8179450035095215,
      "num_tokens": 40909931.0,
      "step": 257
    },
    {
      "epoch": 0.13123092573753814,
      "grad_norm": 1.133946418762207,
      "learning_rate": 4.355932203389831e-06,
      "loss": 0.616,
      "mean_token_accuracy": 0.8172526359558105,
      "num_tokens": 41064135.0,
      "step": 258
    },
    {
      "epoch": 0.13173957273652084,
      "grad_norm": 1.3298335075378418,
      "learning_rate": 4.372881355932203e-06,
      "loss": 0.6107,
      "mean_token_accuracy": 0.8188907504081726,
      "num_tokens": 41215121.0,
      "step": 259
    },
    {
      "epoch": 0.13224821973550355,
      "grad_norm": 1.2448811531066895,
      "learning_rate": 4.3898305084745765e-06,
      "loss": 0.6229,
      "mean_token_accuracy": 0.817119836807251,
      "num_tokens": 41373300.0,
      "step": 260
    },
    {
      "epoch": 0.13275686673448628,
      "grad_norm": 1.1549959182739258,
      "learning_rate": 4.40677966101695e-06,
      "loss": 0.6281,
      "mean_token_accuracy": 0.8145886659622192,
      "num_tokens": 41538822.0,
      "step": 261
    },
    {
      "epoch": 0.13326551373346898,
      "grad_norm": 1.3290430307388306,
      "learning_rate": 4.423728813559323e-06,
      "loss": 0.6715,
      "mean_token_accuracy": 0.8023037910461426,
      "num_tokens": 41699217.0,
      "step": 262
    },
    {
      "epoch": 0.13377416073245169,
      "grad_norm": 1.1931039094924927,
      "learning_rate": 4.440677966101695e-06,
      "loss": 0.6279,
      "mean_token_accuracy": 0.8135525584220886,
      "num_tokens": 41870291.0,
      "step": 263
    },
    {
      "epoch": 0.1342828077314344,
      "grad_norm": 1.2231743335723877,
      "learning_rate": 4.457627118644068e-06,
      "loss": 0.6595,
      "mean_token_accuracy": 0.8069267272949219,
      "num_tokens": 42029303.0,
      "step": 264
    },
    {
      "epoch": 0.1347914547304171,
      "grad_norm": 1.1536788940429688,
      "learning_rate": 4.474576271186441e-06,
      "loss": 0.6467,
      "mean_token_accuracy": 0.8107583522796631,
      "num_tokens": 42197688.0,
      "step": 265
    },
    {
      "epoch": 0.1353001017293998,
      "grad_norm": 1.23910391330719,
      "learning_rate": 4.491525423728814e-06,
      "loss": 0.6301,
      "mean_token_accuracy": 0.8137621879577637,
      "num_tokens": 42349174.0,
      "step": 266
    },
    {
      "epoch": 0.1358087487283825,
      "grad_norm": 1.164609432220459,
      "learning_rate": 4.508474576271187e-06,
      "loss": 0.6095,
      "mean_token_accuracy": 0.8188919425010681,
      "num_tokens": 42506878.0,
      "step": 267
    },
    {
      "epoch": 0.1363173957273652,
      "grad_norm": 1.158076286315918,
      "learning_rate": 4.52542372881356e-06,
      "loss": 0.6489,
      "mean_token_accuracy": 0.8091750741004944,
      "num_tokens": 42676569.0,
      "step": 268
    },
    {
      "epoch": 0.1368260427263479,
      "grad_norm": 1.172566533088684,
      "learning_rate": 4.542372881355933e-06,
      "loss": 0.639,
      "mean_token_accuracy": 0.8108097910881042,
      "num_tokens": 42831149.0,
      "step": 269
    },
    {
      "epoch": 0.1373346897253306,
      "grad_norm": 1.0729190111160278,
      "learning_rate": 4.559322033898305e-06,
      "loss": 0.6385,
      "mean_token_accuracy": 0.8113300800323486,
      "num_tokens": 42992073.0,
      "step": 270
    },
    {
      "epoch": 0.13784333672431331,
      "grad_norm": 1.0952930450439453,
      "learning_rate": 4.576271186440678e-06,
      "loss": 0.6504,
      "mean_token_accuracy": 0.8107742071151733,
      "num_tokens": 43157608.0,
      "step": 271
    },
    {
      "epoch": 0.13835198372329605,
      "grad_norm": 1.1373825073242188,
      "learning_rate": 4.5932203389830506e-06,
      "loss": 0.6053,
      "mean_token_accuracy": 0.8202582597732544,
      "num_tokens": 43319905.0,
      "step": 272
    },
    {
      "epoch": 0.13886063072227875,
      "grad_norm": 1.1396281719207764,
      "learning_rate": 4.610169491525424e-06,
      "loss": 0.6442,
      "mean_token_accuracy": 0.81023108959198,
      "num_tokens": 43483878.0,
      "step": 273
    },
    {
      "epoch": 0.13936927772126145,
      "grad_norm": 1.0831595659255981,
      "learning_rate": 4.627118644067797e-06,
      "loss": 0.6014,
      "mean_token_accuracy": 0.8223749995231628,
      "num_tokens": 43642021.0,
      "step": 274
    },
    {
      "epoch": 0.13987792472024416,
      "grad_norm": 1.2875261306762695,
      "learning_rate": 4.64406779661017e-06,
      "loss": 0.6157,
      "mean_token_accuracy": 0.8163620233535767,
      "num_tokens": 43788563.0,
      "step": 275
    },
    {
      "epoch": 0.14038657171922686,
      "grad_norm": 1.1488323211669922,
      "learning_rate": 4.6610169491525425e-06,
      "loss": 0.6473,
      "mean_token_accuracy": 0.8100237846374512,
      "num_tokens": 43948725.0,
      "step": 276
    },
    {
      "epoch": 0.14089521871820956,
      "grad_norm": 1.088779091835022,
      "learning_rate": 4.677966101694916e-06,
      "loss": 0.6068,
      "mean_token_accuracy": 0.819377064704895,
      "num_tokens": 44107222.0,
      "step": 277
    },
    {
      "epoch": 0.14140386571719227,
      "grad_norm": 1.162422776222229,
      "learning_rate": 4.694915254237288e-06,
      "loss": 0.6296,
      "mean_token_accuracy": 0.8144094944000244,
      "num_tokens": 44265793.0,
      "step": 278
    },
    {
      "epoch": 0.14191251271617497,
      "grad_norm": 1.285162091255188,
      "learning_rate": 4.711864406779661e-06,
      "loss": 0.6396,
      "mean_token_accuracy": 0.8107025027275085,
      "num_tokens": 44416663.0,
      "step": 279
    },
    {
      "epoch": 0.14242115971515767,
      "grad_norm": 1.2526861429214478,
      "learning_rate": 4.728813559322034e-06,
      "loss": 0.629,
      "mean_token_accuracy": 0.8151564598083496,
      "num_tokens": 44585677.0,
      "step": 280
    },
    {
      "epoch": 0.14292980671414038,
      "grad_norm": 1.271353006362915,
      "learning_rate": 4.745762711864408e-06,
      "loss": 0.6325,
      "mean_token_accuracy": 0.8143658638000488,
      "num_tokens": 44733031.0,
      "step": 281
    },
    {
      "epoch": 0.14343845371312308,
      "grad_norm": 1.6910256147384644,
      "learning_rate": 4.76271186440678e-06,
      "loss": 0.6524,
      "mean_token_accuracy": 0.8086466789245605,
      "num_tokens": 44896422.0,
      "step": 282
    },
    {
      "epoch": 0.14394710071210579,
      "grad_norm": 1.168068766593933,
      "learning_rate": 4.779661016949153e-06,
      "loss": 0.6153,
      "mean_token_accuracy": 0.8179129958152771,
      "num_tokens": 45063428.0,
      "step": 283
    },
    {
      "epoch": 0.14445574771108852,
      "grad_norm": 1.2636305093765259,
      "learning_rate": 4.7966101694915255e-06,
      "loss": 0.6059,
      "mean_token_accuracy": 0.8190077543258667,
      "num_tokens": 45222963.0,
      "step": 284
    },
    {
      "epoch": 0.14496439471007122,
      "grad_norm": 1.121416687965393,
      "learning_rate": 4.813559322033899e-06,
      "loss": 0.6207,
      "mean_token_accuracy": 0.8159518241882324,
      "num_tokens": 45369420.0,
      "step": 285
    },
    {
      "epoch": 0.14547304170905392,
      "grad_norm": 1.2177066802978516,
      "learning_rate": 4.830508474576272e-06,
      "loss": 0.589,
      "mean_token_accuracy": 0.8277342319488525,
      "num_tokens": 45525565.0,
      "step": 286
    },
    {
      "epoch": 0.14598168870803663,
      "grad_norm": 1.2821828126907349,
      "learning_rate": 4.847457627118645e-06,
      "loss": 0.6338,
      "mean_token_accuracy": 0.8130690455436707,
      "num_tokens": 45688704.0,
      "step": 287
    },
    {
      "epoch": 0.14649033570701933,
      "grad_norm": 1.1307625770568848,
      "learning_rate": 4.864406779661017e-06,
      "loss": 0.6274,
      "mean_token_accuracy": 0.813611626625061,
      "num_tokens": 45852191.0,
      "step": 288
    },
    {
      "epoch": 0.14699898270600203,
      "grad_norm": 1.269042730331421,
      "learning_rate": 4.881355932203391e-06,
      "loss": 0.6412,
      "mean_token_accuracy": 0.8145066499710083,
      "num_tokens": 46009944.0,
      "step": 289
    },
    {
      "epoch": 0.14750762970498474,
      "grad_norm": 1.25006103515625,
      "learning_rate": 4.898305084745763e-06,
      "loss": 0.6723,
      "mean_token_accuracy": 0.8050320148468018,
      "num_tokens": 46163464.0,
      "step": 290
    },
    {
      "epoch": 0.14801627670396744,
      "grad_norm": 1.1930915117263794,
      "learning_rate": 4.915254237288136e-06,
      "loss": 0.6295,
      "mean_token_accuracy": 0.8156224489212036,
      "num_tokens": 46315469.0,
      "step": 291
    },
    {
      "epoch": 0.14852492370295015,
      "grad_norm": 1.1504136323928833,
      "learning_rate": 4.9322033898305085e-06,
      "loss": 0.6161,
      "mean_token_accuracy": 0.8182114958763123,
      "num_tokens": 46475098.0,
      "step": 292
    },
    {
      "epoch": 0.14903357070193285,
      "grad_norm": 1.1462980508804321,
      "learning_rate": 4.949152542372882e-06,
      "loss": 0.6266,
      "mean_token_accuracy": 0.8146628141403198,
      "num_tokens": 46645510.0,
      "step": 293
    },
    {
      "epoch": 0.14954221770091555,
      "grad_norm": 1.1933327913284302,
      "learning_rate": 4.966101694915255e-06,
      "loss": 0.612,
      "mean_token_accuracy": 0.8185099959373474,
      "num_tokens": 46799760.0,
      "step": 294
    },
    {
      "epoch": 0.15005086469989828,
      "grad_norm": 1.2216078042984009,
      "learning_rate": 4.983050847457628e-06,
      "loss": 0.615,
      "mean_token_accuracy": 0.8170180320739746,
      "num_tokens": 46965987.0,
      "step": 295
    },
    {
      "epoch": 0.150559511698881,
      "grad_norm": 1.2336102724075317,
      "learning_rate": 5e-06,
      "loss": 0.5949,
      "mean_token_accuracy": 0.8221638202667236,
      "num_tokens": 47128771.0,
      "step": 296
    },
    {
      "epoch": 0.1510681586978637,
      "grad_norm": 1.1848920583724976,
      "learning_rate": 5.016949152542373e-06,
      "loss": 0.6477,
      "mean_token_accuracy": 0.8114307522773743,
      "num_tokens": 47288316.0,
      "step": 297
    },
    {
      "epoch": 0.1515768056968464,
      "grad_norm": 1.2134908437728882,
      "learning_rate": 5.033898305084746e-06,
      "loss": 0.6137,
      "mean_token_accuracy": 0.8162137269973755,
      "num_tokens": 47455167.0,
      "step": 298
    },
    {
      "epoch": 0.1520854526958291,
      "grad_norm": 1.1854346990585327,
      "learning_rate": 5.050847457627119e-06,
      "loss": 0.6353,
      "mean_token_accuracy": 0.8133347034454346,
      "num_tokens": 47600859.0,
      "step": 299
    },
    {
      "epoch": 0.1525940996948118,
      "grad_norm": 1.2845247983932495,
      "learning_rate": 5.067796610169492e-06,
      "loss": 0.6107,
      "mean_token_accuracy": 0.8182547688484192,
      "num_tokens": 47768491.0,
      "step": 300
    },
    {
      "epoch": 0.1531027466937945,
      "grad_norm": 1.2308179140090942,
      "learning_rate": 5.084745762711865e-06,
      "loss": 0.6199,
      "mean_token_accuracy": 0.8154051303863525,
      "num_tokens": 47918622.0,
      "step": 301
    },
    {
      "epoch": 0.1536113936927772,
      "grad_norm": 1.2153412103652954,
      "learning_rate": 5.101694915254237e-06,
      "loss": 0.5681,
      "mean_token_accuracy": 0.8302782773971558,
      "num_tokens": 48076479.0,
      "step": 302
    },
    {
      "epoch": 0.1541200406917599,
      "grad_norm": 1.2243883609771729,
      "learning_rate": 5.118644067796611e-06,
      "loss": 0.6139,
      "mean_token_accuracy": 0.8183121681213379,
      "num_tokens": 48226249.0,
      "step": 303
    },
    {
      "epoch": 0.15462868769074262,
      "grad_norm": 1.0889983177185059,
      "learning_rate": 5.135593220338983e-06,
      "loss": 0.5715,
      "mean_token_accuracy": 0.8308095335960388,
      "num_tokens": 48390048.0,
      "step": 304
    },
    {
      "epoch": 0.15513733468972532,
      "grad_norm": 1.2139381170272827,
      "learning_rate": 5.152542372881356e-06,
      "loss": 0.5961,
      "mean_token_accuracy": 0.822356104850769,
      "num_tokens": 48558970.0,
      "step": 305
    },
    {
      "epoch": 0.15564598168870802,
      "grad_norm": 1.1127007007598877,
      "learning_rate": 5.169491525423729e-06,
      "loss": 0.6033,
      "mean_token_accuracy": 0.8220300674438477,
      "num_tokens": 48712282.0,
      "step": 306
    },
    {
      "epoch": 0.15615462868769076,
      "grad_norm": 1.232642412185669,
      "learning_rate": 5.186440677966102e-06,
      "loss": 0.6247,
      "mean_token_accuracy": 0.8168089985847473,
      "num_tokens": 48874280.0,
      "step": 307
    },
    {
      "epoch": 0.15666327568667346,
      "grad_norm": 1.2171428203582764,
      "learning_rate": 5.203389830508475e-06,
      "loss": 0.6643,
      "mean_token_accuracy": 0.8060603141784668,
      "num_tokens": 49032570.0,
      "step": 308
    },
    {
      "epoch": 0.15717192268565616,
      "grad_norm": 1.1634498834609985,
      "learning_rate": 5.220338983050848e-06,
      "loss": 0.6029,
      "mean_token_accuracy": 0.8210587501525879,
      "num_tokens": 49206160.0,
      "step": 309
    },
    {
      "epoch": 0.15768056968463887,
      "grad_norm": 1.1851032972335815,
      "learning_rate": 5.23728813559322e-06,
      "loss": 0.6534,
      "mean_token_accuracy": 0.8060614466667175,
      "num_tokens": 49357658.0,
      "step": 310
    },
    {
      "epoch": 0.15818921668362157,
      "grad_norm": 1.223835825920105,
      "learning_rate": 5.254237288135594e-06,
      "loss": 0.6275,
      "mean_token_accuracy": 0.8136088252067566,
      "num_tokens": 49513574.0,
      "step": 311
    },
    {
      "epoch": 0.15869786368260427,
      "grad_norm": 1.1291638612747192,
      "learning_rate": 5.271186440677966e-06,
      "loss": 0.5827,
      "mean_token_accuracy": 0.8268251419067383,
      "num_tokens": 49678022.0,
      "step": 312
    },
    {
      "epoch": 0.15920651068158698,
      "grad_norm": 1.2867546081542969,
      "learning_rate": 5.28813559322034e-06,
      "loss": 0.6355,
      "mean_token_accuracy": 0.8134222030639648,
      "num_tokens": 49844519.0,
      "step": 313
    },
    {
      "epoch": 0.15971515768056968,
      "grad_norm": 1.2789193391799927,
      "learning_rate": 5.305084745762712e-06,
      "loss": 0.5766,
      "mean_token_accuracy": 0.8284808397293091,
      "num_tokens": 50003658.0,
      "step": 314
    },
    {
      "epoch": 0.16022380467955238,
      "grad_norm": 1.2516961097717285,
      "learning_rate": 5.322033898305086e-06,
      "loss": 0.6287,
      "mean_token_accuracy": 0.8147469758987427,
      "num_tokens": 50169430.0,
      "step": 315
    },
    {
      "epoch": 0.1607324516785351,
      "grad_norm": 1.2042748928070068,
      "learning_rate": 5.338983050847458e-06,
      "loss": 0.6007,
      "mean_token_accuracy": 0.8214442729949951,
      "num_tokens": 50337188.0,
      "step": 316
    },
    {
      "epoch": 0.1612410986775178,
      "grad_norm": 1.3558809757232666,
      "learning_rate": 5.355932203389831e-06,
      "loss": 0.6417,
      "mean_token_accuracy": 0.8108523488044739,
      "num_tokens": 50488304.0,
      "step": 317
    },
    {
      "epoch": 0.1617497456765005,
      "grad_norm": 1.2530821561813354,
      "learning_rate": 5.372881355932204e-06,
      "loss": 0.6376,
      "mean_token_accuracy": 0.8133925199508667,
      "num_tokens": 50649690.0,
      "step": 318
    },
    {
      "epoch": 0.16225839267548323,
      "grad_norm": 1.2648253440856934,
      "learning_rate": 5.389830508474577e-06,
      "loss": 0.5934,
      "mean_token_accuracy": 0.8214468955993652,
      "num_tokens": 50809934.0,
      "step": 319
    },
    {
      "epoch": 0.16276703967446593,
      "grad_norm": 1.289933204650879,
      "learning_rate": 5.40677966101695e-06,
      "loss": 0.6457,
      "mean_token_accuracy": 0.8087766170501709,
      "num_tokens": 50973651.0,
      "step": 320
    },
    {
      "epoch": 0.16327568667344863,
      "grad_norm": 1.2105209827423096,
      "learning_rate": 5.423728813559323e-06,
      "loss": 0.6661,
      "mean_token_accuracy": 0.8031639456748962,
      "num_tokens": 51152576.0,
      "step": 321
    },
    {
      "epoch": 0.16378433367243134,
      "grad_norm": 3.8657143115997314,
      "learning_rate": 5.440677966101695e-06,
      "loss": 0.5896,
      "mean_token_accuracy": 0.8247578740119934,
      "num_tokens": 51302446.0,
      "step": 322
    },
    {
      "epoch": 0.16429298067141404,
      "grad_norm": 1.6237268447875977,
      "learning_rate": 5.457627118644067e-06,
      "loss": 0.626,
      "mean_token_accuracy": 0.8159285187721252,
      "num_tokens": 51443754.0,
      "step": 323
    },
    {
      "epoch": 0.16480162767039674,
      "grad_norm": 1.179121971130371,
      "learning_rate": 5.474576271186441e-06,
      "loss": 0.6163,
      "mean_token_accuracy": 0.8178367018699646,
      "num_tokens": 51613127.0,
      "step": 324
    },
    {
      "epoch": 0.16531027466937945,
      "grad_norm": 1.1720213890075684,
      "learning_rate": 5.491525423728814e-06,
      "loss": 0.6054,
      "mean_token_accuracy": 0.8200127482414246,
      "num_tokens": 51778162.0,
      "step": 325
    },
    {
      "epoch": 0.16581892166836215,
      "grad_norm": 1.1967053413391113,
      "learning_rate": 5.508474576271187e-06,
      "loss": 0.602,
      "mean_token_accuracy": 0.8214297890663147,
      "num_tokens": 51935798.0,
      "step": 326
    },
    {
      "epoch": 0.16632756866734486,
      "grad_norm": 1.2503341436386108,
      "learning_rate": 5.525423728813559e-06,
      "loss": 0.6247,
      "mean_token_accuracy": 0.8145067691802979,
      "num_tokens": 52089803.0,
      "step": 327
    },
    {
      "epoch": 0.16683621566632756,
      "grad_norm": 1.2478593587875366,
      "learning_rate": 5.542372881355933e-06,
      "loss": 0.6118,
      "mean_token_accuracy": 0.8154499530792236,
      "num_tokens": 52241386.0,
      "step": 328
    },
    {
      "epoch": 0.16734486266531026,
      "grad_norm": 1.168043613433838,
      "learning_rate": 5.559322033898306e-06,
      "loss": 0.6269,
      "mean_token_accuracy": 0.8142362236976624,
      "num_tokens": 52397569.0,
      "step": 329
    },
    {
      "epoch": 0.167853509664293,
      "grad_norm": 1.290001630783081,
      "learning_rate": 5.576271186440678e-06,
      "loss": 0.6198,
      "mean_token_accuracy": 0.8157522678375244,
      "num_tokens": 52550468.0,
      "step": 330
    },
    {
      "epoch": 0.1683621566632757,
      "grad_norm": 1.2450792789459229,
      "learning_rate": 5.593220338983051e-06,
      "loss": 0.6044,
      "mean_token_accuracy": 0.8214341402053833,
      "num_tokens": 52705147.0,
      "step": 331
    },
    {
      "epoch": 0.1688708036622584,
      "grad_norm": 1.2368236780166626,
      "learning_rate": 5.610169491525424e-06,
      "loss": 0.6489,
      "mean_token_accuracy": 0.8084084987640381,
      "num_tokens": 52870603.0,
      "step": 332
    },
    {
      "epoch": 0.1693794506612411,
      "grad_norm": 1.301377296447754,
      "learning_rate": 5.6271186440677975e-06,
      "loss": 0.6287,
      "mean_token_accuracy": 0.8143856525421143,
      "num_tokens": 53027492.0,
      "step": 333
    },
    {
      "epoch": 0.1698880976602238,
      "grad_norm": 1.2675710916519165,
      "learning_rate": 5.64406779661017e-06,
      "loss": 0.6272,
      "mean_token_accuracy": 0.8145819902420044,
      "num_tokens": 53186596.0,
      "step": 334
    },
    {
      "epoch": 0.1703967446592065,
      "grad_norm": 1.2474238872528076,
      "learning_rate": 5.661016949152542e-06,
      "loss": 0.6111,
      "mean_token_accuracy": 0.8180166482925415,
      "num_tokens": 53341286.0,
      "step": 335
    },
    {
      "epoch": 0.17090539165818922,
      "grad_norm": 1.1779944896697998,
      "learning_rate": 5.677966101694916e-06,
      "loss": 0.6307,
      "mean_token_accuracy": 0.8118512034416199,
      "num_tokens": 53511107.0,
      "step": 336
    },
    {
      "epoch": 0.17141403865717192,
      "grad_norm": 1.151164174079895,
      "learning_rate": 5.694915254237289e-06,
      "loss": 0.5874,
      "mean_token_accuracy": 0.8255366086959839,
      "num_tokens": 53673861.0,
      "step": 337
    },
    {
      "epoch": 0.17192268565615462,
      "grad_norm": 1.3543330430984497,
      "learning_rate": 5.711864406779661e-06,
      "loss": 0.6194,
      "mean_token_accuracy": 0.8178123831748962,
      "num_tokens": 53829950.0,
      "step": 338
    },
    {
      "epoch": 0.17243133265513733,
      "grad_norm": 1.2115367650985718,
      "learning_rate": 5.728813559322034e-06,
      "loss": 0.6358,
      "mean_token_accuracy": 0.8141074180603027,
      "num_tokens": 53992561.0,
      "step": 339
    },
    {
      "epoch": 0.17293997965412003,
      "grad_norm": 1.124480128288269,
      "learning_rate": 5.745762711864407e-06,
      "loss": 0.6296,
      "mean_token_accuracy": 0.8138951063156128,
      "num_tokens": 54156414.0,
      "step": 340
    },
    {
      "epoch": 0.17344862665310273,
      "grad_norm": 1.212111234664917,
      "learning_rate": 5.7627118644067805e-06,
      "loss": 0.6212,
      "mean_token_accuracy": 0.8161656856536865,
      "num_tokens": 54315479.0,
      "step": 341
    },
    {
      "epoch": 0.17395727365208546,
      "grad_norm": 1.1296113729476929,
      "learning_rate": 5.779661016949153e-06,
      "loss": 0.6132,
      "mean_token_accuracy": 0.8164753913879395,
      "num_tokens": 54487854.0,
      "step": 342
    },
    {
      "epoch": 0.17446592065106817,
      "grad_norm": 1.1760289669036865,
      "learning_rate": 5.796610169491525e-06,
      "loss": 0.6085,
      "mean_token_accuracy": 0.8185662031173706,
      "num_tokens": 54637940.0,
      "step": 343
    },
    {
      "epoch": 0.17497456765005087,
      "grad_norm": 1.222051978111267,
      "learning_rate": 5.813559322033899e-06,
      "loss": 0.6142,
      "mean_token_accuracy": 0.8181614279747009,
      "num_tokens": 54799217.0,
      "step": 344
    },
    {
      "epoch": 0.17548321464903358,
      "grad_norm": 1.1629087924957275,
      "learning_rate": 5.830508474576272e-06,
      "loss": 0.601,
      "mean_token_accuracy": 0.8221548795700073,
      "num_tokens": 54952384.0,
      "step": 345
    },
    {
      "epoch": 0.17599186164801628,
      "grad_norm": 1.1986356973648071,
      "learning_rate": 5.847457627118645e-06,
      "loss": 0.622,
      "mean_token_accuracy": 0.8151795864105225,
      "num_tokens": 55126417.0,
      "step": 346
    },
    {
      "epoch": 0.17650050864699898,
      "grad_norm": 1.1795095205307007,
      "learning_rate": 5.864406779661017e-06,
      "loss": 0.6564,
      "mean_token_accuracy": 0.8053359985351562,
      "num_tokens": 55296549.0,
      "step": 347
    },
    {
      "epoch": 0.1770091556459817,
      "grad_norm": 1.2172720432281494,
      "learning_rate": 5.881355932203391e-06,
      "loss": 0.5921,
      "mean_token_accuracy": 0.8246195316314697,
      "num_tokens": 55457557.0,
      "step": 348
    },
    {
      "epoch": 0.1775178026449644,
      "grad_norm": 1.1696581840515137,
      "learning_rate": 5.8983050847457635e-06,
      "loss": 0.6151,
      "mean_token_accuracy": 0.8157768249511719,
      "num_tokens": 55615322.0,
      "step": 349
    },
    {
      "epoch": 0.1780264496439471,
      "grad_norm": 1.2739365100860596,
      "learning_rate": 5.915254237288136e-06,
      "loss": 0.6278,
      "mean_token_accuracy": 0.8150952458381653,
      "num_tokens": 55780975.0,
      "step": 350
    },
    {
      "epoch": 0.1785350966429298,
      "grad_norm": 1.2735313177108765,
      "learning_rate": 5.932203389830509e-06,
      "loss": 0.6269,
      "mean_token_accuracy": 0.8135517835617065,
      "num_tokens": 55931561.0,
      "step": 351
    },
    {
      "epoch": 0.1790437436419125,
      "grad_norm": 1.2213759422302246,
      "learning_rate": 5.949152542372881e-06,
      "loss": 0.6508,
      "mean_token_accuracy": 0.8071233034133911,
      "num_tokens": 56097949.0,
      "step": 352
    },
    {
      "epoch": 0.17955239064089523,
      "grad_norm": 1.164584994316101,
      "learning_rate": 5.9661016949152555e-06,
      "loss": 0.6596,
      "mean_token_accuracy": 0.8046541213989258,
      "num_tokens": 56256104.0,
      "step": 353
    },
    {
      "epoch": 0.18006103763987794,
      "grad_norm": 1.2391185760498047,
      "learning_rate": 5.983050847457628e-06,
      "loss": 0.6413,
      "mean_token_accuracy": 0.8112727403640747,
      "num_tokens": 56427795.0,
      "step": 354
    },
    {
      "epoch": 0.18056968463886064,
      "grad_norm": 1.2065702676773071,
      "learning_rate": 6e-06,
      "loss": 0.6143,
      "mean_token_accuracy": 0.8167493343353271,
      "num_tokens": 56580755.0,
      "step": 355
    },
    {
      "epoch": 0.18107833163784334,
      "grad_norm": 1.212113380432129,
      "learning_rate": 6.0169491525423725e-06,
      "loss": 0.6219,
      "mean_token_accuracy": 0.8156697750091553,
      "num_tokens": 56725887.0,
      "step": 356
    },
    {
      "epoch": 0.18158697863682605,
      "grad_norm": 1.1100784540176392,
      "learning_rate": 6.0338983050847465e-06,
      "loss": 0.6086,
      "mean_token_accuracy": 0.8173103332519531,
      "num_tokens": 56893128.0,
      "step": 357
    },
    {
      "epoch": 0.18209562563580875,
      "grad_norm": 1.0563573837280273,
      "learning_rate": 6.050847457627119e-06,
      "loss": 0.6011,
      "mean_token_accuracy": 0.821357250213623,
      "num_tokens": 57066741.0,
      "step": 358
    },
    {
      "epoch": 0.18260427263479145,
      "grad_norm": 1.2157350778579712,
      "learning_rate": 6.067796610169492e-06,
      "loss": 0.5994,
      "mean_token_accuracy": 0.8220726847648621,
      "num_tokens": 57215386.0,
      "step": 359
    },
    {
      "epoch": 0.18311291963377416,
      "grad_norm": 1.2231732606887817,
      "learning_rate": 6.084745762711864e-06,
      "loss": 0.6222,
      "mean_token_accuracy": 0.8157521486282349,
      "num_tokens": 57382943.0,
      "step": 360
    },
    {
      "epoch": 0.18362156663275686,
      "grad_norm": 1.1992939710617065,
      "learning_rate": 6.1016949152542385e-06,
      "loss": 0.6004,
      "mean_token_accuracy": 0.8207837343215942,
      "num_tokens": 57533799.0,
      "step": 361
    },
    {
      "epoch": 0.18413021363173956,
      "grad_norm": 1.2414939403533936,
      "learning_rate": 6.118644067796611e-06,
      "loss": 0.6133,
      "mean_token_accuracy": 0.8167545199394226,
      "num_tokens": 57697387.0,
      "step": 362
    },
    {
      "epoch": 0.18463886063072227,
      "grad_norm": 1.1865259408950806,
      "learning_rate": 6.135593220338983e-06,
      "loss": 0.6199,
      "mean_token_accuracy": 0.8168951869010925,
      "num_tokens": 57855776.0,
      "step": 363
    },
    {
      "epoch": 0.18514750762970497,
      "grad_norm": 1.1268796920776367,
      "learning_rate": 6.152542372881356e-06,
      "loss": 0.6,
      "mean_token_accuracy": 0.8215888738632202,
      "num_tokens": 58024333.0,
      "step": 364
    },
    {
      "epoch": 0.1856561546286877,
      "grad_norm": 1.254247784614563,
      "learning_rate": 6.1694915254237295e-06,
      "loss": 0.6002,
      "mean_token_accuracy": 0.8212365508079529,
      "num_tokens": 58184182.0,
      "step": 365
    },
    {
      "epoch": 0.1861648016276704,
      "grad_norm": 1.1949880123138428,
      "learning_rate": 6.186440677966103e-06,
      "loss": 0.6216,
      "mean_token_accuracy": 0.8140779733657837,
      "num_tokens": 58338761.0,
      "step": 366
    },
    {
      "epoch": 0.1866734486266531,
      "grad_norm": 1.2143034934997559,
      "learning_rate": 6.203389830508475e-06,
      "loss": 0.621,
      "mean_token_accuracy": 0.8147321939468384,
      "num_tokens": 58485678.0,
      "step": 367
    },
    {
      "epoch": 0.1871820956256358,
      "grad_norm": 1.2522412538528442,
      "learning_rate": 6.2203389830508474e-06,
      "loss": 0.6444,
      "mean_token_accuracy": 0.8117758631706238,
      "num_tokens": 58643138.0,
      "step": 368
    },
    {
      "epoch": 0.18769074262461852,
      "grad_norm": 1.1688337326049805,
      "learning_rate": 6.2372881355932215e-06,
      "loss": 0.5978,
      "mean_token_accuracy": 0.821524441242218,
      "num_tokens": 58804789.0,
      "step": 369
    },
    {
      "epoch": 0.18819938962360122,
      "grad_norm": 1.1917411088943481,
      "learning_rate": 6.254237288135594e-06,
      "loss": 0.6014,
      "mean_token_accuracy": 0.8203130960464478,
      "num_tokens": 58978624.0,
      "step": 370
    },
    {
      "epoch": 0.18870803662258392,
      "grad_norm": 1.1307294368743896,
      "learning_rate": 6.271186440677966e-06,
      "loss": 0.5915,
      "mean_token_accuracy": 0.8264414668083191,
      "num_tokens": 59150156.0,
      "step": 371
    },
    {
      "epoch": 0.18921668362156663,
      "grad_norm": 1.1567331552505493,
      "learning_rate": 6.288135593220339e-06,
      "loss": 0.5645,
      "mean_token_accuracy": 0.8300540447235107,
      "num_tokens": 59307831.0,
      "step": 372
    },
    {
      "epoch": 0.18972533062054933,
      "grad_norm": 1.2923393249511719,
      "learning_rate": 6.3050847457627125e-06,
      "loss": 0.6242,
      "mean_token_accuracy": 0.8148703575134277,
      "num_tokens": 59458464.0,
      "step": 373
    },
    {
      "epoch": 0.19023397761953204,
      "grad_norm": 1.2552546262741089,
      "learning_rate": 6.322033898305086e-06,
      "loss": 0.6044,
      "mean_token_accuracy": 0.821365237236023,
      "num_tokens": 59612951.0,
      "step": 374
    },
    {
      "epoch": 0.19074262461851474,
      "grad_norm": 1.2645580768585205,
      "learning_rate": 6.338983050847458e-06,
      "loss": 0.5737,
      "mean_token_accuracy": 0.8263217806816101,
      "num_tokens": 59774372.0,
      "step": 375
    },
    {
      "epoch": 0.19125127161749747,
      "grad_norm": 1.2015972137451172,
      "learning_rate": 6.3559322033898304e-06,
      "loss": 0.5775,
      "mean_token_accuracy": 0.8266362547874451,
      "num_tokens": 59940444.0,
      "step": 376
    },
    {
      "epoch": 0.19175991861648017,
      "grad_norm": 1.3558928966522217,
      "learning_rate": 6.372881355932204e-06,
      "loss": 0.6211,
      "mean_token_accuracy": 0.8140040636062622,
      "num_tokens": 60097403.0,
      "step": 377
    },
    {
      "epoch": 0.19226856561546288,
      "grad_norm": 1.1792396306991577,
      "learning_rate": 6.389830508474577e-06,
      "loss": 0.5988,
      "mean_token_accuracy": 0.820154070854187,
      "num_tokens": 60270518.0,
      "step": 378
    },
    {
      "epoch": 0.19277721261444558,
      "grad_norm": 1.29557204246521,
      "learning_rate": 6.40677966101695e-06,
      "loss": 0.5914,
      "mean_token_accuracy": 0.8230335712432861,
      "num_tokens": 60434705.0,
      "step": 379
    },
    {
      "epoch": 0.19328585961342828,
      "grad_norm": 1.251295804977417,
      "learning_rate": 6.423728813559322e-06,
      "loss": 0.6274,
      "mean_token_accuracy": 0.814145028591156,
      "num_tokens": 60587162.0,
      "step": 380
    },
    {
      "epoch": 0.193794506612411,
      "grad_norm": 1.1563841104507446,
      "learning_rate": 6.440677966101695e-06,
      "loss": 0.5998,
      "mean_token_accuracy": 0.8219859600067139,
      "num_tokens": 60749142.0,
      "step": 381
    },
    {
      "epoch": 0.1943031536113937,
      "grad_norm": 1.303562879562378,
      "learning_rate": 6.457627118644069e-06,
      "loss": 0.5886,
      "mean_token_accuracy": 0.823674201965332,
      "num_tokens": 60917294.0,
      "step": 382
    },
    {
      "epoch": 0.1948118006103764,
      "grad_norm": 1.3838107585906982,
      "learning_rate": 6.474576271186441e-06,
      "loss": 0.5929,
      "mean_token_accuracy": 0.8225725889205933,
      "num_tokens": 61084543.0,
      "step": 383
    },
    {
      "epoch": 0.1953204476093591,
      "grad_norm": 1.2146414518356323,
      "learning_rate": 6.491525423728814e-06,
      "loss": 0.6255,
      "mean_token_accuracy": 0.8173696398735046,
      "num_tokens": 61248579.0,
      "step": 384
    },
    {
      "epoch": 0.1958290946083418,
      "grad_norm": 1.264757752418518,
      "learning_rate": 6.508474576271187e-06,
      "loss": 0.6135,
      "mean_token_accuracy": 0.8174546360969543,
      "num_tokens": 61403557.0,
      "step": 385
    },
    {
      "epoch": 0.1963377416073245,
      "grad_norm": 1.2781434059143066,
      "learning_rate": 6.52542372881356e-06,
      "loss": 0.6065,
      "mean_token_accuracy": 0.8191805481910706,
      "num_tokens": 61568050.0,
      "step": 386
    },
    {
      "epoch": 0.1968463886063072,
      "grad_norm": 1.1514405012130737,
      "learning_rate": 6.542372881355933e-06,
      "loss": 0.6281,
      "mean_token_accuracy": 0.8139955997467041,
      "num_tokens": 61735941.0,
      "step": 387
    },
    {
      "epoch": 0.19735503560528994,
      "grad_norm": 1.1844816207885742,
      "learning_rate": 6.559322033898305e-06,
      "loss": 0.5633,
      "mean_token_accuracy": 0.8317704200744629,
      "num_tokens": 61895500.0,
      "step": 388
    },
    {
      "epoch": 0.19786368260427264,
      "grad_norm": 1.1167271137237549,
      "learning_rate": 6.576271186440678e-06,
      "loss": 0.6092,
      "mean_token_accuracy": 0.8186742067337036,
      "num_tokens": 62062619.0,
      "step": 389
    },
    {
      "epoch": 0.19837232960325535,
      "grad_norm": 1.2830709218978882,
      "learning_rate": 6.593220338983052e-06,
      "loss": 0.6137,
      "mean_token_accuracy": 0.8206446766853333,
      "num_tokens": 62247115.0,
      "step": 390
    },
    {
      "epoch": 0.19888097660223805,
      "grad_norm": 1.105200171470642,
      "learning_rate": 6.610169491525424e-06,
      "loss": 0.6043,
      "mean_token_accuracy": 0.8199334740638733,
      "num_tokens": 62404051.0,
      "step": 391
    },
    {
      "epoch": 0.19938962360122076,
      "grad_norm": 4.479599952697754,
      "learning_rate": 6.627118644067797e-06,
      "loss": 0.6184,
      "mean_token_accuracy": 0.8156611919403076,
      "num_tokens": 62559441.0,
      "step": 392
    },
    {
      "epoch": 0.19989827060020346,
      "grad_norm": 1.32205069065094,
      "learning_rate": 6.64406779661017e-06,
      "loss": 0.6477,
      "mean_token_accuracy": 0.8093985915184021,
      "num_tokens": 62719791.0,
      "step": 393
    },
    {
      "epoch": 0.20040691759918616,
      "grad_norm": 1.2664779424667358,
      "learning_rate": 6.661016949152544e-06,
      "loss": 0.6088,
      "mean_token_accuracy": 0.818210780620575,
      "num_tokens": 62863338.0,
      "step": 394
    },
    {
      "epoch": 0.20091556459816887,
      "grad_norm": 1.157365083694458,
      "learning_rate": 6.677966101694916e-06,
      "loss": 0.5954,
      "mean_token_accuracy": 0.8220131397247314,
      "num_tokens": 63024566.0,
      "step": 395
    },
    {
      "epoch": 0.20142421159715157,
      "grad_norm": 1.1239126920700073,
      "learning_rate": 6.694915254237288e-06,
      "loss": 0.5648,
      "mean_token_accuracy": 0.8296006917953491,
      "num_tokens": 63189778.0,
      "step": 396
    },
    {
      "epoch": 0.20193285859613427,
      "grad_norm": 1.2238695621490479,
      "learning_rate": 6.7118644067796615e-06,
      "loss": 0.6229,
      "mean_token_accuracy": 0.8139328956604004,
      "num_tokens": 63347541.0,
      "step": 397
    },
    {
      "epoch": 0.20244150559511698,
      "grad_norm": 1.1842625141143799,
      "learning_rate": 6.728813559322035e-06,
      "loss": 0.6311,
      "mean_token_accuracy": 0.8118071556091309,
      "num_tokens": 63513614.0,
      "step": 398
    },
    {
      "epoch": 0.2029501525940997,
      "grad_norm": 1.3057259321212769,
      "learning_rate": 6.745762711864408e-06,
      "loss": 0.6071,
      "mean_token_accuracy": 0.8176935315132141,
      "num_tokens": 63657699.0,
      "step": 399
    },
    {
      "epoch": 0.2034587995930824,
      "grad_norm": 1.132493495941162,
      "learning_rate": 6.76271186440678e-06,
      "loss": 0.6029,
      "mean_token_accuracy": 0.8205714821815491,
      "num_tokens": 63810862.0,
      "step": 400
    },
    {
      "epoch": 0.20396744659206512,
      "grad_norm": 1.1689492464065552,
      "learning_rate": 6.779661016949153e-06,
      "loss": 0.6263,
      "mean_token_accuracy": 0.8131099939346313,
      "num_tokens": 63976337.0,
      "step": 401
    },
    {
      "epoch": 0.20447609359104782,
      "grad_norm": 1.246856927871704,
      "learning_rate": 6.796610169491527e-06,
      "loss": 0.6094,
      "mean_token_accuracy": 0.8199894428253174,
      "num_tokens": 64131809.0,
      "step": 402
    },
    {
      "epoch": 0.20498474059003052,
      "grad_norm": 1.1247804164886475,
      "learning_rate": 6.813559322033899e-06,
      "loss": 0.6244,
      "mean_token_accuracy": 0.8156639337539673,
      "num_tokens": 64301531.0,
      "step": 403
    },
    {
      "epoch": 0.20549338758901323,
      "grad_norm": 1.205808401107788,
      "learning_rate": 6.830508474576271e-06,
      "loss": 0.6042,
      "mean_token_accuracy": 0.8193472623825073,
      "num_tokens": 64453734.0,
      "step": 404
    },
    {
      "epoch": 0.20600203458799593,
      "grad_norm": 1.227185845375061,
      "learning_rate": 6.8474576271186445e-06,
      "loss": 0.6337,
      "mean_token_accuracy": 0.814788818359375,
      "num_tokens": 64614638.0,
      "step": 405
    },
    {
      "epoch": 0.20651068158697863,
      "grad_norm": 1.187564730644226,
      "learning_rate": 6.864406779661017e-06,
      "loss": 0.5787,
      "mean_token_accuracy": 0.8256047368049622,
      "num_tokens": 64783877.0,
      "step": 406
    },
    {
      "epoch": 0.20701932858596134,
      "grad_norm": 1.1679338216781616,
      "learning_rate": 6.881355932203391e-06,
      "loss": 0.6134,
      "mean_token_accuracy": 0.8183737993240356,
      "num_tokens": 64946232.0,
      "step": 407
    },
    {
      "epoch": 0.20752797558494404,
      "grad_norm": 1.1042274236679077,
      "learning_rate": 6.898305084745763e-06,
      "loss": 0.5988,
      "mean_token_accuracy": 0.821270227432251,
      "num_tokens": 65105547.0,
      "step": 408
    },
    {
      "epoch": 0.20803662258392674,
      "grad_norm": 1.2437533140182495,
      "learning_rate": 6.915254237288136e-06,
      "loss": 0.6193,
      "mean_token_accuracy": 0.8171321153640747,
      "num_tokens": 65262260.0,
      "step": 409
    },
    {
      "epoch": 0.20854526958290945,
      "grad_norm": 1.145785927772522,
      "learning_rate": 6.932203389830509e-06,
      "loss": 0.6059,
      "mean_token_accuracy": 0.8211365938186646,
      "num_tokens": 65425596.0,
      "step": 410
    },
    {
      "epoch": 0.20905391658189218,
      "grad_norm": 1.1632986068725586,
      "learning_rate": 6.949152542372882e-06,
      "loss": 0.6114,
      "mean_token_accuracy": 0.8181918263435364,
      "num_tokens": 65582379.0,
      "step": 411
    },
    {
      "epoch": 0.20956256358087488,
      "grad_norm": 1.200835108757019,
      "learning_rate": 6.966101694915255e-06,
      "loss": 0.6299,
      "mean_token_accuracy": 0.8122112154960632,
      "num_tokens": 65753543.0,
      "step": 412
    },
    {
      "epoch": 0.2100712105798576,
      "grad_norm": 1.144328236579895,
      "learning_rate": 6.9830508474576275e-06,
      "loss": 0.5904,
      "mean_token_accuracy": 0.8234543800354004,
      "num_tokens": 65920741.0,
      "step": 413
    },
    {
      "epoch": 0.2105798575788403,
      "grad_norm": 1.059288740158081,
      "learning_rate": 7e-06,
      "loss": 0.5766,
      "mean_token_accuracy": 0.8277970552444458,
      "num_tokens": 66079261.0,
      "step": 414
    },
    {
      "epoch": 0.211088504577823,
      "grad_norm": 1.083725929260254,
      "learning_rate": 7.016949152542374e-06,
      "loss": 0.6064,
      "mean_token_accuracy": 0.8182491064071655,
      "num_tokens": 66249455.0,
      "step": 415
    },
    {
      "epoch": 0.2115971515768057,
      "grad_norm": 1.195258617401123,
      "learning_rate": 7.033898305084746e-06,
      "loss": 0.6339,
      "mean_token_accuracy": 0.811680793762207,
      "num_tokens": 66410765.0,
      "step": 416
    },
    {
      "epoch": 0.2121057985757884,
      "grad_norm": 1.1450555324554443,
      "learning_rate": 7.0508474576271195e-06,
      "loss": 0.5983,
      "mean_token_accuracy": 0.8226056098937988,
      "num_tokens": 66564476.0,
      "step": 417
    },
    {
      "epoch": 0.2126144455747711,
      "grad_norm": 1.1363308429718018,
      "learning_rate": 7.067796610169492e-06,
      "loss": 0.6271,
      "mean_token_accuracy": 0.8152729272842407,
      "num_tokens": 66737804.0,
      "step": 418
    },
    {
      "epoch": 0.2131230925737538,
      "grad_norm": 1.1649593114852905,
      "learning_rate": 7.084745762711865e-06,
      "loss": 0.6233,
      "mean_token_accuracy": 0.8146108388900757,
      "num_tokens": 66895240.0,
      "step": 419
    },
    {
      "epoch": 0.2136317395727365,
      "grad_norm": 1.182517647743225,
      "learning_rate": 7.101694915254238e-06,
      "loss": 0.6034,
      "mean_token_accuracy": 0.8202614188194275,
      "num_tokens": 67056288.0,
      "step": 420
    },
    {
      "epoch": 0.21414038657171922,
      "grad_norm": 1.193202018737793,
      "learning_rate": 7.1186440677966106e-06,
      "loss": 0.6147,
      "mean_token_accuracy": 0.8172149658203125,
      "num_tokens": 67206207.0,
      "step": 421
    },
    {
      "epoch": 0.21464903357070192,
      "grad_norm": 1.1349166631698608,
      "learning_rate": 7.135593220338983e-06,
      "loss": 0.5734,
      "mean_token_accuracy": 0.8272489309310913,
      "num_tokens": 67369756.0,
      "step": 422
    },
    {
      "epoch": 0.21515768056968465,
      "grad_norm": 1.1669480800628662,
      "learning_rate": 7.152542372881357e-06,
      "loss": 0.586,
      "mean_token_accuracy": 0.822455883026123,
      "num_tokens": 67527516.0,
      "step": 423
    },
    {
      "epoch": 0.21566632756866735,
      "grad_norm": 1.1813911199569702,
      "learning_rate": 7.169491525423729e-06,
      "loss": 0.5945,
      "mean_token_accuracy": 0.8220915794372559,
      "num_tokens": 67688042.0,
      "step": 424
    },
    {
      "epoch": 0.21617497456765006,
      "grad_norm": 1.2895721197128296,
      "learning_rate": 7.1864406779661025e-06,
      "loss": 0.5985,
      "mean_token_accuracy": 0.8207123279571533,
      "num_tokens": 67854191.0,
      "step": 425
    },
    {
      "epoch": 0.21668362156663276,
      "grad_norm": 1.1939265727996826,
      "learning_rate": 7.203389830508475e-06,
      "loss": 0.5893,
      "mean_token_accuracy": 0.8222360610961914,
      "num_tokens": 68008533.0,
      "step": 426
    },
    {
      "epoch": 0.21719226856561547,
      "grad_norm": 1.2244884967803955,
      "learning_rate": 7.220338983050849e-06,
      "loss": 0.6094,
      "mean_token_accuracy": 0.8172198534011841,
      "num_tokens": 68162894.0,
      "step": 427
    },
    {
      "epoch": 0.21770091556459817,
      "grad_norm": 1.1649295091629028,
      "learning_rate": 7.237288135593221e-06,
      "loss": 0.5621,
      "mean_token_accuracy": 0.8311367630958557,
      "num_tokens": 68332401.0,
      "step": 428
    },
    {
      "epoch": 0.21820956256358087,
      "grad_norm": 1.2211418151855469,
      "learning_rate": 7.2542372881355936e-06,
      "loss": 0.5759,
      "mean_token_accuracy": 0.826878547668457,
      "num_tokens": 68489170.0,
      "step": 429
    },
    {
      "epoch": 0.21871820956256358,
      "grad_norm": 1.3775643110275269,
      "learning_rate": 7.271186440677967e-06,
      "loss": 0.6185,
      "mean_token_accuracy": 0.8159877061843872,
      "num_tokens": 68643444.0,
      "step": 430
    },
    {
      "epoch": 0.21922685656154628,
      "grad_norm": 1.2026543617248535,
      "learning_rate": 7.288135593220339e-06,
      "loss": 0.6023,
      "mean_token_accuracy": 0.8209705352783203,
      "num_tokens": 68812499.0,
      "step": 431
    },
    {
      "epoch": 0.21973550356052898,
      "grad_norm": 1.309005856513977,
      "learning_rate": 7.305084745762713e-06,
      "loss": 0.5912,
      "mean_token_accuracy": 0.8220794200897217,
      "num_tokens": 68976455.0,
      "step": 432
    },
    {
      "epoch": 0.2202441505595117,
      "grad_norm": 1.4292582273483276,
      "learning_rate": 7.3220338983050855e-06,
      "loss": 0.6372,
      "mean_token_accuracy": 0.8120876550674438,
      "num_tokens": 69142925.0,
      "step": 433
    },
    {
      "epoch": 0.22075279755849442,
      "grad_norm": 1.242922306060791,
      "learning_rate": 7.338983050847458e-06,
      "loss": 0.6253,
      "mean_token_accuracy": 0.8133541345596313,
      "num_tokens": 69307832.0,
      "step": 434
    },
    {
      "epoch": 0.22126144455747712,
      "grad_norm": 1.163408637046814,
      "learning_rate": 7.355932203389831e-06,
      "loss": 0.5749,
      "mean_token_accuracy": 0.8260928392410278,
      "num_tokens": 69466104.0,
      "step": 435
    },
    {
      "epoch": 0.22177009155645983,
      "grad_norm": 1.1773909330368042,
      "learning_rate": 7.372881355932204e-06,
      "loss": 0.5611,
      "mean_token_accuracy": 0.8318485021591187,
      "num_tokens": 69619859.0,
      "step": 436
    },
    {
      "epoch": 0.22227873855544253,
      "grad_norm": 1.2546722888946533,
      "learning_rate": 7.3898305084745766e-06,
      "loss": 0.5674,
      "mean_token_accuracy": 0.8298258781433105,
      "num_tokens": 69774687.0,
      "step": 437
    },
    {
      "epoch": 0.22278738555442523,
      "grad_norm": 1.3101048469543457,
      "learning_rate": 7.40677966101695e-06,
      "loss": 0.6228,
      "mean_token_accuracy": 0.8147264719009399,
      "num_tokens": 69924843.0,
      "step": 438
    },
    {
      "epoch": 0.22329603255340794,
      "grad_norm": 1.2783973217010498,
      "learning_rate": 7.423728813559322e-06,
      "loss": 0.6056,
      "mean_token_accuracy": 0.8204407691955566,
      "num_tokens": 70088441.0,
      "step": 439
    },
    {
      "epoch": 0.22380467955239064,
      "grad_norm": 1.2047439813613892,
      "learning_rate": 7.440677966101696e-06,
      "loss": 0.5955,
      "mean_token_accuracy": 0.8198369741439819,
      "num_tokens": 70244433.0,
      "step": 440
    },
    {
      "epoch": 0.22431332655137334,
      "grad_norm": 1.2373347282409668,
      "learning_rate": 7.4576271186440685e-06,
      "loss": 0.6113,
      "mean_token_accuracy": 0.8163229823112488,
      "num_tokens": 70412934.0,
      "step": 441
    },
    {
      "epoch": 0.22482197355035605,
      "grad_norm": 1.2173701524734497,
      "learning_rate": 7.474576271186441e-06,
      "loss": 0.6004,
      "mean_token_accuracy": 0.8211686015129089,
      "num_tokens": 70568120.0,
      "step": 442
    },
    {
      "epoch": 0.22533062054933875,
      "grad_norm": 1.2755416631698608,
      "learning_rate": 7.491525423728814e-06,
      "loss": 0.6173,
      "mean_token_accuracy": 0.8141643404960632,
      "num_tokens": 70725251.0,
      "step": 443
    },
    {
      "epoch": 0.22583926754832145,
      "grad_norm": 1.3401257991790771,
      "learning_rate": 7.508474576271187e-06,
      "loss": 0.613,
      "mean_token_accuracy": 0.8169766068458557,
      "num_tokens": 70881374.0,
      "step": 444
    },
    {
      "epoch": 0.22634791454730416,
      "grad_norm": 1.2987544536590576,
      "learning_rate": 7.52542372881356e-06,
      "loss": 0.6096,
      "mean_token_accuracy": 0.817240834236145,
      "num_tokens": 71029051.0,
      "step": 445
    },
    {
      "epoch": 0.2268565615462869,
      "grad_norm": 1.3045681715011597,
      "learning_rate": 7.542372881355933e-06,
      "loss": 0.5953,
      "mean_token_accuracy": 0.8227115869522095,
      "num_tokens": 71192310.0,
      "step": 446
    },
    {
      "epoch": 0.2273652085452696,
      "grad_norm": 1.2596161365509033,
      "learning_rate": 7.559322033898305e-06,
      "loss": 0.6205,
      "mean_token_accuracy": 0.8148484230041504,
      "num_tokens": 71355710.0,
      "step": 447
    },
    {
      "epoch": 0.2278738555442523,
      "grad_norm": 1.2501027584075928,
      "learning_rate": 7.576271186440679e-06,
      "loss": 0.624,
      "mean_token_accuracy": 0.8155069351196289,
      "num_tokens": 71509580.0,
      "step": 448
    },
    {
      "epoch": 0.228382502543235,
      "grad_norm": 1.1821587085723877,
      "learning_rate": 7.5932203389830515e-06,
      "loss": 0.5762,
      "mean_token_accuracy": 0.8275756239891052,
      "num_tokens": 71666995.0,
      "step": 449
    },
    {
      "epoch": 0.2288911495422177,
      "grad_norm": 1.2694047689437866,
      "learning_rate": 7.610169491525425e-06,
      "loss": 0.5938,
      "mean_token_accuracy": 0.8225254416465759,
      "num_tokens": 71825201.0,
      "step": 450
    },
    {
      "epoch": 0.2293997965412004,
      "grad_norm": 1.2232170104980469,
      "learning_rate": 7.627118644067797e-06,
      "loss": 0.5787,
      "mean_token_accuracy": 0.8244859576225281,
      "num_tokens": 71984961.0,
      "step": 451
    },
    {
      "epoch": 0.2299084435401831,
      "grad_norm": 1.261027455329895,
      "learning_rate": 7.64406779661017e-06,
      "loss": 0.6008,
      "mean_token_accuracy": 0.8211561441421509,
      "num_tokens": 72158603.0,
      "step": 452
    },
    {
      "epoch": 0.23041709053916581,
      "grad_norm": 1.2971307039260864,
      "learning_rate": 7.661016949152543e-06,
      "loss": 0.5777,
      "mean_token_accuracy": 0.8278082609176636,
      "num_tokens": 72313855.0,
      "step": 453
    },
    {
      "epoch": 0.23092573753814852,
      "grad_norm": 1.1784662008285522,
      "learning_rate": 7.677966101694917e-06,
      "loss": 0.5981,
      "mean_token_accuracy": 0.8198465704917908,
      "num_tokens": 72476307.0,
      "step": 454
    },
    {
      "epoch": 0.23143438453713122,
      "grad_norm": 1.2892508506774902,
      "learning_rate": 7.694915254237289e-06,
      "loss": 0.5903,
      "mean_token_accuracy": 0.8216632604598999,
      "num_tokens": 72630768.0,
      "step": 455
    },
    {
      "epoch": 0.23194303153611392,
      "grad_norm": 1.192068338394165,
      "learning_rate": 7.711864406779663e-06,
      "loss": 0.5911,
      "mean_token_accuracy": 0.8217214345932007,
      "num_tokens": 72805014.0,
      "step": 456
    },
    {
      "epoch": 0.23245167853509666,
      "grad_norm": 1.2026872634887695,
      "learning_rate": 7.728813559322035e-06,
      "loss": 0.5546,
      "mean_token_accuracy": 0.8341190218925476,
      "num_tokens": 72965078.0,
      "step": 457
    },
    {
      "epoch": 0.23296032553407936,
      "grad_norm": 1.1919342279434204,
      "learning_rate": 7.745762711864408e-06,
      "loss": 0.574,
      "mean_token_accuracy": 0.8285409808158875,
      "num_tokens": 73123811.0,
      "step": 458
    },
    {
      "epoch": 0.23346897253306206,
      "grad_norm": 1.2350062131881714,
      "learning_rate": 7.76271186440678e-06,
      "loss": 0.5785,
      "mean_token_accuracy": 0.8264639377593994,
      "num_tokens": 73286856.0,
      "step": 459
    },
    {
      "epoch": 0.23397761953204477,
      "grad_norm": 1.1893889904022217,
      "learning_rate": 7.779661016949152e-06,
      "loss": 0.6018,
      "mean_token_accuracy": 0.8182100057601929,
      "num_tokens": 73432474.0,
      "step": 460
    },
    {
      "epoch": 0.23448626653102747,
      "grad_norm": 1.2224371433258057,
      "learning_rate": 7.796610169491526e-06,
      "loss": 0.5944,
      "mean_token_accuracy": 0.8205611705780029,
      "num_tokens": 73589719.0,
      "step": 461
    },
    {
      "epoch": 0.23499491353001017,
      "grad_norm": 1.1163181066513062,
      "learning_rate": 7.813559322033899e-06,
      "loss": 0.5929,
      "mean_token_accuracy": 0.8211892247200012,
      "num_tokens": 73759617.0,
      "step": 462
    },
    {
      "epoch": 0.23550356052899288,
      "grad_norm": 1.317673683166504,
      "learning_rate": 7.830508474576271e-06,
      "loss": 0.6155,
      "mean_token_accuracy": 0.8157134056091309,
      "num_tokens": 73911743.0,
      "step": 463
    },
    {
      "epoch": 0.23601220752797558,
      "grad_norm": 1.2159312963485718,
      "learning_rate": 7.847457627118643e-06,
      "loss": 0.6012,
      "mean_token_accuracy": 0.8198647499084473,
      "num_tokens": 74082923.0,
      "step": 464
    },
    {
      "epoch": 0.23652085452695829,
      "grad_norm": 1.2829540967941284,
      "learning_rate": 7.864406779661017e-06,
      "loss": 0.5893,
      "mean_token_accuracy": 0.8235445022583008,
      "num_tokens": 74243087.0,
      "step": 465
    },
    {
      "epoch": 0.237029501525941,
      "grad_norm": 1.17428457736969,
      "learning_rate": 7.88135593220339e-06,
      "loss": 0.5769,
      "mean_token_accuracy": 0.8249427080154419,
      "num_tokens": 74410401.0,
      "step": 466
    },
    {
      "epoch": 0.2375381485249237,
      "grad_norm": 1.2670059204101562,
      "learning_rate": 7.898305084745764e-06,
      "loss": 0.5923,
      "mean_token_accuracy": 0.8222395777702332,
      "num_tokens": 74568603.0,
      "step": 467
    },
    {
      "epoch": 0.2380467955239064,
      "grad_norm": 1.3219184875488281,
      "learning_rate": 7.915254237288136e-06,
      "loss": 0.6009,
      "mean_token_accuracy": 0.8215954303741455,
      "num_tokens": 74727709.0,
      "step": 468
    },
    {
      "epoch": 0.23855544252288913,
      "grad_norm": 1.1674619913101196,
      "learning_rate": 7.93220338983051e-06,
      "loss": 0.5822,
      "mean_token_accuracy": 0.8243297338485718,
      "num_tokens": 74890980.0,
      "step": 469
    },
    {
      "epoch": 0.23906408952187183,
      "grad_norm": 1.2517329454421997,
      "learning_rate": 7.949152542372883e-06,
      "loss": 0.5495,
      "mean_token_accuracy": 0.8352354764938354,
      "num_tokens": 75056158.0,
      "step": 470
    },
    {
      "epoch": 0.23957273652085453,
      "grad_norm": 1.224470615386963,
      "learning_rate": 7.966101694915255e-06,
      "loss": 0.5858,
      "mean_token_accuracy": 0.8256843090057373,
      "num_tokens": 75217454.0,
      "step": 471
    },
    {
      "epoch": 0.24008138351983724,
      "grad_norm": 1.2624529600143433,
      "learning_rate": 7.983050847457627e-06,
      "loss": 0.5861,
      "mean_token_accuracy": 0.8228200674057007,
      "num_tokens": 75380129.0,
      "step": 472
    },
    {
      "epoch": 0.24059003051881994,
      "grad_norm": 1.237319827079773,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.5642,
      "mean_token_accuracy": 0.8284198045730591,
      "num_tokens": 75538207.0,
      "step": 473
    },
    {
      "epoch": 0.24109867751780265,
      "grad_norm": 1.1995705366134644,
      "learning_rate": 8.016949152542374e-06,
      "loss": 0.5607,
      "mean_token_accuracy": 0.8310052156448364,
      "num_tokens": 75692158.0,
      "step": 474
    },
    {
      "epoch": 0.24160732451678535,
      "grad_norm": 1.2999194860458374,
      "learning_rate": 8.033898305084746e-06,
      "loss": 0.5739,
      "mean_token_accuracy": 0.8259695768356323,
      "num_tokens": 75850325.0,
      "step": 475
    },
    {
      "epoch": 0.24211597151576805,
      "grad_norm": 1.207257628440857,
      "learning_rate": 8.050847457627118e-06,
      "loss": 0.5581,
      "mean_token_accuracy": 0.8317427635192871,
      "num_tokens": 76015367.0,
      "step": 476
    },
    {
      "epoch": 0.24262461851475076,
      "grad_norm": 1.1974592208862305,
      "learning_rate": 8.067796610169492e-06,
      "loss": 0.5692,
      "mean_token_accuracy": 0.8268818855285645,
      "num_tokens": 76165209.0,
      "step": 477
    },
    {
      "epoch": 0.24313326551373346,
      "grad_norm": 1.3411630392074585,
      "learning_rate": 8.084745762711865e-06,
      "loss": 0.6101,
      "mean_token_accuracy": 0.8171261548995972,
      "num_tokens": 76309618.0,
      "step": 478
    },
    {
      "epoch": 0.24364191251271616,
      "grad_norm": 1.2520160675048828,
      "learning_rate": 8.101694915254237e-06,
      "loss": 0.5645,
      "mean_token_accuracy": 0.8279448747634888,
      "num_tokens": 76468138.0,
      "step": 479
    },
    {
      "epoch": 0.2441505595116989,
      "grad_norm": 1.327293038368225,
      "learning_rate": 8.118644067796611e-06,
      "loss": 0.5847,
      "mean_token_accuracy": 0.8226466178894043,
      "num_tokens": 76638347.0,
      "step": 480
    },
    {
      "epoch": 0.2446592065106816,
      "grad_norm": 1.2002809047698975,
      "learning_rate": 8.135593220338983e-06,
      "loss": 0.6024,
      "mean_token_accuracy": 0.8184869289398193,
      "num_tokens": 76807912.0,
      "step": 481
    },
    {
      "epoch": 0.2451678535096643,
      "grad_norm": 1.269645094871521,
      "learning_rate": 8.152542372881358e-06,
      "loss": 0.5857,
      "mean_token_accuracy": 0.8238763809204102,
      "num_tokens": 76960377.0,
      "step": 482
    },
    {
      "epoch": 0.245676500508647,
      "grad_norm": 1.2257829904556274,
      "learning_rate": 8.16949152542373e-06,
      "loss": 0.644,
      "mean_token_accuracy": 0.8092539310455322,
      "num_tokens": 77135219.0,
      "step": 483
    },
    {
      "epoch": 0.2461851475076297,
      "grad_norm": 1.2210476398468018,
      "learning_rate": 8.186440677966102e-06,
      "loss": 0.6071,
      "mean_token_accuracy": 0.8186261653900146,
      "num_tokens": 77284277.0,
      "step": 484
    },
    {
      "epoch": 0.2466937945066124,
      "grad_norm": 1.1874899864196777,
      "learning_rate": 8.203389830508475e-06,
      "loss": 0.5955,
      "mean_token_accuracy": 0.8232449293136597,
      "num_tokens": 77463576.0,
      "step": 485
    },
    {
      "epoch": 0.24720244150559512,
      "grad_norm": 1.288148283958435,
      "learning_rate": 8.220338983050849e-06,
      "loss": 0.5746,
      "mean_token_accuracy": 0.8276445865631104,
      "num_tokens": 77624014.0,
      "step": 486
    },
    {
      "epoch": 0.24771108850457782,
      "grad_norm": 1.3062938451766968,
      "learning_rate": 8.237288135593221e-06,
      "loss": 0.5965,
      "mean_token_accuracy": 0.8204386234283447,
      "num_tokens": 77786805.0,
      "step": 487
    },
    {
      "epoch": 0.24821973550356052,
      "grad_norm": 1.3395079374313354,
      "learning_rate": 8.254237288135593e-06,
      "loss": 0.6057,
      "mean_token_accuracy": 0.8180563449859619,
      "num_tokens": 77955105.0,
      "step": 488
    },
    {
      "epoch": 0.24872838250254323,
      "grad_norm": 1.370811104774475,
      "learning_rate": 8.271186440677966e-06,
      "loss": 0.5758,
      "mean_token_accuracy": 0.8270718455314636,
      "num_tokens": 78103025.0,
      "step": 489
    },
    {
      "epoch": 0.24923702950152593,
      "grad_norm": 1.405150294303894,
      "learning_rate": 8.28813559322034e-06,
      "loss": 0.5757,
      "mean_token_accuracy": 0.826675534248352,
      "num_tokens": 78259934.0,
      "step": 490
    },
    {
      "epoch": 0.24974567650050863,
      "grad_norm": 1.173972249031067,
      "learning_rate": 8.305084745762712e-06,
      "loss": 0.5774,
      "mean_token_accuracy": 0.8275264501571655,
      "num_tokens": 78420450.0,
      "step": 491
    },
    {
      "epoch": 0.25025432349949134,
      "grad_norm": 1.3530043363571167,
      "learning_rate": 8.322033898305086e-06,
      "loss": 0.5926,
      "mean_token_accuracy": 0.8219928741455078,
      "num_tokens": 78575928.0,
      "step": 492
    },
    {
      "epoch": 0.25076297049847407,
      "grad_norm": 1.2971301078796387,
      "learning_rate": 8.338983050847458e-06,
      "loss": 0.5863,
      "mean_token_accuracy": 0.8221102356910706,
      "num_tokens": 78723238.0,
      "step": 493
    },
    {
      "epoch": 0.25127161749745675,
      "grad_norm": 1.2696784734725952,
      "learning_rate": 8.35593220338983e-06,
      "loss": 0.5965,
      "mean_token_accuracy": 0.8210808038711548,
      "num_tokens": 78897539.0,
      "step": 494
    },
    {
      "epoch": 0.2517802644964395,
      "grad_norm": 1.2298940420150757,
      "learning_rate": 8.372881355932205e-06,
      "loss": 0.5512,
      "mean_token_accuracy": 0.8342494964599609,
      "num_tokens": 79059834.0,
      "step": 495
    },
    {
      "epoch": 0.25228891149542215,
      "grad_norm": 1.2379639148712158,
      "learning_rate": 8.389830508474577e-06,
      "loss": 0.5971,
      "mean_token_accuracy": 0.8209534883499146,
      "num_tokens": 79206412.0,
      "step": 496
    },
    {
      "epoch": 0.2527975584944049,
      "grad_norm": 1.1856814622879028,
      "learning_rate": 8.40677966101695e-06,
      "loss": 0.5919,
      "mean_token_accuracy": 0.8220817446708679,
      "num_tokens": 79379375.0,
      "step": 497
    },
    {
      "epoch": 0.2533062054933876,
      "grad_norm": 5.187645435333252,
      "learning_rate": 8.423728813559324e-06,
      "loss": 0.5795,
      "mean_token_accuracy": 0.8255907297134399,
      "num_tokens": 79550225.0,
      "step": 498
    },
    {
      "epoch": 0.2538148524923703,
      "grad_norm": 1.6230169534683228,
      "learning_rate": 8.440677966101696e-06,
      "loss": 0.5654,
      "mean_token_accuracy": 0.8300318121910095,
      "num_tokens": 79705365.0,
      "step": 499
    },
    {
      "epoch": 0.254323499491353,
      "grad_norm": 1.5571225881576538,
      "learning_rate": 8.457627118644068e-06,
      "loss": 0.5913,
      "mean_token_accuracy": 0.8222099542617798,
      "num_tokens": 79861100.0,
      "step": 500
    },
    {
      "epoch": 0.2548321464903357,
      "grad_norm": 1.3249210119247437,
      "learning_rate": 8.47457627118644e-06,
      "loss": 0.5905,
      "mean_token_accuracy": 0.8226123452186584,
      "num_tokens": 80017875.0,
      "step": 501
    },
    {
      "epoch": 0.25534079348931843,
      "grad_norm": 1.379267930984497,
      "learning_rate": 8.491525423728815e-06,
      "loss": 0.5906,
      "mean_token_accuracy": 0.8216269016265869,
      "num_tokens": 80184953.0,
      "step": 502
    },
    {
      "epoch": 0.2558494404883011,
      "grad_norm": 1.2471497058868408,
      "learning_rate": 8.508474576271187e-06,
      "loss": 0.5887,
      "mean_token_accuracy": 0.8234388828277588,
      "num_tokens": 80342864.0,
      "step": 503
    },
    {
      "epoch": 0.25635808748728384,
      "grad_norm": 1.3340461254119873,
      "learning_rate": 8.52542372881356e-06,
      "loss": 0.5575,
      "mean_token_accuracy": 0.8304026126861572,
      "num_tokens": 80497644.0,
      "step": 504
    },
    {
      "epoch": 0.2568667344862665,
      "grad_norm": 1.4473865032196045,
      "learning_rate": 8.542372881355933e-06,
      "loss": 0.6242,
      "mean_token_accuracy": 0.8131608963012695,
      "num_tokens": 80652132.0,
      "step": 505
    },
    {
      "epoch": 0.25737538148524924,
      "grad_norm": 1.4180216789245605,
      "learning_rate": 8.559322033898306e-06,
      "loss": 0.584,
      "mean_token_accuracy": 0.8273070454597473,
      "num_tokens": 80815501.0,
      "step": 506
    },
    {
      "epoch": 0.2578840284842319,
      "grad_norm": 1.296979546546936,
      "learning_rate": 8.57627118644068e-06,
      "loss": 0.6033,
      "mean_token_accuracy": 0.818980872631073,
      "num_tokens": 80982538.0,
      "step": 507
    },
    {
      "epoch": 0.25839267548321465,
      "grad_norm": 1.4863985776901245,
      "learning_rate": 8.593220338983052e-06,
      "loss": 0.6042,
      "mean_token_accuracy": 0.8181418776512146,
      "num_tokens": 81137163.0,
      "step": 508
    },
    {
      "epoch": 0.2589013224821974,
      "grad_norm": 1.335146427154541,
      "learning_rate": 8.610169491525424e-06,
      "loss": 0.6109,
      "mean_token_accuracy": 0.8161569833755493,
      "num_tokens": 81304752.0,
      "step": 509
    },
    {
      "epoch": 0.25940996948118006,
      "grad_norm": 1.484490990638733,
      "learning_rate": 8.627118644067798e-06,
      "loss": 0.5753,
      "mean_token_accuracy": 0.8248833417892456,
      "num_tokens": 81448192.0,
      "step": 510
    },
    {
      "epoch": 0.2599186164801628,
      "grad_norm": 1.4455455541610718,
      "learning_rate": 8.64406779661017e-06,
      "loss": 0.6117,
      "mean_token_accuracy": 0.816500723361969,
      "num_tokens": 81608372.0,
      "step": 511
    },
    {
      "epoch": 0.26042726347914547,
      "grad_norm": 1.2930644750595093,
      "learning_rate": 8.661016949152543e-06,
      "loss": 0.5756,
      "mean_token_accuracy": 0.8255362510681152,
      "num_tokens": 81757788.0,
      "step": 512
    },
    {
      "epoch": 0.2609359104781282,
      "grad_norm": 1.2497493028640747,
      "learning_rate": 8.677966101694915e-06,
      "loss": 0.5735,
      "mean_token_accuracy": 0.8275926113128662,
      "num_tokens": 81917301.0,
      "step": 513
    },
    {
      "epoch": 0.2614445574771109,
      "grad_norm": 1.1094409227371216,
      "learning_rate": 8.694915254237288e-06,
      "loss": 0.5829,
      "mean_token_accuracy": 0.8236455917358398,
      "num_tokens": 82085431.0,
      "step": 514
    },
    {
      "epoch": 0.2619532044760936,
      "grad_norm": 1.2383760213851929,
      "learning_rate": 8.711864406779662e-06,
      "loss": 0.5753,
      "mean_token_accuracy": 0.8278086185455322,
      "num_tokens": 82238284.0,
      "step": 515
    },
    {
      "epoch": 0.2624618514750763,
      "grad_norm": 1.1112157106399536,
      "learning_rate": 8.728813559322034e-06,
      "loss": 0.6045,
      "mean_token_accuracy": 0.8183934688568115,
      "num_tokens": 82404647.0,
      "step": 516
    },
    {
      "epoch": 0.262970498474059,
      "grad_norm": 1.2201389074325562,
      "learning_rate": 8.745762711864407e-06,
      "loss": 0.5801,
      "mean_token_accuracy": 0.8245712518692017,
      "num_tokens": 82572960.0,
      "step": 517
    },
    {
      "epoch": 0.2634791454730417,
      "grad_norm": 1.177706003189087,
      "learning_rate": 8.76271186440678e-06,
      "loss": 0.5881,
      "mean_token_accuracy": 0.8239415884017944,
      "num_tokens": 82738083.0,
      "step": 518
    },
    {
      "epoch": 0.2639877924720244,
      "grad_norm": 1.3163994550704956,
      "learning_rate": 8.779661016949153e-06,
      "loss": 0.5782,
      "mean_token_accuracy": 0.8258814215660095,
      "num_tokens": 82887087.0,
      "step": 519
    },
    {
      "epoch": 0.2644964394710071,
      "grad_norm": 1.3369812965393066,
      "learning_rate": 8.796610169491527e-06,
      "loss": 0.5901,
      "mean_token_accuracy": 0.8220720291137695,
      "num_tokens": 83033584.0,
      "step": 520
    },
    {
      "epoch": 0.2650050864699898,
      "grad_norm": 1.3403503894805908,
      "learning_rate": 8.8135593220339e-06,
      "loss": 0.5989,
      "mean_token_accuracy": 0.8187283277511597,
      "num_tokens": 83195499.0,
      "step": 521
    },
    {
      "epoch": 0.26551373346897256,
      "grad_norm": 1.2609878778457642,
      "learning_rate": 8.830508474576272e-06,
      "loss": 0.5901,
      "mean_token_accuracy": 0.8238098621368408,
      "num_tokens": 83363928.0,
      "step": 522
    },
    {
      "epoch": 0.26602238046795523,
      "grad_norm": 1.2445399761199951,
      "learning_rate": 8.847457627118646e-06,
      "loss": 0.5802,
      "mean_token_accuracy": 0.8244575262069702,
      "num_tokens": 83520177.0,
      "step": 523
    },
    {
      "epoch": 0.26653102746693796,
      "grad_norm": 1.289551854133606,
      "learning_rate": 8.864406779661018e-06,
      "loss": 0.5654,
      "mean_token_accuracy": 0.8277870416641235,
      "num_tokens": 83680255.0,
      "step": 524
    },
    {
      "epoch": 0.26703967446592064,
      "grad_norm": 1.2121199369430542,
      "learning_rate": 8.88135593220339e-06,
      "loss": 0.5824,
      "mean_token_accuracy": 0.8237113952636719,
      "num_tokens": 83852400.0,
      "step": 525
    },
    {
      "epoch": 0.26754832146490337,
      "grad_norm": 1.316636562347412,
      "learning_rate": 8.898305084745763e-06,
      "loss": 0.611,
      "mean_token_accuracy": 0.8159763813018799,
      "num_tokens": 84014162.0,
      "step": 526
    },
    {
      "epoch": 0.26805696846388605,
      "grad_norm": 1.1623677015304565,
      "learning_rate": 8.915254237288137e-06,
      "loss": 0.5672,
      "mean_token_accuracy": 0.8281941413879395,
      "num_tokens": 84170100.0,
      "step": 527
    },
    {
      "epoch": 0.2685656154628688,
      "grad_norm": 1.1831893920898438,
      "learning_rate": 8.932203389830509e-06,
      "loss": 0.5499,
      "mean_token_accuracy": 0.833954930305481,
      "num_tokens": 84329460.0,
      "step": 528
    },
    {
      "epoch": 0.26907426246185145,
      "grad_norm": 1.1870083808898926,
      "learning_rate": 8.949152542372881e-06,
      "loss": 0.5441,
      "mean_token_accuracy": 0.8352469801902771,
      "num_tokens": 84486375.0,
      "step": 529
    },
    {
      "epoch": 0.2695829094608342,
      "grad_norm": 1.3223252296447754,
      "learning_rate": 8.966101694915254e-06,
      "loss": 0.5534,
      "mean_token_accuracy": 0.8336907625198364,
      "num_tokens": 84643661.0,
      "step": 530
    },
    {
      "epoch": 0.27009155645981686,
      "grad_norm": 1.1877634525299072,
      "learning_rate": 8.983050847457628e-06,
      "loss": 0.5679,
      "mean_token_accuracy": 0.829535186290741,
      "num_tokens": 84812504.0,
      "step": 531
    },
    {
      "epoch": 0.2706002034587996,
      "grad_norm": 1.2430506944656372,
      "learning_rate": 9e-06,
      "loss": 0.5689,
      "mean_token_accuracy": 0.8278781771659851,
      "num_tokens": 84968874.0,
      "step": 532
    },
    {
      "epoch": 0.2711088504577823,
      "grad_norm": 1.1894301176071167,
      "learning_rate": 9.016949152542374e-06,
      "loss": 0.5746,
      "mean_token_accuracy": 0.824907660484314,
      "num_tokens": 85129918.0,
      "step": 533
    },
    {
      "epoch": 0.271617497456765,
      "grad_norm": 1.2176803350448608,
      "learning_rate": 9.033898305084747e-06,
      "loss": 0.5669,
      "mean_token_accuracy": 0.8289027214050293,
      "num_tokens": 85292100.0,
      "step": 534
    },
    {
      "epoch": 0.27212614445574773,
      "grad_norm": 1.238106608390808,
      "learning_rate": 9.05084745762712e-06,
      "loss": 0.552,
      "mean_token_accuracy": 0.8348509073257446,
      "num_tokens": 85445236.0,
      "step": 535
    },
    {
      "epoch": 0.2726347914547304,
      "grad_norm": 1.2461129426956177,
      "learning_rate": 9.067796610169493e-06,
      "loss": 0.587,
      "mean_token_accuracy": 0.822391927242279,
      "num_tokens": 85605945.0,
      "step": 536
    },
    {
      "epoch": 0.27314343845371314,
      "grad_norm": 1.171604871749878,
      "learning_rate": 9.084745762711865e-06,
      "loss": 0.5675,
      "mean_token_accuracy": 0.8278486728668213,
      "num_tokens": 85760177.0,
      "step": 537
    },
    {
      "epoch": 0.2736520854526958,
      "grad_norm": 1.1491703987121582,
      "learning_rate": 9.101694915254238e-06,
      "loss": 0.5855,
      "mean_token_accuracy": 0.8246716260910034,
      "num_tokens": 85924331.0,
      "step": 538
    },
    {
      "epoch": 0.27416073245167855,
      "grad_norm": 1.1518174409866333,
      "learning_rate": 9.11864406779661e-06,
      "loss": 0.6048,
      "mean_token_accuracy": 0.8184250593185425,
      "num_tokens": 86087686.0,
      "step": 539
    },
    {
      "epoch": 0.2746693794506612,
      "grad_norm": 1.2020530700683594,
      "learning_rate": 9.135593220338984e-06,
      "loss": 0.6024,
      "mean_token_accuracy": 0.8186061978340149,
      "num_tokens": 86231923.0,
      "step": 540
    },
    {
      "epoch": 0.27517802644964395,
      "grad_norm": 1.246408462524414,
      "learning_rate": 9.152542372881356e-06,
      "loss": 0.5731,
      "mean_token_accuracy": 0.8254373669624329,
      "num_tokens": 86393599.0,
      "step": 541
    },
    {
      "epoch": 0.27568667344862663,
      "grad_norm": 1.2908697128295898,
      "learning_rate": 9.169491525423729e-06,
      "loss": 0.5437,
      "mean_token_accuracy": 0.8329447507858276,
      "num_tokens": 86546912.0,
      "step": 542
    },
    {
      "epoch": 0.27619532044760936,
      "grad_norm": 1.3099312782287598,
      "learning_rate": 9.186440677966101e-06,
      "loss": 0.6402,
      "mean_token_accuracy": 0.8103553056716919,
      "num_tokens": 86710627.0,
      "step": 543
    },
    {
      "epoch": 0.2767039674465921,
      "grad_norm": 1.2442010641098022,
      "learning_rate": 9.203389830508475e-06,
      "loss": 0.5928,
      "mean_token_accuracy": 0.8202736377716064,
      "num_tokens": 86872441.0,
      "step": 544
    },
    {
      "epoch": 0.27721261444557477,
      "grad_norm": 1.3643654584884644,
      "learning_rate": 9.220338983050847e-06,
      "loss": 0.5981,
      "mean_token_accuracy": 0.8186464309692383,
      "num_tokens": 87040426.0,
      "step": 545
    },
    {
      "epoch": 0.2777212614445575,
      "grad_norm": 1.2602033615112305,
      "learning_rate": 9.237288135593222e-06,
      "loss": 0.5838,
      "mean_token_accuracy": 0.8244240283966064,
      "num_tokens": 87204457.0,
      "step": 546
    },
    {
      "epoch": 0.2782299084435402,
      "grad_norm": 1.3273673057556152,
      "learning_rate": 9.254237288135594e-06,
      "loss": 0.5934,
      "mean_token_accuracy": 0.8224654197692871,
      "num_tokens": 87350790.0,
      "step": 547
    },
    {
      "epoch": 0.2787385554425229,
      "grad_norm": 1.3505678176879883,
      "learning_rate": 9.271186440677968e-06,
      "loss": 0.5992,
      "mean_token_accuracy": 0.8188058137893677,
      "num_tokens": 87501470.0,
      "step": 548
    },
    {
      "epoch": 0.2792472024415056,
      "grad_norm": 1.3277742862701416,
      "learning_rate": 9.28813559322034e-06,
      "loss": 0.5902,
      "mean_token_accuracy": 0.8227229118347168,
      "num_tokens": 87654280.0,
      "step": 549
    },
    {
      "epoch": 0.2797558494404883,
      "grad_norm": 1.2117854356765747,
      "learning_rate": 9.305084745762713e-06,
      "loss": 0.5726,
      "mean_token_accuracy": 0.8264171481132507,
      "num_tokens": 87816423.0,
      "step": 550
    },
    {
      "epoch": 0.280264496439471,
      "grad_norm": 1.4077842235565186,
      "learning_rate": 9.322033898305085e-06,
      "loss": 0.6106,
      "mean_token_accuracy": 0.8154207468032837,
      "num_tokens": 87960156.0,
      "step": 551
    },
    {
      "epoch": 0.2807731434384537,
      "grad_norm": 1.3092397451400757,
      "learning_rate": 9.338983050847459e-06,
      "loss": 0.5554,
      "mean_token_accuracy": 0.8319989442825317,
      "num_tokens": 88105701.0,
      "step": 552
    },
    {
      "epoch": 0.2812817904374364,
      "grad_norm": 1.2441811561584473,
      "learning_rate": 9.355932203389831e-06,
      "loss": 0.5804,
      "mean_token_accuracy": 0.8244675397872925,
      "num_tokens": 88261189.0,
      "step": 553
    },
    {
      "epoch": 0.28179043743641913,
      "grad_norm": 1.3278307914733887,
      "learning_rate": 9.372881355932204e-06,
      "loss": 0.592,
      "mean_token_accuracy": 0.8210278153419495,
      "num_tokens": 88429993.0,
      "step": 554
    },
    {
      "epoch": 0.2822990844354018,
      "grad_norm": 1.2581342458724976,
      "learning_rate": 9.389830508474576e-06,
      "loss": 0.578,
      "mean_token_accuracy": 0.8260815143585205,
      "num_tokens": 88598116.0,
      "step": 555
    },
    {
      "epoch": 0.28280773143438453,
      "grad_norm": 1.350409746170044,
      "learning_rate": 9.40677966101695e-06,
      "loss": 0.6042,
      "mean_token_accuracy": 0.8164364099502563,
      "num_tokens": 88754109.0,
      "step": 556
    },
    {
      "epoch": 0.28331637843336727,
      "grad_norm": 1.393012285232544,
      "learning_rate": 9.423728813559322e-06,
      "loss": 0.5888,
      "mean_token_accuracy": 0.8246204853057861,
      "num_tokens": 88911407.0,
      "step": 557
    },
    {
      "epoch": 0.28382502543234994,
      "grad_norm": 1.4434447288513184,
      "learning_rate": 9.440677966101696e-06,
      "loss": 0.6229,
      "mean_token_accuracy": 0.8114631175994873,
      "num_tokens": 89075556.0,
      "step": 558
    },
    {
      "epoch": 0.2843336724313327,
      "grad_norm": 1.2736214399337769,
      "learning_rate": 9.457627118644069e-06,
      "loss": 0.5838,
      "mean_token_accuracy": 0.8232846260070801,
      "num_tokens": 89234090.0,
      "step": 559
    },
    {
      "epoch": 0.28484231943031535,
      "grad_norm": 1.4164948463439941,
      "learning_rate": 9.474576271186441e-06,
      "loss": 0.596,
      "mean_token_accuracy": 0.8208360075950623,
      "num_tokens": 89394941.0,
      "step": 560
    },
    {
      "epoch": 0.2853509664292981,
      "grad_norm": 1.3688839673995972,
      "learning_rate": 9.491525423728815e-06,
      "loss": 0.5935,
      "mean_token_accuracy": 0.8203197717666626,
      "num_tokens": 89560585.0,
      "step": 561
    },
    {
      "epoch": 0.28585961342828076,
      "grad_norm": 1.2683659791946411,
      "learning_rate": 9.508474576271188e-06,
      "loss": 0.5913,
      "mean_token_accuracy": 0.8229166865348816,
      "num_tokens": 89726320.0,
      "step": 562
    },
    {
      "epoch": 0.2863682604272635,
      "grad_norm": 1.242645502090454,
      "learning_rate": 9.52542372881356e-06,
      "loss": 0.5569,
      "mean_token_accuracy": 0.8335822224617004,
      "num_tokens": 89880236.0,
      "step": 563
    },
    {
      "epoch": 0.28687690742624616,
      "grad_norm": 1.1361278295516968,
      "learning_rate": 9.542372881355934e-06,
      "loss": 0.6169,
      "mean_token_accuracy": 0.8147539496421814,
      "num_tokens": 90046963.0,
      "step": 564
    },
    {
      "epoch": 0.2873855544252289,
      "grad_norm": 1.174190640449524,
      "learning_rate": 9.559322033898306e-06,
      "loss": 0.5615,
      "mean_token_accuracy": 0.8308785557746887,
      "num_tokens": 90208654.0,
      "step": 565
    },
    {
      "epoch": 0.28789420142421157,
      "grad_norm": 1.269404649734497,
      "learning_rate": 9.576271186440679e-06,
      "loss": 0.5913,
      "mean_token_accuracy": 0.8214564323425293,
      "num_tokens": 90360145.0,
      "step": 566
    },
    {
      "epoch": 0.2884028484231943,
      "grad_norm": 1.4272724390029907,
      "learning_rate": 9.593220338983051e-06,
      "loss": 0.5825,
      "mean_token_accuracy": 0.8251073360443115,
      "num_tokens": 90517341.0,
      "step": 567
    },
    {
      "epoch": 0.28891149542217703,
      "grad_norm": 1.2705575227737427,
      "learning_rate": 9.610169491525423e-06,
      "loss": 0.5811,
      "mean_token_accuracy": 0.8249518275260925,
      "num_tokens": 90666248.0,
      "step": 568
    },
    {
      "epoch": 0.2894201424211597,
      "grad_norm": 1.2269147634506226,
      "learning_rate": 9.627118644067797e-06,
      "loss": 0.5657,
      "mean_token_accuracy": 0.8296713829040527,
      "num_tokens": 90824390.0,
      "step": 569
    },
    {
      "epoch": 0.28992878942014244,
      "grad_norm": 1.2506111860275269,
      "learning_rate": 9.64406779661017e-06,
      "loss": 0.5953,
      "mean_token_accuracy": 0.8209437727928162,
      "num_tokens": 90990168.0,
      "step": 570
    },
    {
      "epoch": 0.2904374364191251,
      "grad_norm": 1.2558051347732544,
      "learning_rate": 9.661016949152544e-06,
      "loss": 0.5779,
      "mean_token_accuracy": 0.8250951170921326,
      "num_tokens": 91130945.0,
      "step": 571
    },
    {
      "epoch": 0.29094608341810785,
      "grad_norm": 1.2120747566223145,
      "learning_rate": 9.677966101694916e-06,
      "loss": 0.5693,
      "mean_token_accuracy": 0.8284271955490112,
      "num_tokens": 91287120.0,
      "step": 572
    },
    {
      "epoch": 0.2914547304170905,
      "grad_norm": 1.3007855415344238,
      "learning_rate": 9.69491525423729e-06,
      "loss": 0.5856,
      "mean_token_accuracy": 0.82295161485672,
      "num_tokens": 91445888.0,
      "step": 573
    },
    {
      "epoch": 0.29196337741607326,
      "grad_norm": 1.2200978994369507,
      "learning_rate": 9.711864406779662e-06,
      "loss": 0.5471,
      "mean_token_accuracy": 0.8336622714996338,
      "num_tokens": 91605888.0,
      "step": 574
    },
    {
      "epoch": 0.29247202441505593,
      "grad_norm": 1.305680274963379,
      "learning_rate": 9.728813559322035e-06,
      "loss": 0.6212,
      "mean_token_accuracy": 0.8143027424812317,
      "num_tokens": 91766382.0,
      "step": 575
    },
    {
      "epoch": 0.29298067141403866,
      "grad_norm": 1.149032711982727,
      "learning_rate": 9.745762711864407e-06,
      "loss": 0.5802,
      "mean_token_accuracy": 0.8255360126495361,
      "num_tokens": 91923111.0,
      "step": 576
    },
    {
      "epoch": 0.29348931841302134,
      "grad_norm": 1.212319254875183,
      "learning_rate": 9.762711864406781e-06,
      "loss": 0.602,
      "mean_token_accuracy": 0.8194811344146729,
      "num_tokens": 92076217.0,
      "step": 577
    },
    {
      "epoch": 0.29399796541200407,
      "grad_norm": 1.2744840383529663,
      "learning_rate": 9.779661016949154e-06,
      "loss": 0.5347,
      "mean_token_accuracy": 0.836704432964325,
      "num_tokens": 92233100.0,
      "step": 578
    },
    {
      "epoch": 0.2945066124109868,
      "grad_norm": 1.250063419342041,
      "learning_rate": 9.796610169491526e-06,
      "loss": 0.6222,
      "mean_token_accuracy": 0.8125181198120117,
      "num_tokens": 92384392.0,
      "step": 579
    },
    {
      "epoch": 0.2950152594099695,
      "grad_norm": 1.2378636598587036,
      "learning_rate": 9.813559322033898e-06,
      "loss": 0.5395,
      "mean_token_accuracy": 0.8373066186904907,
      "num_tokens": 92538563.0,
      "step": 580
    },
    {
      "epoch": 0.2955239064089522,
      "grad_norm": 1.3510687351226807,
      "learning_rate": 9.830508474576272e-06,
      "loss": 0.5621,
      "mean_token_accuracy": 0.8300359845161438,
      "num_tokens": 92702514.0,
      "step": 581
    },
    {
      "epoch": 0.2960325534079349,
      "grad_norm": 1.1700639724731445,
      "learning_rate": 9.847457627118645e-06,
      "loss": 0.5268,
      "mean_token_accuracy": 0.8393378257751465,
      "num_tokens": 92861897.0,
      "step": 582
    },
    {
      "epoch": 0.2965412004069176,
      "grad_norm": 1.4047576189041138,
      "learning_rate": 9.864406779661017e-06,
      "loss": 0.5431,
      "mean_token_accuracy": 0.8350200653076172,
      "num_tokens": 93021485.0,
      "step": 583
    },
    {
      "epoch": 0.2970498474059003,
      "grad_norm": 1.2136011123657227,
      "learning_rate": 9.881355932203391e-06,
      "loss": 0.5115,
      "mean_token_accuracy": 0.8411509990692139,
      "num_tokens": 93169971.0,
      "step": 584
    },
    {
      "epoch": 0.297558494404883,
      "grad_norm": 1.3670368194580078,
      "learning_rate": 9.898305084745763e-06,
      "loss": 0.5697,
      "mean_token_accuracy": 0.8275981545448303,
      "num_tokens": 93321364.0,
      "step": 585
    },
    {
      "epoch": 0.2980671414038657,
      "grad_norm": 1.283612847328186,
      "learning_rate": 9.915254237288137e-06,
      "loss": 0.5931,
      "mean_token_accuracy": 0.8204793930053711,
      "num_tokens": 93476129.0,
      "step": 586
    },
    {
      "epoch": 0.29857578840284843,
      "grad_norm": 1.3070842027664185,
      "learning_rate": 9.93220338983051e-06,
      "loss": 0.5813,
      "mean_token_accuracy": 0.8225874304771423,
      "num_tokens": 93637944.0,
      "step": 587
    },
    {
      "epoch": 0.2990844354018311,
      "grad_norm": 1.1480412483215332,
      "learning_rate": 9.949152542372882e-06,
      "loss": 0.5669,
      "mean_token_accuracy": 0.8301283121109009,
      "num_tokens": 93795891.0,
      "step": 588
    },
    {
      "epoch": 0.29959308240081384,
      "grad_norm": 1.242099404335022,
      "learning_rate": 9.966101694915256e-06,
      "loss": 0.6161,
      "mean_token_accuracy": 0.8146437406539917,
      "num_tokens": 93949753.0,
      "step": 589
    },
    {
      "epoch": 0.30010172939979657,
      "grad_norm": 1.2934678792953491,
      "learning_rate": 9.983050847457628e-06,
      "loss": 0.5363,
      "mean_token_accuracy": 0.8385515213012695,
      "num_tokens": 94099353.0,
      "step": 590
    },
    {
      "epoch": 0.30061037639877924,
      "grad_norm": 1.4157675504684448,
      "learning_rate": 1e-05,
      "loss": 0.6421,
      "mean_token_accuracy": 0.808651328086853,
      "num_tokens": 94262648.0,
      "step": 591
    },
    {
      "epoch": 0.301119023397762,
      "grad_norm": 1.358555793762207,
      "learning_rate": 1e-05,
      "loss": 0.582,
      "mean_token_accuracy": 0.8241909742355347,
      "num_tokens": 94420660.0,
      "step": 592
    },
    {
      "epoch": 0.30162767039674465,
      "grad_norm": 1.3359708786010742,
      "learning_rate": 1e-05,
      "loss": 0.5721,
      "mean_token_accuracy": 0.8296599388122559,
      "num_tokens": 94580578.0,
      "step": 593
    },
    {
      "epoch": 0.3021363173957274,
      "grad_norm": 1.4047656059265137,
      "learning_rate": 1e-05,
      "loss": 0.5795,
      "mean_token_accuracy": 0.8244053721427917,
      "num_tokens": 94745810.0,
      "step": 594
    },
    {
      "epoch": 0.30264496439471006,
      "grad_norm": 1.289602518081665,
      "learning_rate": 1e-05,
      "loss": 0.5907,
      "mean_token_accuracy": 0.8213446140289307,
      "num_tokens": 94904762.0,
      "step": 595
    },
    {
      "epoch": 0.3031536113936928,
      "grad_norm": 1.376883625984192,
      "learning_rate": 1e-05,
      "loss": 0.5718,
      "mean_token_accuracy": 0.8260505199432373,
      "num_tokens": 95053065.0,
      "step": 596
    },
    {
      "epoch": 0.30366225839267547,
      "grad_norm": 1.2694602012634277,
      "learning_rate": 1e-05,
      "loss": 0.5914,
      "mean_token_accuracy": 0.8208502531051636,
      "num_tokens": 95207324.0,
      "step": 597
    },
    {
      "epoch": 0.3041709053916582,
      "grad_norm": 1.331419825553894,
      "learning_rate": 1e-05,
      "loss": 0.5597,
      "mean_token_accuracy": 0.8295787572860718,
      "num_tokens": 95365564.0,
      "step": 598
    },
    {
      "epoch": 0.3046795523906409,
      "grad_norm": 1.2676899433135986,
      "learning_rate": 1e-05,
      "loss": 0.5765,
      "mean_token_accuracy": 0.8234357833862305,
      "num_tokens": 95522523.0,
      "step": 599
    },
    {
      "epoch": 0.3051881993896236,
      "grad_norm": 1.3706655502319336,
      "learning_rate": 1e-05,
      "loss": 0.5494,
      "mean_token_accuracy": 0.8335732221603394,
      "num_tokens": 95677617.0,
      "step": 600
    },
    {
      "epoch": 0.3056968463886063,
      "grad_norm": 1.1500375270843506,
      "learning_rate": 1e-05,
      "loss": 0.581,
      "mean_token_accuracy": 0.8245093822479248,
      "num_tokens": 95855477.0,
      "step": 601
    },
    {
      "epoch": 0.306205493387589,
      "grad_norm": 1.2335515022277832,
      "learning_rate": 1e-05,
      "loss": 0.6158,
      "mean_token_accuracy": 0.8146917819976807,
      "num_tokens": 96017692.0,
      "step": 602
    },
    {
      "epoch": 0.30671414038657174,
      "grad_norm": 1.223585605621338,
      "learning_rate": 1e-05,
      "loss": 0.573,
      "mean_token_accuracy": 0.8267310857772827,
      "num_tokens": 96175479.0,
      "step": 603
    },
    {
      "epoch": 0.3072227873855544,
      "grad_norm": 1.2164987325668335,
      "learning_rate": 1e-05,
      "loss": 0.5855,
      "mean_token_accuracy": 0.8224807977676392,
      "num_tokens": 96337729.0,
      "step": 604
    },
    {
      "epoch": 0.30773143438453715,
      "grad_norm": 1.1218758821487427,
      "learning_rate": 1e-05,
      "loss": 0.5696,
      "mean_token_accuracy": 0.827846884727478,
      "num_tokens": 96492659.0,
      "step": 605
    },
    {
      "epoch": 0.3082400813835198,
      "grad_norm": 1.2200160026550293,
      "learning_rate": 1e-05,
      "loss": 0.5835,
      "mean_token_accuracy": 0.8219121694564819,
      "num_tokens": 96644965.0,
      "step": 606
    },
    {
      "epoch": 0.30874872838250256,
      "grad_norm": 1.1906354427337646,
      "learning_rate": 1e-05,
      "loss": 0.544,
      "mean_token_accuracy": 0.8338629007339478,
      "num_tokens": 96801670.0,
      "step": 607
    },
    {
      "epoch": 0.30925737538148523,
      "grad_norm": 1.2348442077636719,
      "learning_rate": 1e-05,
      "loss": 0.5751,
      "mean_token_accuracy": 0.8240618705749512,
      "num_tokens": 96966380.0,
      "step": 608
    },
    {
      "epoch": 0.30976602238046796,
      "grad_norm": 1.207759976387024,
      "learning_rate": 1e-05,
      "loss": 0.5759,
      "mean_token_accuracy": 0.8252097368240356,
      "num_tokens": 97116653.0,
      "step": 609
    },
    {
      "epoch": 0.31027466937945064,
      "grad_norm": 1.3898676633834839,
      "learning_rate": 1e-05,
      "loss": 0.5701,
      "mean_token_accuracy": 0.8274800777435303,
      "num_tokens": 97279451.0,
      "step": 610
    },
    {
      "epoch": 0.31078331637843337,
      "grad_norm": 1.2286509275436401,
      "learning_rate": 1e-05,
      "loss": 0.6084,
      "mean_token_accuracy": 0.8180156350135803,
      "num_tokens": 97439569.0,
      "step": 611
    },
    {
      "epoch": 0.31129196337741605,
      "grad_norm": 1.1610603332519531,
      "learning_rate": 1e-05,
      "loss": 0.5625,
      "mean_token_accuracy": 0.8294017910957336,
      "num_tokens": 97595648.0,
      "step": 612
    },
    {
      "epoch": 0.3118006103763988,
      "grad_norm": 1.1723705530166626,
      "learning_rate": 1e-05,
      "loss": 0.602,
      "mean_token_accuracy": 0.8190659284591675,
      "num_tokens": 97754844.0,
      "step": 613
    },
    {
      "epoch": 0.3123092573753815,
      "grad_norm": 1.1151776313781738,
      "learning_rate": 1e-05,
      "loss": 0.5651,
      "mean_token_accuracy": 0.8298476338386536,
      "num_tokens": 97916479.0,
      "step": 614
    },
    {
      "epoch": 0.3128179043743642,
      "grad_norm": 1.2659906148910522,
      "learning_rate": 1e-05,
      "loss": 0.5978,
      "mean_token_accuracy": 0.8206175565719604,
      "num_tokens": 98083229.0,
      "step": 615
    },
    {
      "epoch": 0.3133265513733469,
      "grad_norm": 1.1788604259490967,
      "learning_rate": 1e-05,
      "loss": 0.5743,
      "mean_token_accuracy": 0.8246487379074097,
      "num_tokens": 98234766.0,
      "step": 616
    },
    {
      "epoch": 0.3138351983723296,
      "grad_norm": 1.1908249855041504,
      "learning_rate": 1e-05,
      "loss": 0.564,
      "mean_token_accuracy": 0.8283237218856812,
      "num_tokens": 98401537.0,
      "step": 617
    },
    {
      "epoch": 0.3143438453713123,
      "grad_norm": 1.152085781097412,
      "learning_rate": 1e-05,
      "loss": 0.5548,
      "mean_token_accuracy": 0.8315668702125549,
      "num_tokens": 98567715.0,
      "step": 618
    },
    {
      "epoch": 0.314852492370295,
      "grad_norm": 1.1295671463012695,
      "learning_rate": 1e-05,
      "loss": 0.5742,
      "mean_token_accuracy": 0.8245363235473633,
      "num_tokens": 98735978.0,
      "step": 619
    },
    {
      "epoch": 0.31536113936927773,
      "grad_norm": 1.2672114372253418,
      "learning_rate": 1e-05,
      "loss": 0.6023,
      "mean_token_accuracy": 0.8201767206192017,
      "num_tokens": 98906165.0,
      "step": 620
    },
    {
      "epoch": 0.3158697863682604,
      "grad_norm": 1.220890760421753,
      "learning_rate": 1e-05,
      "loss": 0.5827,
      "mean_token_accuracy": 0.8241899013519287,
      "num_tokens": 99056653.0,
      "step": 621
    },
    {
      "epoch": 0.31637843336724314,
      "grad_norm": 1.1554323434829712,
      "learning_rate": 1e-05,
      "loss": 0.5379,
      "mean_token_accuracy": 0.835025429725647,
      "num_tokens": 99211579.0,
      "step": 622
    },
    {
      "epoch": 0.3168870803662258,
      "grad_norm": 1.2123159170150757,
      "learning_rate": 1e-05,
      "loss": 0.5864,
      "mean_token_accuracy": 0.8231475353240967,
      "num_tokens": 99361670.0,
      "step": 623
    },
    {
      "epoch": 0.31739572736520855,
      "grad_norm": 1.1268131732940674,
      "learning_rate": 1e-05,
      "loss": 0.6015,
      "mean_token_accuracy": 0.8182185888290405,
      "num_tokens": 99530349.0,
      "step": 624
    },
    {
      "epoch": 0.3179043743641913,
      "grad_norm": 1.1105821132659912,
      "learning_rate": 1e-05,
      "loss": 0.5594,
      "mean_token_accuracy": 0.8318790197372437,
      "num_tokens": 99693143.0,
      "step": 625
    },
    {
      "epoch": 0.31841302136317395,
      "grad_norm": 1.0727262496948242,
      "learning_rate": 1e-05,
      "loss": 0.6265,
      "mean_token_accuracy": 0.8118510246276855,
      "num_tokens": 99867681.0,
      "step": 626
    },
    {
      "epoch": 0.3189216683621567,
      "grad_norm": 1.245917558670044,
      "learning_rate": 1e-05,
      "loss": 0.5739,
      "mean_token_accuracy": 0.8248531222343445,
      "num_tokens": 100026848.0,
      "step": 627
    },
    {
      "epoch": 0.31943031536113936,
      "grad_norm": 1.0944061279296875,
      "learning_rate": 1e-05,
      "loss": 0.5631,
      "mean_token_accuracy": 0.8312759399414062,
      "num_tokens": 100198041.0,
      "step": 628
    },
    {
      "epoch": 0.3199389623601221,
      "grad_norm": 1.275220274925232,
      "learning_rate": 1e-05,
      "loss": 0.5433,
      "mean_token_accuracy": 0.8339123725891113,
      "num_tokens": 100352919.0,
      "step": 629
    },
    {
      "epoch": 0.32044760935910477,
      "grad_norm": 1.1218000650405884,
      "learning_rate": 1e-05,
      "loss": 0.5449,
      "mean_token_accuracy": 0.8353885412216187,
      "num_tokens": 100509537.0,
      "step": 630
    },
    {
      "epoch": 0.3209562563580875,
      "grad_norm": 1.2154606580734253,
      "learning_rate": 1e-05,
      "loss": 0.5481,
      "mean_token_accuracy": 0.8327236771583557,
      "num_tokens": 100663059.0,
      "step": 631
    },
    {
      "epoch": 0.3214649033570702,
      "grad_norm": 1.4552714824676514,
      "learning_rate": 1e-05,
      "loss": 0.6006,
      "mean_token_accuracy": 0.8186696767807007,
      "num_tokens": 100831452.0,
      "step": 632
    },
    {
      "epoch": 0.3219735503560529,
      "grad_norm": 1.2636407613754272,
      "learning_rate": 1e-05,
      "loss": 0.5707,
      "mean_token_accuracy": 0.8284827470779419,
      "num_tokens": 100984332.0,
      "step": 633
    },
    {
      "epoch": 0.3224821973550356,
      "grad_norm": 1.3425766229629517,
      "learning_rate": 1e-05,
      "loss": 0.5877,
      "mean_token_accuracy": 0.8216133117675781,
      "num_tokens": 101131963.0,
      "step": 634
    },
    {
      "epoch": 0.3229908443540183,
      "grad_norm": 1.1513028144836426,
      "learning_rate": 1e-05,
      "loss": 0.5675,
      "mean_token_accuracy": 0.8274757862091064,
      "num_tokens": 101289236.0,
      "step": 635
    },
    {
      "epoch": 0.323499491353001,
      "grad_norm": 1.3217164278030396,
      "learning_rate": 1e-05,
      "loss": 0.569,
      "mean_token_accuracy": 0.8281552791595459,
      "num_tokens": 101448301.0,
      "step": 636
    },
    {
      "epoch": 0.3240081383519837,
      "grad_norm": 1.212892770767212,
      "learning_rate": 1e-05,
      "loss": 0.5739,
      "mean_token_accuracy": 0.8250631093978882,
      "num_tokens": 101601568.0,
      "step": 637
    },
    {
      "epoch": 0.32451678535096645,
      "grad_norm": 1.3648805618286133,
      "learning_rate": 1e-05,
      "loss": 0.5734,
      "mean_token_accuracy": 0.824871301651001,
      "num_tokens": 101750506.0,
      "step": 638
    },
    {
      "epoch": 0.32502543234994913,
      "grad_norm": 1.2872990369796753,
      "learning_rate": 1e-05,
      "loss": 0.5893,
      "mean_token_accuracy": 0.8241720199584961,
      "num_tokens": 101915028.0,
      "step": 639
    },
    {
      "epoch": 0.32553407934893186,
      "grad_norm": 1.2380262613296509,
      "learning_rate": 1e-05,
      "loss": 0.5814,
      "mean_token_accuracy": 0.8232290744781494,
      "num_tokens": 102079659.0,
      "step": 640
    },
    {
      "epoch": 0.32604272634791454,
      "grad_norm": 1.1958699226379395,
      "learning_rate": 1e-05,
      "loss": 0.5496,
      "mean_token_accuracy": 0.8321324586868286,
      "num_tokens": 102247764.0,
      "step": 641
    },
    {
      "epoch": 0.32655137334689727,
      "grad_norm": 1.235268473625183,
      "learning_rate": 1e-05,
      "loss": 0.5493,
      "mean_token_accuracy": 0.8314043283462524,
      "num_tokens": 102405734.0,
      "step": 642
    },
    {
      "epoch": 0.32706002034587994,
      "grad_norm": 1.3108789920806885,
      "learning_rate": 1e-05,
      "loss": 0.5702,
      "mean_token_accuracy": 0.8258588314056396,
      "num_tokens": 102566525.0,
      "step": 643
    },
    {
      "epoch": 0.3275686673448627,
      "grad_norm": 1.1007673740386963,
      "learning_rate": 1e-05,
      "loss": 0.6158,
      "mean_token_accuracy": 0.8150623440742493,
      "num_tokens": 102739870.0,
      "step": 644
    },
    {
      "epoch": 0.32807731434384535,
      "grad_norm": 1.3075753450393677,
      "learning_rate": 1e-05,
      "loss": 0.5888,
      "mean_token_accuracy": 0.8222059011459351,
      "num_tokens": 102895661.0,
      "step": 645
    },
    {
      "epoch": 0.3285859613428281,
      "grad_norm": 1.282270073890686,
      "learning_rate": 1e-05,
      "loss": 0.5613,
      "mean_token_accuracy": 0.830429196357727,
      "num_tokens": 103055954.0,
      "step": 646
    },
    {
      "epoch": 0.32909460834181076,
      "grad_norm": 1.1636611223220825,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.8302597999572754,
      "num_tokens": 103214101.0,
      "step": 647
    },
    {
      "epoch": 0.3296032553407935,
      "grad_norm": 1.2605210542678833,
      "learning_rate": 1e-05,
      "loss": 0.5941,
      "mean_token_accuracy": 0.8216217756271362,
      "num_tokens": 103373341.0,
      "step": 648
    },
    {
      "epoch": 0.3301119023397762,
      "grad_norm": 1.1215887069702148,
      "learning_rate": 1e-05,
      "loss": 0.5567,
      "mean_token_accuracy": 0.8300506472587585,
      "num_tokens": 103546489.0,
      "step": 649
    },
    {
      "epoch": 0.3306205493387589,
      "grad_norm": 1.2471184730529785,
      "learning_rate": 1e-05,
      "loss": 0.5582,
      "mean_token_accuracy": 0.8286738991737366,
      "num_tokens": 103709672.0,
      "step": 650
    },
    {
      "epoch": 0.3311291963377416,
      "grad_norm": 1.2193257808685303,
      "learning_rate": 1e-05,
      "loss": 0.5431,
      "mean_token_accuracy": 0.8333966732025146,
      "num_tokens": 103875427.0,
      "step": 651
    },
    {
      "epoch": 0.3316378433367243,
      "grad_norm": 1.2777115106582642,
      "learning_rate": 1e-05,
      "loss": 0.568,
      "mean_token_accuracy": 0.8270902633666992,
      "num_tokens": 104027661.0,
      "step": 652
    },
    {
      "epoch": 0.33214649033570703,
      "grad_norm": 1.2929627895355225,
      "learning_rate": 1e-05,
      "loss": 0.6048,
      "mean_token_accuracy": 0.8180147409439087,
      "num_tokens": 104180735.0,
      "step": 653
    },
    {
      "epoch": 0.3326551373346897,
      "grad_norm": 1.2298574447631836,
      "learning_rate": 1e-05,
      "loss": 0.5642,
      "mean_token_accuracy": 0.8300544023513794,
      "num_tokens": 104353014.0,
      "step": 654
    },
    {
      "epoch": 0.33316378433367244,
      "grad_norm": 1.287480354309082,
      "learning_rate": 1e-05,
      "loss": 0.5738,
      "mean_token_accuracy": 0.8261777758598328,
      "num_tokens": 104517143.0,
      "step": 655
    },
    {
      "epoch": 0.3336724313326551,
      "grad_norm": 1.1733216047286987,
      "learning_rate": 1e-05,
      "loss": 0.6069,
      "mean_token_accuracy": 0.8148963451385498,
      "num_tokens": 104669203.0,
      "step": 656
    },
    {
      "epoch": 0.33418107833163785,
      "grad_norm": 1.364167332649231,
      "learning_rate": 1e-05,
      "loss": 0.5939,
      "mean_token_accuracy": 0.8200170993804932,
      "num_tokens": 104810466.0,
      "step": 657
    },
    {
      "epoch": 0.3346897253306205,
      "grad_norm": 1.1087124347686768,
      "learning_rate": 1e-05,
      "loss": 0.5942,
      "mean_token_accuracy": 0.8199101686477661,
      "num_tokens": 104980367.0,
      "step": 658
    },
    {
      "epoch": 0.33519837232960326,
      "grad_norm": 1.5046215057373047,
      "learning_rate": 1e-05,
      "loss": 0.5528,
      "mean_token_accuracy": 0.8326549530029297,
      "num_tokens": 105131160.0,
      "step": 659
    },
    {
      "epoch": 0.335707019328586,
      "grad_norm": 1.1873219013214111,
      "learning_rate": 1e-05,
      "loss": 0.6084,
      "mean_token_accuracy": 0.818149983882904,
      "num_tokens": 105298008.0,
      "step": 660
    },
    {
      "epoch": 0.33621566632756866,
      "grad_norm": 1.428198218345642,
      "learning_rate": 1e-05,
      "loss": 0.6123,
      "mean_token_accuracy": 0.8152676820755005,
      "num_tokens": 105452271.0,
      "step": 661
    },
    {
      "epoch": 0.3367243133265514,
      "grad_norm": 1.202708125114441,
      "learning_rate": 1e-05,
      "loss": 0.5469,
      "mean_token_accuracy": 0.8334332704544067,
      "num_tokens": 105619978.0,
      "step": 662
    },
    {
      "epoch": 0.33723296032553407,
      "grad_norm": 1.1747825145721436,
      "learning_rate": 1e-05,
      "loss": 0.5598,
      "mean_token_accuracy": 0.829487681388855,
      "num_tokens": 105786355.0,
      "step": 663
    },
    {
      "epoch": 0.3377416073245168,
      "grad_norm": 1.3457754850387573,
      "learning_rate": 1e-05,
      "loss": 0.5984,
      "mean_token_accuracy": 0.819837749004364,
      "num_tokens": 105942503.0,
      "step": 664
    },
    {
      "epoch": 0.3382502543234995,
      "grad_norm": 1.232384204864502,
      "learning_rate": 1e-05,
      "loss": 0.5735,
      "mean_token_accuracy": 0.8262651562690735,
      "num_tokens": 106088220.0,
      "step": 665
    },
    {
      "epoch": 0.3387589013224822,
      "grad_norm": 1.2785651683807373,
      "learning_rate": 1e-05,
      "loss": 0.5332,
      "mean_token_accuracy": 0.8363296985626221,
      "num_tokens": 106246812.0,
      "step": 666
    },
    {
      "epoch": 0.3392675483214649,
      "grad_norm": 1.1574139595031738,
      "learning_rate": 1e-05,
      "loss": 0.5324,
      "mean_token_accuracy": 0.8359112739562988,
      "num_tokens": 106405912.0,
      "step": 667
    },
    {
      "epoch": 0.3397761953204476,
      "grad_norm": 1.167323350906372,
      "learning_rate": 1e-05,
      "loss": 0.6198,
      "mean_token_accuracy": 0.8138099908828735,
      "num_tokens": 106556412.0,
      "step": 668
    },
    {
      "epoch": 0.3402848423194303,
      "grad_norm": 1.261059284210205,
      "learning_rate": 1e-05,
      "loss": 0.5882,
      "mean_token_accuracy": 0.8206124305725098,
      "num_tokens": 106712640.0,
      "step": 669
    },
    {
      "epoch": 0.340793489318413,
      "grad_norm": 1.237960696220398,
      "learning_rate": 1e-05,
      "loss": 0.5775,
      "mean_token_accuracy": 0.8242462873458862,
      "num_tokens": 106869710.0,
      "step": 670
    },
    {
      "epoch": 0.34130213631739575,
      "grad_norm": 1.2517688274383545,
      "learning_rate": 1e-05,
      "loss": 0.5506,
      "mean_token_accuracy": 0.8329342603683472,
      "num_tokens": 107034402.0,
      "step": 671
    },
    {
      "epoch": 0.34181078331637843,
      "grad_norm": 1.1442136764526367,
      "learning_rate": 1e-05,
      "loss": 0.5805,
      "mean_token_accuracy": 0.8242872357368469,
      "num_tokens": 107197492.0,
      "step": 672
    },
    {
      "epoch": 0.34231943031536116,
      "grad_norm": 1.2725573778152466,
      "learning_rate": 1e-05,
      "loss": 0.5624,
      "mean_token_accuracy": 0.8281676769256592,
      "num_tokens": 107355096.0,
      "step": 673
    },
    {
      "epoch": 0.34282807731434384,
      "grad_norm": 1.2090405225753784,
      "learning_rate": 1e-05,
      "loss": 0.5604,
      "mean_token_accuracy": 0.8302109241485596,
      "num_tokens": 107520176.0,
      "step": 674
    },
    {
      "epoch": 0.34333672431332657,
      "grad_norm": 1.1444567441940308,
      "learning_rate": 1e-05,
      "loss": 0.55,
      "mean_token_accuracy": 0.8324422836303711,
      "num_tokens": 107676895.0,
      "step": 675
    },
    {
      "epoch": 0.34384537131230924,
      "grad_norm": 1.2566196918487549,
      "learning_rate": 1e-05,
      "loss": 0.5666,
      "mean_token_accuracy": 0.8281424045562744,
      "num_tokens": 107829848.0,
      "step": 676
    },
    {
      "epoch": 0.344354018311292,
      "grad_norm": 1.143470048904419,
      "learning_rate": 1e-05,
      "loss": 0.6006,
      "mean_token_accuracy": 0.8164473176002502,
      "num_tokens": 108003143.0,
      "step": 677
    },
    {
      "epoch": 0.34486266531027465,
      "grad_norm": 1.2039549350738525,
      "learning_rate": 1e-05,
      "loss": 0.5863,
      "mean_token_accuracy": 0.823419451713562,
      "num_tokens": 108169162.0,
      "step": 678
    },
    {
      "epoch": 0.3453713123092574,
      "grad_norm": 1.162801742553711,
      "learning_rate": 1e-05,
      "loss": 0.548,
      "mean_token_accuracy": 0.8335537910461426,
      "num_tokens": 108326285.0,
      "step": 679
    },
    {
      "epoch": 0.34587995930824006,
      "grad_norm": 1.1526139974594116,
      "learning_rate": 1e-05,
      "loss": 0.5487,
      "mean_token_accuracy": 0.8326566219329834,
      "num_tokens": 108481647.0,
      "step": 680
    },
    {
      "epoch": 0.3463886063072228,
      "grad_norm": 1.1998027563095093,
      "learning_rate": 1e-05,
      "loss": 0.5556,
      "mean_token_accuracy": 0.8318803906440735,
      "num_tokens": 108632219.0,
      "step": 681
    },
    {
      "epoch": 0.34689725330620547,
      "grad_norm": 1.1241905689239502,
      "learning_rate": 1e-05,
      "loss": 0.5693,
      "mean_token_accuracy": 0.8260743618011475,
      "num_tokens": 108795594.0,
      "step": 682
    },
    {
      "epoch": 0.3474059003051882,
      "grad_norm": 1.1517409086227417,
      "learning_rate": 1e-05,
      "loss": 0.5577,
      "mean_token_accuracy": 0.8296923637390137,
      "num_tokens": 108947589.0,
      "step": 683
    },
    {
      "epoch": 0.34791454730417093,
      "grad_norm": 1.1438885927200317,
      "learning_rate": 1e-05,
      "loss": 0.5731,
      "mean_token_accuracy": 0.8259981274604797,
      "num_tokens": 109100848.0,
      "step": 684
    },
    {
      "epoch": 0.3484231943031536,
      "grad_norm": 1.2787206172943115,
      "learning_rate": 1e-05,
      "loss": 0.6061,
      "mean_token_accuracy": 0.8187931180000305,
      "num_tokens": 109268616.0,
      "step": 685
    },
    {
      "epoch": 0.34893184130213634,
      "grad_norm": 1.1225264072418213,
      "learning_rate": 1e-05,
      "loss": 0.5434,
      "mean_token_accuracy": 0.833125650882721,
      "num_tokens": 109425497.0,
      "step": 686
    },
    {
      "epoch": 0.349440488301119,
      "grad_norm": 1.3023948669433594,
      "learning_rate": 1e-05,
      "loss": 0.5646,
      "mean_token_accuracy": 0.82758629322052,
      "num_tokens": 109596743.0,
      "step": 687
    },
    {
      "epoch": 0.34994913530010174,
      "grad_norm": 1.1158123016357422,
      "learning_rate": 1e-05,
      "loss": 0.54,
      "mean_token_accuracy": 0.8356212377548218,
      "num_tokens": 109752861.0,
      "step": 688
    },
    {
      "epoch": 0.3504577822990844,
      "grad_norm": 1.1276473999023438,
      "learning_rate": 1e-05,
      "loss": 0.5434,
      "mean_token_accuracy": 0.8341058492660522,
      "num_tokens": 109912156.0,
      "step": 689
    },
    {
      "epoch": 0.35096642929806715,
      "grad_norm": 1.326568365097046,
      "learning_rate": 1e-05,
      "loss": 0.5503,
      "mean_token_accuracy": 0.8320469260215759,
      "num_tokens": 110075837.0,
      "step": 690
    },
    {
      "epoch": 0.3514750762970498,
      "grad_norm": 1.1527599096298218,
      "learning_rate": 1e-05,
      "loss": 0.587,
      "mean_token_accuracy": 0.8230042457580566,
      "num_tokens": 110230103.0,
      "step": 691
    },
    {
      "epoch": 0.35198372329603256,
      "grad_norm": 1.2510868310928345,
      "learning_rate": 1e-05,
      "loss": 0.5628,
      "mean_token_accuracy": 0.8295976519584656,
      "num_tokens": 110388178.0,
      "step": 692
    },
    {
      "epoch": 0.35249237029501523,
      "grad_norm": 1.2448580265045166,
      "learning_rate": 1e-05,
      "loss": 0.575,
      "mean_token_accuracy": 0.8250375390052795,
      "num_tokens": 110556832.0,
      "step": 693
    },
    {
      "epoch": 0.35300101729399797,
      "grad_norm": 1.3243716955184937,
      "learning_rate": 1e-05,
      "loss": 0.5814,
      "mean_token_accuracy": 0.8233970403671265,
      "num_tokens": 110712545.0,
      "step": 694
    },
    {
      "epoch": 0.3535096642929807,
      "grad_norm": 1.2904678583145142,
      "learning_rate": 1e-05,
      "loss": 0.547,
      "mean_token_accuracy": 0.8315588235855103,
      "num_tokens": 110855683.0,
      "step": 695
    },
    {
      "epoch": 0.3540183112919634,
      "grad_norm": 1.129303216934204,
      "learning_rate": 1e-05,
      "loss": 0.5724,
      "mean_token_accuracy": 0.8251450657844543,
      "num_tokens": 111014148.0,
      "step": 696
    },
    {
      "epoch": 0.3545269582909461,
      "grad_norm": 1.3024146556854248,
      "learning_rate": 1e-05,
      "loss": 0.5699,
      "mean_token_accuracy": 0.8282129764556885,
      "num_tokens": 111168794.0,
      "step": 697
    },
    {
      "epoch": 0.3550356052899288,
      "grad_norm": 1.2099299430847168,
      "learning_rate": 1e-05,
      "loss": 0.5728,
      "mean_token_accuracy": 0.8253839612007141,
      "num_tokens": 111318802.0,
      "step": 698
    },
    {
      "epoch": 0.3555442522889115,
      "grad_norm": 1.365613579750061,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.831680417060852,
      "num_tokens": 111482438.0,
      "step": 699
    },
    {
      "epoch": 0.3560528992878942,
      "grad_norm": 1.1956545114517212,
      "learning_rate": 1e-05,
      "loss": 0.5713,
      "mean_token_accuracy": 0.8248770833015442,
      "num_tokens": 111631108.0,
      "step": 700
    },
    {
      "epoch": 0.3565615462868769,
      "grad_norm": 1.1379332542419434,
      "learning_rate": 1e-05,
      "loss": 0.5813,
      "mean_token_accuracy": 0.8245474696159363,
      "num_tokens": 111792659.0,
      "step": 701
    },
    {
      "epoch": 0.3570701932858596,
      "grad_norm": 1.311396837234497,
      "learning_rate": 1e-05,
      "loss": 0.575,
      "mean_token_accuracy": 0.8252518177032471,
      "num_tokens": 111956492.0,
      "step": 702
    },
    {
      "epoch": 0.3575788402848423,
      "grad_norm": 1.1396161317825317,
      "learning_rate": 1e-05,
      "loss": 0.5696,
      "mean_token_accuracy": 0.82741379737854,
      "num_tokens": 112111925.0,
      "step": 703
    },
    {
      "epoch": 0.358087487283825,
      "grad_norm": 1.2385365962982178,
      "learning_rate": 1e-05,
      "loss": 0.5718,
      "mean_token_accuracy": 0.8264093399047852,
      "num_tokens": 112271716.0,
      "step": 704
    },
    {
      "epoch": 0.35859613428280773,
      "grad_norm": 1.1056485176086426,
      "learning_rate": 1e-05,
      "loss": 0.5399,
      "mean_token_accuracy": 0.833939790725708,
      "num_tokens": 112433508.0,
      "step": 705
    },
    {
      "epoch": 0.35910478128179046,
      "grad_norm": 1.142451524734497,
      "learning_rate": 1e-05,
      "loss": 0.5366,
      "mean_token_accuracy": 0.8368534445762634,
      "num_tokens": 112595775.0,
      "step": 706
    },
    {
      "epoch": 0.35961342828077314,
      "grad_norm": 1.3010401725769043,
      "learning_rate": 1e-05,
      "loss": 0.5498,
      "mean_token_accuracy": 0.8305310010910034,
      "num_tokens": 112747723.0,
      "step": 707
    },
    {
      "epoch": 0.36012207527975587,
      "grad_norm": 1.1268346309661865,
      "learning_rate": 1e-05,
      "loss": 0.5627,
      "mean_token_accuracy": 0.828546404838562,
      "num_tokens": 112906425.0,
      "step": 708
    },
    {
      "epoch": 0.36063072227873855,
      "grad_norm": 1.151298999786377,
      "learning_rate": 1e-05,
      "loss": 0.5531,
      "mean_token_accuracy": 0.8322466015815735,
      "num_tokens": 113062355.0,
      "step": 709
    },
    {
      "epoch": 0.3611393692777213,
      "grad_norm": 1.1866037845611572,
      "learning_rate": 1e-05,
      "loss": 0.5839,
      "mean_token_accuracy": 0.8216893672943115,
      "num_tokens": 113237249.0,
      "step": 710
    },
    {
      "epoch": 0.36164801627670395,
      "grad_norm": 1.164064884185791,
      "learning_rate": 1e-05,
      "loss": 0.5826,
      "mean_token_accuracy": 0.8220279216766357,
      "num_tokens": 113399443.0,
      "step": 711
    },
    {
      "epoch": 0.3621566632756867,
      "grad_norm": 1.190131425857544,
      "learning_rate": 1e-05,
      "loss": 0.5407,
      "mean_token_accuracy": 0.8338578343391418,
      "num_tokens": 113555518.0,
      "step": 712
    },
    {
      "epoch": 0.36266531027466936,
      "grad_norm": 1.2120224237442017,
      "learning_rate": 1e-05,
      "loss": 0.5716,
      "mean_token_accuracy": 0.8261467218399048,
      "num_tokens": 113698492.0,
      "step": 713
    },
    {
      "epoch": 0.3631739572736521,
      "grad_norm": 1.243842363357544,
      "learning_rate": 1e-05,
      "loss": 0.6163,
      "mean_token_accuracy": 0.8146119713783264,
      "num_tokens": 113869795.0,
      "step": 714
    },
    {
      "epoch": 0.36368260427263477,
      "grad_norm": 1.1908296346664429,
      "learning_rate": 1e-05,
      "loss": 0.5441,
      "mean_token_accuracy": 0.833182692527771,
      "num_tokens": 114035400.0,
      "step": 715
    },
    {
      "epoch": 0.3641912512716175,
      "grad_norm": 1.2101727724075317,
      "learning_rate": 1e-05,
      "loss": 0.5856,
      "mean_token_accuracy": 0.8229519128799438,
      "num_tokens": 114180922.0,
      "step": 716
    },
    {
      "epoch": 0.3646998982706002,
      "grad_norm": 1.2070703506469727,
      "learning_rate": 1e-05,
      "loss": 0.5583,
      "mean_token_accuracy": 0.8307386636734009,
      "num_tokens": 114351814.0,
      "step": 717
    },
    {
      "epoch": 0.3652085452695829,
      "grad_norm": 1.186836838722229,
      "learning_rate": 1e-05,
      "loss": 0.5417,
      "mean_token_accuracy": 0.832840085029602,
      "num_tokens": 114503125.0,
      "step": 718
    },
    {
      "epoch": 0.36571719226856564,
      "grad_norm": 1.173302173614502,
      "learning_rate": 1e-05,
      "loss": 0.5824,
      "mean_token_accuracy": 0.8242769241333008,
      "num_tokens": 114658004.0,
      "step": 719
    },
    {
      "epoch": 0.3662258392675483,
      "grad_norm": 1.3086676597595215,
      "learning_rate": 1e-05,
      "loss": 0.5966,
      "mean_token_accuracy": 0.8192692995071411,
      "num_tokens": 114829230.0,
      "step": 720
    },
    {
      "epoch": 0.36673448626653105,
      "grad_norm": 1.2513558864593506,
      "learning_rate": 1e-05,
      "loss": 0.5651,
      "mean_token_accuracy": 0.8275444507598877,
      "num_tokens": 114987847.0,
      "step": 721
    },
    {
      "epoch": 0.3672431332655137,
      "grad_norm": 1.1308586597442627,
      "learning_rate": 1e-05,
      "loss": 0.5543,
      "mean_token_accuracy": 0.8325884938240051,
      "num_tokens": 115154558.0,
      "step": 722
    },
    {
      "epoch": 0.36775178026449645,
      "grad_norm": 1.2177181243896484,
      "learning_rate": 1e-05,
      "loss": 0.5585,
      "mean_token_accuracy": 0.8294394612312317,
      "num_tokens": 115315702.0,
      "step": 723
    },
    {
      "epoch": 0.36826042726347913,
      "grad_norm": 1.2306625843048096,
      "learning_rate": 1e-05,
      "loss": 0.5576,
      "mean_token_accuracy": 0.8295246958732605,
      "num_tokens": 115471150.0,
      "step": 724
    },
    {
      "epoch": 0.36876907426246186,
      "grad_norm": 1.1928412914276123,
      "learning_rate": 1e-05,
      "loss": 0.5782,
      "mean_token_accuracy": 0.8248293399810791,
      "num_tokens": 115630359.0,
      "step": 725
    },
    {
      "epoch": 0.36927772126144454,
      "grad_norm": 1.2332453727722168,
      "learning_rate": 1e-05,
      "loss": 0.5759,
      "mean_token_accuracy": 0.8268725872039795,
      "num_tokens": 115802068.0,
      "step": 726
    },
    {
      "epoch": 0.36978636826042727,
      "grad_norm": 1.221351981163025,
      "learning_rate": 1e-05,
      "loss": 0.5523,
      "mean_token_accuracy": 0.831588625907898,
      "num_tokens": 115954597.0,
      "step": 727
    },
    {
      "epoch": 0.37029501525940994,
      "grad_norm": 1.178653359413147,
      "learning_rate": 1e-05,
      "loss": 0.5899,
      "mean_token_accuracy": 0.8207197189331055,
      "num_tokens": 116121996.0,
      "step": 728
    },
    {
      "epoch": 0.3708036622583927,
      "grad_norm": 1.2622578144073486,
      "learning_rate": 1e-05,
      "loss": 0.5705,
      "mean_token_accuracy": 0.8262423276901245,
      "num_tokens": 116286650.0,
      "step": 729
    },
    {
      "epoch": 0.3713123092573754,
      "grad_norm": 1.219468116760254,
      "learning_rate": 1e-05,
      "loss": 0.551,
      "mean_token_accuracy": 0.8329716324806213,
      "num_tokens": 116456648.0,
      "step": 730
    },
    {
      "epoch": 0.3718209562563581,
      "grad_norm": 1.1355637311935425,
      "learning_rate": 1e-05,
      "loss": 0.5541,
      "mean_token_accuracy": 0.8318668603897095,
      "num_tokens": 116621931.0,
      "step": 731
    },
    {
      "epoch": 0.3723296032553408,
      "grad_norm": 1.310593605041504,
      "learning_rate": 1e-05,
      "loss": 0.6152,
      "mean_token_accuracy": 0.8147826790809631,
      "num_tokens": 116787232.0,
      "step": 732
    },
    {
      "epoch": 0.3728382502543235,
      "grad_norm": 1.2487748861312866,
      "learning_rate": 1e-05,
      "loss": 0.6179,
      "mean_token_accuracy": 0.8136881589889526,
      "num_tokens": 116949019.0,
      "step": 733
    },
    {
      "epoch": 0.3733468972533062,
      "grad_norm": 1.1336181163787842,
      "learning_rate": 1e-05,
      "loss": 0.5902,
      "mean_token_accuracy": 0.8222196102142334,
      "num_tokens": 117122455.0,
      "step": 734
    },
    {
      "epoch": 0.3738555442522889,
      "grad_norm": 1.2856602668762207,
      "learning_rate": 1e-05,
      "loss": 0.5639,
      "mean_token_accuracy": 0.828788161277771,
      "num_tokens": 117267738.0,
      "step": 735
    },
    {
      "epoch": 0.3743641912512716,
      "grad_norm": 1.1545137166976929,
      "learning_rate": 1e-05,
      "loss": 0.5654,
      "mean_token_accuracy": 0.8256614208221436,
      "num_tokens": 117430802.0,
      "step": 736
    },
    {
      "epoch": 0.3748728382502543,
      "grad_norm": 1.143939733505249,
      "learning_rate": 1e-05,
      "loss": 0.5776,
      "mean_token_accuracy": 0.8231573104858398,
      "num_tokens": 117587032.0,
      "step": 737
    },
    {
      "epoch": 0.37538148524923703,
      "grad_norm": 1.060882568359375,
      "learning_rate": 1e-05,
      "loss": 0.5871,
      "mean_token_accuracy": 0.8218531608581543,
      "num_tokens": 117747701.0,
      "step": 738
    },
    {
      "epoch": 0.3758901322482197,
      "grad_norm": 1.1606186628341675,
      "learning_rate": 1e-05,
      "loss": 0.5624,
      "mean_token_accuracy": 0.8296399116516113,
      "num_tokens": 117912178.0,
      "step": 739
    },
    {
      "epoch": 0.37639877924720244,
      "grad_norm": 1.1886235475540161,
      "learning_rate": 1e-05,
      "loss": 0.5331,
      "mean_token_accuracy": 0.8355053663253784,
      "num_tokens": 118069631.0,
      "step": 740
    },
    {
      "epoch": 0.3769074262461852,
      "grad_norm": 1.2048909664154053,
      "learning_rate": 1e-05,
      "loss": 0.5472,
      "mean_token_accuracy": 0.832554817199707,
      "num_tokens": 118216192.0,
      "step": 741
    },
    {
      "epoch": 0.37741607324516785,
      "grad_norm": 1.2597253322601318,
      "learning_rate": 1e-05,
      "loss": 0.5484,
      "mean_token_accuracy": 0.8324501514434814,
      "num_tokens": 118364107.0,
      "step": 742
    },
    {
      "epoch": 0.3779247202441506,
      "grad_norm": 1.1237263679504395,
      "learning_rate": 1e-05,
      "loss": 0.5928,
      "mean_token_accuracy": 0.8186395168304443,
      "num_tokens": 118532657.0,
      "step": 743
    },
    {
      "epoch": 0.37843336724313326,
      "grad_norm": 1.2016260623931885,
      "learning_rate": 1e-05,
      "loss": 0.5557,
      "mean_token_accuracy": 0.8294884562492371,
      "num_tokens": 118689942.0,
      "step": 744
    },
    {
      "epoch": 0.378942014242116,
      "grad_norm": 1.329368233680725,
      "learning_rate": 1e-05,
      "loss": 0.5555,
      "mean_token_accuracy": 0.8295120596885681,
      "num_tokens": 118861717.0,
      "step": 745
    },
    {
      "epoch": 0.37945066124109866,
      "grad_norm": 1.2445459365844727,
      "learning_rate": 1e-05,
      "loss": 0.568,
      "mean_token_accuracy": 0.8270270228385925,
      "num_tokens": 119027905.0,
      "step": 746
    },
    {
      "epoch": 0.3799593082400814,
      "grad_norm": 1.2161781787872314,
      "learning_rate": 1e-05,
      "loss": 0.5114,
      "mean_token_accuracy": 0.8412526845932007,
      "num_tokens": 119173593.0,
      "step": 747
    },
    {
      "epoch": 0.38046795523906407,
      "grad_norm": 1.3197194337844849,
      "learning_rate": 1e-05,
      "loss": 0.6057,
      "mean_token_accuracy": 0.8171380162239075,
      "num_tokens": 119333549.0,
      "step": 748
    },
    {
      "epoch": 0.3809766022380468,
      "grad_norm": 1.2615206241607666,
      "learning_rate": 1e-05,
      "loss": 0.552,
      "mean_token_accuracy": 0.8305937647819519,
      "num_tokens": 119485196.0,
      "step": 749
    },
    {
      "epoch": 0.3814852492370295,
      "grad_norm": 1.241072416305542,
      "learning_rate": 1e-05,
      "loss": 0.5502,
      "mean_token_accuracy": 0.8316583633422852,
      "num_tokens": 119635566.0,
      "step": 750
    },
    {
      "epoch": 0.3819938962360122,
      "grad_norm": 1.3704606294631958,
      "learning_rate": 1e-05,
      "loss": 0.5569,
      "mean_token_accuracy": 0.8296416997909546,
      "num_tokens": 119796534.0,
      "step": 751
    },
    {
      "epoch": 0.38250254323499494,
      "grad_norm": 1.2328382730484009,
      "learning_rate": 1e-05,
      "loss": 0.6011,
      "mean_token_accuracy": 0.8162603378295898,
      "num_tokens": 119965071.0,
      "step": 752
    },
    {
      "epoch": 0.3830111902339776,
      "grad_norm": 1.320970892906189,
      "learning_rate": 1e-05,
      "loss": 0.5441,
      "mean_token_accuracy": 0.8339331746101379,
      "num_tokens": 120116111.0,
      "step": 753
    },
    {
      "epoch": 0.38351983723296035,
      "grad_norm": 1.234978199005127,
      "learning_rate": 1e-05,
      "loss": 0.585,
      "mean_token_accuracy": 0.8218483924865723,
      "num_tokens": 120275408.0,
      "step": 754
    },
    {
      "epoch": 0.384028484231943,
      "grad_norm": 1.2488471269607544,
      "learning_rate": 1e-05,
      "loss": 0.5602,
      "mean_token_accuracy": 0.8297809362411499,
      "num_tokens": 120432203.0,
      "step": 755
    },
    {
      "epoch": 0.38453713123092575,
      "grad_norm": 1.2408305406570435,
      "learning_rate": 1e-05,
      "loss": 0.5966,
      "mean_token_accuracy": 0.8207588791847229,
      "num_tokens": 120600556.0,
      "step": 756
    },
    {
      "epoch": 0.38504577822990843,
      "grad_norm": 1.1832354068756104,
      "learning_rate": 1e-05,
      "loss": 0.5647,
      "mean_token_accuracy": 0.8281243443489075,
      "num_tokens": 120758883.0,
      "step": 757
    },
    {
      "epoch": 0.38555442522889116,
      "grad_norm": 1.1637436151504517,
      "learning_rate": 1e-05,
      "loss": 0.5468,
      "mean_token_accuracy": 0.8347381949424744,
      "num_tokens": 120922100.0,
      "step": 758
    },
    {
      "epoch": 0.38606307222787384,
      "grad_norm": 1.2801555395126343,
      "learning_rate": 1e-05,
      "loss": 0.5651,
      "mean_token_accuracy": 0.8274706602096558,
      "num_tokens": 121075897.0,
      "step": 759
    },
    {
      "epoch": 0.38657171922685657,
      "grad_norm": 1.390296459197998,
      "learning_rate": 1e-05,
      "loss": 0.551,
      "mean_token_accuracy": 0.8321919441223145,
      "num_tokens": 121235661.0,
      "step": 760
    },
    {
      "epoch": 0.38708036622583925,
      "grad_norm": 1.1444047689437866,
      "learning_rate": 1e-05,
      "loss": 0.5786,
      "mean_token_accuracy": 0.8248952627182007,
      "num_tokens": 121393640.0,
      "step": 761
    },
    {
      "epoch": 0.387589013224822,
      "grad_norm": 1.3989663124084473,
      "learning_rate": 1e-05,
      "loss": 0.5751,
      "mean_token_accuracy": 0.8251594305038452,
      "num_tokens": 121539889.0,
      "step": 762
    },
    {
      "epoch": 0.38809766022380465,
      "grad_norm": 1.1752854585647583,
      "learning_rate": 1e-05,
      "loss": 0.5567,
      "mean_token_accuracy": 0.8292431235313416,
      "num_tokens": 121699739.0,
      "step": 763
    },
    {
      "epoch": 0.3886063072227874,
      "grad_norm": 1.4760440587997437,
      "learning_rate": 1e-05,
      "loss": 0.5179,
      "mean_token_accuracy": 0.8403042554855347,
      "num_tokens": 121844594.0,
      "step": 764
    },
    {
      "epoch": 0.3891149542217701,
      "grad_norm": 1.1970382928848267,
      "learning_rate": 1e-05,
      "loss": 0.5733,
      "mean_token_accuracy": 0.8265695571899414,
      "num_tokens": 122018838.0,
      "step": 765
    },
    {
      "epoch": 0.3896236012207528,
      "grad_norm": 1.3410618305206299,
      "learning_rate": 1e-05,
      "loss": 0.5922,
      "mean_token_accuracy": 0.8223105669021606,
      "num_tokens": 122181626.0,
      "step": 766
    },
    {
      "epoch": 0.3901322482197355,
      "grad_norm": 1.3157473802566528,
      "learning_rate": 1e-05,
      "loss": 0.5569,
      "mean_token_accuracy": 0.8304774761199951,
      "num_tokens": 122336319.0,
      "step": 767
    },
    {
      "epoch": 0.3906408952187182,
      "grad_norm": 1.3343852758407593,
      "learning_rate": 1e-05,
      "loss": 0.5262,
      "mean_token_accuracy": 0.8391714096069336,
      "num_tokens": 122486556.0,
      "step": 768
    },
    {
      "epoch": 0.39114954221770093,
      "grad_norm": 1.2602474689483643,
      "learning_rate": 1e-05,
      "loss": 0.5704,
      "mean_token_accuracy": 0.8253237009048462,
      "num_tokens": 122643871.0,
      "step": 769
    },
    {
      "epoch": 0.3916581892166836,
      "grad_norm": 1.2566087245941162,
      "learning_rate": 1e-05,
      "loss": 0.5638,
      "mean_token_accuracy": 0.8279736042022705,
      "num_tokens": 122806517.0,
      "step": 770
    },
    {
      "epoch": 0.39216683621566634,
      "grad_norm": 1.23412024974823,
      "learning_rate": 1e-05,
      "loss": 0.56,
      "mean_token_accuracy": 0.8286000490188599,
      "num_tokens": 122963330.0,
      "step": 771
    },
    {
      "epoch": 0.392675483214649,
      "grad_norm": 1.1621471643447876,
      "learning_rate": 1e-05,
      "loss": 0.5703,
      "mean_token_accuracy": 0.8259589076042175,
      "num_tokens": 123130884.0,
      "step": 772
    },
    {
      "epoch": 0.39318413021363174,
      "grad_norm": 1.1377054452896118,
      "learning_rate": 1e-05,
      "loss": 0.5535,
      "mean_token_accuracy": 0.8336734175682068,
      "num_tokens": 123291806.0,
      "step": 773
    },
    {
      "epoch": 0.3936927772126144,
      "grad_norm": 1.12781822681427,
      "learning_rate": 1e-05,
      "loss": 0.51,
      "mean_token_accuracy": 0.8437920212745667,
      "num_tokens": 123442859.0,
      "step": 774
    },
    {
      "epoch": 0.39420142421159715,
      "grad_norm": 1.2181215286254883,
      "learning_rate": 1e-05,
      "loss": 0.5546,
      "mean_token_accuracy": 0.8290634155273438,
      "num_tokens": 123592681.0,
      "step": 775
    },
    {
      "epoch": 0.3947100712105799,
      "grad_norm": 1.3110220432281494,
      "learning_rate": 1e-05,
      "loss": 0.5395,
      "mean_token_accuracy": 0.8334207534790039,
      "num_tokens": 123764752.0,
      "step": 776
    },
    {
      "epoch": 0.39521871820956256,
      "grad_norm": 1.2244796752929688,
      "learning_rate": 1e-05,
      "loss": 0.5725,
      "mean_token_accuracy": 0.8263367414474487,
      "num_tokens": 123922538.0,
      "step": 777
    },
    {
      "epoch": 0.3957273652085453,
      "grad_norm": 1.2441754341125488,
      "learning_rate": 1e-05,
      "loss": 0.543,
      "mean_token_accuracy": 0.8332741260528564,
      "num_tokens": 124083040.0,
      "step": 778
    },
    {
      "epoch": 0.39623601220752797,
      "grad_norm": 1.3164564371109009,
      "learning_rate": 1e-05,
      "loss": 0.5608,
      "mean_token_accuracy": 0.8291635513305664,
      "num_tokens": 124239220.0,
      "step": 779
    },
    {
      "epoch": 0.3967446592065107,
      "grad_norm": 1.1562912464141846,
      "learning_rate": 1e-05,
      "loss": 0.5583,
      "mean_token_accuracy": 0.8303772807121277,
      "num_tokens": 124398859.0,
      "step": 780
    },
    {
      "epoch": 0.3972533062054934,
      "grad_norm": 1.2362325191497803,
      "learning_rate": 1e-05,
      "loss": 0.582,
      "mean_token_accuracy": 0.8240430951118469,
      "num_tokens": 124549761.0,
      "step": 781
    },
    {
      "epoch": 0.3977619532044761,
      "grad_norm": 1.1560744047164917,
      "learning_rate": 1e-05,
      "loss": 0.5853,
      "mean_token_accuracy": 0.8228815793991089,
      "num_tokens": 124705300.0,
      "step": 782
    },
    {
      "epoch": 0.3982706002034588,
      "grad_norm": 1.2574455738067627,
      "learning_rate": 1e-05,
      "loss": 0.5631,
      "mean_token_accuracy": 0.8280406594276428,
      "num_tokens": 124863824.0,
      "step": 783
    },
    {
      "epoch": 0.3987792472024415,
      "grad_norm": 1.2234323024749756,
      "learning_rate": 1e-05,
      "loss": 0.5901,
      "mean_token_accuracy": 0.8210317492485046,
      "num_tokens": 125017794.0,
      "step": 784
    },
    {
      "epoch": 0.3992878942014242,
      "grad_norm": 1.2508187294006348,
      "learning_rate": 1e-05,
      "loss": 0.5789,
      "mean_token_accuracy": 0.8266098499298096,
      "num_tokens": 125178609.0,
      "step": 785
    },
    {
      "epoch": 0.3997965412004069,
      "grad_norm": 1.2754849195480347,
      "learning_rate": 1e-05,
      "loss": 0.5759,
      "mean_token_accuracy": 0.8228343725204468,
      "num_tokens": 125331610.0,
      "step": 786
    },
    {
      "epoch": 0.40030518819938965,
      "grad_norm": 1.1867464780807495,
      "learning_rate": 1e-05,
      "loss": 0.5324,
      "mean_token_accuracy": 0.8351470232009888,
      "num_tokens": 125481250.0,
      "step": 787
    },
    {
      "epoch": 0.4008138351983723,
      "grad_norm": 1.333287239074707,
      "learning_rate": 1e-05,
      "loss": 0.5767,
      "mean_token_accuracy": 0.825285017490387,
      "num_tokens": 125650464.0,
      "step": 788
    },
    {
      "epoch": 0.40132248219735506,
      "grad_norm": 1.4091664552688599,
      "learning_rate": 1e-05,
      "loss": 0.5532,
      "mean_token_accuracy": 0.8286569118499756,
      "num_tokens": 125817057.0,
      "step": 789
    },
    {
      "epoch": 0.40183112919633773,
      "grad_norm": 1.0249348878860474,
      "learning_rate": 1e-05,
      "loss": 0.5409,
      "mean_token_accuracy": 0.833737850189209,
      "num_tokens": 125964354.0,
      "step": 790
    },
    {
      "epoch": 0.40233977619532046,
      "grad_norm": 1.2201231718063354,
      "learning_rate": 1e-05,
      "loss": 0.5382,
      "mean_token_accuracy": 0.8345334529876709,
      "num_tokens": 126119274.0,
      "step": 791
    },
    {
      "epoch": 0.40284842319430314,
      "grad_norm": 1.3263883590698242,
      "learning_rate": 1e-05,
      "loss": 0.5606,
      "mean_token_accuracy": 0.8295636773109436,
      "num_tokens": 126266332.0,
      "step": 792
    },
    {
      "epoch": 0.40335707019328587,
      "grad_norm": 1.1370054483413696,
      "learning_rate": 1e-05,
      "loss": 0.5671,
      "mean_token_accuracy": 0.8283527493476868,
      "num_tokens": 126416815.0,
      "step": 793
    },
    {
      "epoch": 0.40386571719226855,
      "grad_norm": 1.232959508895874,
      "learning_rate": 1e-05,
      "loss": 0.5465,
      "mean_token_accuracy": 0.8347012400627136,
      "num_tokens": 126582617.0,
      "step": 794
    },
    {
      "epoch": 0.4043743641912513,
      "grad_norm": 1.158482551574707,
      "learning_rate": 1e-05,
      "loss": 0.5592,
      "mean_token_accuracy": 0.8307364583015442,
      "num_tokens": 126734456.0,
      "step": 795
    },
    {
      "epoch": 0.40488301119023395,
      "grad_norm": 1.0260847806930542,
      "learning_rate": 1e-05,
      "loss": 0.5504,
      "mean_token_accuracy": 0.832145631313324,
      "num_tokens": 126894951.0,
      "step": 796
    },
    {
      "epoch": 0.4053916581892167,
      "grad_norm": 1.1842215061187744,
      "learning_rate": 1e-05,
      "loss": 0.5751,
      "mean_token_accuracy": 0.8247040510177612,
      "num_tokens": 127040903.0,
      "step": 797
    },
    {
      "epoch": 0.4059003051881994,
      "grad_norm": 1.2335243225097656,
      "learning_rate": 1e-05,
      "loss": 0.5908,
      "mean_token_accuracy": 0.8203405737876892,
      "num_tokens": 127193010.0,
      "step": 798
    },
    {
      "epoch": 0.4064089521871821,
      "grad_norm": 1.0869059562683105,
      "learning_rate": 1e-05,
      "loss": 0.5698,
      "mean_token_accuracy": 0.8263587951660156,
      "num_tokens": 127356124.0,
      "step": 799
    },
    {
      "epoch": 0.4069175991861648,
      "grad_norm": 1.0540488958358765,
      "learning_rate": 1e-05,
      "loss": 0.5488,
      "mean_token_accuracy": 0.8332864046096802,
      "num_tokens": 127530354.0,
      "step": 800
    },
    {
      "epoch": 0.4074262461851475,
      "grad_norm": 1.14442777633667,
      "learning_rate": 1e-05,
      "loss": 0.556,
      "mean_token_accuracy": 0.8312243223190308,
      "num_tokens": 127674617.0,
      "step": 801
    },
    {
      "epoch": 0.40793489318413023,
      "grad_norm": 1.1911969184875488,
      "learning_rate": 1e-05,
      "loss": 0.5508,
      "mean_token_accuracy": 0.8314390182495117,
      "num_tokens": 127833308.0,
      "step": 802
    },
    {
      "epoch": 0.4084435401831129,
      "grad_norm": 1.2103899717330933,
      "learning_rate": 1e-05,
      "loss": 0.5811,
      "mean_token_accuracy": 0.8233762979507446,
      "num_tokens": 127994496.0,
      "step": 803
    },
    {
      "epoch": 0.40895218718209564,
      "grad_norm": 1.246382713317871,
      "learning_rate": 1e-05,
      "loss": 0.5577,
      "mean_token_accuracy": 0.8302004337310791,
      "num_tokens": 128161302.0,
      "step": 804
    },
    {
      "epoch": 0.4094608341810783,
      "grad_norm": 1.2232249975204468,
      "learning_rate": 1e-05,
      "loss": 0.5679,
      "mean_token_accuracy": 0.8264790773391724,
      "num_tokens": 128318321.0,
      "step": 805
    },
    {
      "epoch": 0.40996948118006105,
      "grad_norm": 1.2698410749435425,
      "learning_rate": 1e-05,
      "loss": 0.5626,
      "mean_token_accuracy": 0.8275660872459412,
      "num_tokens": 128485010.0,
      "step": 806
    },
    {
      "epoch": 0.4104781281790437,
      "grad_norm": 1.1866873502731323,
      "learning_rate": 1e-05,
      "loss": 0.5847,
      "mean_token_accuracy": 0.8216865658760071,
      "num_tokens": 128651307.0,
      "step": 807
    },
    {
      "epoch": 0.41098677517802645,
      "grad_norm": 1.2357579469680786,
      "learning_rate": 1e-05,
      "loss": 0.5459,
      "mean_token_accuracy": 0.8317310810089111,
      "num_tokens": 128794807.0,
      "step": 808
    },
    {
      "epoch": 0.41149542217700913,
      "grad_norm": 1.2447888851165771,
      "learning_rate": 1e-05,
      "loss": 0.5525,
      "mean_token_accuracy": 0.8310381770133972,
      "num_tokens": 128954027.0,
      "step": 809
    },
    {
      "epoch": 0.41200406917599186,
      "grad_norm": 1.1477376222610474,
      "learning_rate": 1e-05,
      "loss": 0.5514,
      "mean_token_accuracy": 0.8300490379333496,
      "num_tokens": 129117445.0,
      "step": 810
    },
    {
      "epoch": 0.4125127161749746,
      "grad_norm": 1.1172250509262085,
      "learning_rate": 1e-05,
      "loss": 0.5521,
      "mean_token_accuracy": 0.830201268196106,
      "num_tokens": 129285951.0,
      "step": 811
    },
    {
      "epoch": 0.41302136317395727,
      "grad_norm": 1.1050262451171875,
      "learning_rate": 1e-05,
      "loss": 0.5875,
      "mean_token_accuracy": 0.8233176469802856,
      "num_tokens": 129465561.0,
      "step": 812
    },
    {
      "epoch": 0.41353001017294,
      "grad_norm": 1.23516845703125,
      "learning_rate": 1e-05,
      "loss": 0.5097,
      "mean_token_accuracy": 0.8415330648422241,
      "num_tokens": 129614099.0,
      "step": 813
    },
    {
      "epoch": 0.4140386571719227,
      "grad_norm": 1.1005412340164185,
      "learning_rate": 1e-05,
      "loss": 0.55,
      "mean_token_accuracy": 0.8317654728889465,
      "num_tokens": 129789431.0,
      "step": 814
    },
    {
      "epoch": 0.4145473041709054,
      "grad_norm": 1.0260275602340698,
      "learning_rate": 1e-05,
      "loss": 0.5099,
      "mean_token_accuracy": 0.8431083559989929,
      "num_tokens": 129947839.0,
      "step": 815
    },
    {
      "epoch": 0.4150559511698881,
      "grad_norm": 1.2571784257888794,
      "learning_rate": 1e-05,
      "loss": 0.5925,
      "mean_token_accuracy": 0.8207975029945374,
      "num_tokens": 130099262.0,
      "step": 816
    },
    {
      "epoch": 0.4155645981688708,
      "grad_norm": 1.1722218990325928,
      "learning_rate": 1e-05,
      "loss": 0.5542,
      "mean_token_accuracy": 0.830123245716095,
      "num_tokens": 130262007.0,
      "step": 817
    },
    {
      "epoch": 0.4160732451678535,
      "grad_norm": 1.234281063079834,
      "learning_rate": 1e-05,
      "loss": 0.5842,
      "mean_token_accuracy": 0.8230294585227966,
      "num_tokens": 130412634.0,
      "step": 818
    },
    {
      "epoch": 0.4165818921668362,
      "grad_norm": 1.166526198387146,
      "learning_rate": 1e-05,
      "loss": 0.5665,
      "mean_token_accuracy": 0.8284645080566406,
      "num_tokens": 130576459.0,
      "step": 819
    },
    {
      "epoch": 0.4170905391658189,
      "grad_norm": 1.139403223991394,
      "learning_rate": 1e-05,
      "loss": 0.5654,
      "mean_token_accuracy": 0.8261489272117615,
      "num_tokens": 130738545.0,
      "step": 820
    },
    {
      "epoch": 0.41759918616480163,
      "grad_norm": 1.0868419408798218,
      "learning_rate": 1e-05,
      "loss": 0.5814,
      "mean_token_accuracy": 0.8246670961380005,
      "num_tokens": 130899383.0,
      "step": 821
    },
    {
      "epoch": 0.41810783316378436,
      "grad_norm": 1.1820749044418335,
      "learning_rate": 1e-05,
      "loss": 0.5486,
      "mean_token_accuracy": 0.8303079605102539,
      "num_tokens": 131069889.0,
      "step": 822
    },
    {
      "epoch": 0.41861648016276704,
      "grad_norm": 1.1270747184753418,
      "learning_rate": 1e-05,
      "loss": 0.5625,
      "mean_token_accuracy": 0.8289308547973633,
      "num_tokens": 131230218.0,
      "step": 823
    },
    {
      "epoch": 0.41912512716174977,
      "grad_norm": 1.3298698663711548,
      "learning_rate": 1e-05,
      "loss": 0.5661,
      "mean_token_accuracy": 0.826373815536499,
      "num_tokens": 131374265.0,
      "step": 824
    },
    {
      "epoch": 0.41963377416073244,
      "grad_norm": 1.1981703042984009,
      "learning_rate": 1e-05,
      "loss": 0.5483,
      "mean_token_accuracy": 0.8322027921676636,
      "num_tokens": 131543646.0,
      "step": 825
    },
    {
      "epoch": 0.4201424211597152,
      "grad_norm": 1.183638334274292,
      "learning_rate": 1e-05,
      "loss": 0.5913,
      "mean_token_accuracy": 0.8200170397758484,
      "num_tokens": 131702768.0,
      "step": 826
    },
    {
      "epoch": 0.42065106815869785,
      "grad_norm": 1.2016685009002686,
      "learning_rate": 1e-05,
      "loss": 0.5742,
      "mean_token_accuracy": 0.826677680015564,
      "num_tokens": 131860643.0,
      "step": 827
    },
    {
      "epoch": 0.4211597151576806,
      "grad_norm": 1.2117851972579956,
      "learning_rate": 1e-05,
      "loss": 0.5348,
      "mean_token_accuracy": 0.835245668888092,
      "num_tokens": 132005394.0,
      "step": 828
    },
    {
      "epoch": 0.42166836215666326,
      "grad_norm": 1.2130509614944458,
      "learning_rate": 1e-05,
      "loss": 0.5418,
      "mean_token_accuracy": 0.8323445320129395,
      "num_tokens": 132162321.0,
      "step": 829
    },
    {
      "epoch": 0.422177009155646,
      "grad_norm": 1.1499370336532593,
      "learning_rate": 1e-05,
      "loss": 0.5328,
      "mean_token_accuracy": 0.8365702629089355,
      "num_tokens": 132318929.0,
      "step": 830
    },
    {
      "epoch": 0.42268565615462866,
      "grad_norm": 1.1348990201950073,
      "learning_rate": 1e-05,
      "loss": 0.5476,
      "mean_token_accuracy": 0.8332790732383728,
      "num_tokens": 132479785.0,
      "step": 831
    },
    {
      "epoch": 0.4231943031536114,
      "grad_norm": 1.1613720655441284,
      "learning_rate": 1e-05,
      "loss": 0.5196,
      "mean_token_accuracy": 0.8395429253578186,
      "num_tokens": 132639135.0,
      "step": 832
    },
    {
      "epoch": 0.4237029501525941,
      "grad_norm": 1.314821720123291,
      "learning_rate": 1e-05,
      "loss": 0.6086,
      "mean_token_accuracy": 0.8147256970405579,
      "num_tokens": 132798835.0,
      "step": 833
    },
    {
      "epoch": 0.4242115971515768,
      "grad_norm": 1.0963644981384277,
      "learning_rate": 1e-05,
      "loss": 0.5403,
      "mean_token_accuracy": 0.8340123891830444,
      "num_tokens": 132955884.0,
      "step": 834
    },
    {
      "epoch": 0.42472024415055953,
      "grad_norm": 1.202355980873108,
      "learning_rate": 1e-05,
      "loss": 0.5384,
      "mean_token_accuracy": 0.834297776222229,
      "num_tokens": 133118521.0,
      "step": 835
    },
    {
      "epoch": 0.4252288911495422,
      "grad_norm": 1.1341054439544678,
      "learning_rate": 1e-05,
      "loss": 0.5146,
      "mean_token_accuracy": 0.842079758644104,
      "num_tokens": 133281242.0,
      "step": 836
    },
    {
      "epoch": 0.42573753814852494,
      "grad_norm": 1.160120964050293,
      "learning_rate": 1e-05,
      "loss": 0.5471,
      "mean_token_accuracy": 0.8323448896408081,
      "num_tokens": 133450654.0,
      "step": 837
    },
    {
      "epoch": 0.4262461851475076,
      "grad_norm": 1.1472262144088745,
      "learning_rate": 1e-05,
      "loss": 0.5742,
      "mean_token_accuracy": 0.8254837989807129,
      "num_tokens": 133620766.0,
      "step": 838
    },
    {
      "epoch": 0.42675483214649035,
      "grad_norm": 1.1196868419647217,
      "learning_rate": 1e-05,
      "loss": 0.5786,
      "mean_token_accuracy": 0.8239191174507141,
      "num_tokens": 133790293.0,
      "step": 839
    },
    {
      "epoch": 0.427263479145473,
      "grad_norm": 1.1362370252609253,
      "learning_rate": 1e-05,
      "loss": 0.536,
      "mean_token_accuracy": 0.8357857465744019,
      "num_tokens": 133962681.0,
      "step": 840
    },
    {
      "epoch": 0.42777212614445576,
      "grad_norm": 1.1518752574920654,
      "learning_rate": 1e-05,
      "loss": 0.57,
      "mean_token_accuracy": 0.8252599239349365,
      "num_tokens": 134128875.0,
      "step": 841
    },
    {
      "epoch": 0.42828077314343843,
      "grad_norm": 1.142612099647522,
      "learning_rate": 1e-05,
      "loss": 0.6033,
      "mean_token_accuracy": 0.8177984952926636,
      "num_tokens": 134299862.0,
      "step": 842
    },
    {
      "epoch": 0.42878942014242116,
      "grad_norm": 1.2787010669708252,
      "learning_rate": 1e-05,
      "loss": 0.535,
      "mean_token_accuracy": 0.8356817960739136,
      "num_tokens": 134447887.0,
      "step": 843
    },
    {
      "epoch": 0.42929806714140384,
      "grad_norm": 1.1462864875793457,
      "learning_rate": 1e-05,
      "loss": 0.5543,
      "mean_token_accuracy": 0.8310085535049438,
      "num_tokens": 134603411.0,
      "step": 844
    },
    {
      "epoch": 0.42980671414038657,
      "grad_norm": 1.2700620889663696,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.8362004160881042,
      "num_tokens": 134753687.0,
      "step": 845
    },
    {
      "epoch": 0.4303153611393693,
      "grad_norm": 1.2534452676773071,
      "learning_rate": 1e-05,
      "loss": 0.5627,
      "mean_token_accuracy": 0.8288543224334717,
      "num_tokens": 134910477.0,
      "step": 846
    },
    {
      "epoch": 0.430824008138352,
      "grad_norm": 1.0986623764038086,
      "learning_rate": 1e-05,
      "loss": 0.569,
      "mean_token_accuracy": 0.8261380791664124,
      "num_tokens": 135065619.0,
      "step": 847
    },
    {
      "epoch": 0.4313326551373347,
      "grad_norm": 1.3233753442764282,
      "learning_rate": 1e-05,
      "loss": 0.5968,
      "mean_token_accuracy": 0.8202333450317383,
      "num_tokens": 135215377.0,
      "step": 848
    },
    {
      "epoch": 0.4318413021363174,
      "grad_norm": 1.2474511861801147,
      "learning_rate": 1e-05,
      "loss": 0.5717,
      "mean_token_accuracy": 0.8252692222595215,
      "num_tokens": 135371149.0,
      "step": 849
    },
    {
      "epoch": 0.4323499491353001,
      "grad_norm": 1.084205150604248,
      "learning_rate": 1e-05,
      "loss": 0.5536,
      "mean_token_accuracy": 0.8310810327529907,
      "num_tokens": 135530087.0,
      "step": 850
    },
    {
      "epoch": 0.4328585961342828,
      "grad_norm": 1.149458408355713,
      "learning_rate": 1e-05,
      "loss": 0.5481,
      "mean_token_accuracy": 0.8319704532623291,
      "num_tokens": 135689766.0,
      "step": 851
    },
    {
      "epoch": 0.4333672431332655,
      "grad_norm": 1.1447380781173706,
      "learning_rate": 1e-05,
      "loss": 0.5274,
      "mean_token_accuracy": 0.838167130947113,
      "num_tokens": 135852811.0,
      "step": 852
    },
    {
      "epoch": 0.4338758901322482,
      "grad_norm": 1.070308804512024,
      "learning_rate": 1e-05,
      "loss": 0.55,
      "mean_token_accuracy": 0.8330268859863281,
      "num_tokens": 136011756.0,
      "step": 853
    },
    {
      "epoch": 0.43438453713123093,
      "grad_norm": 1.2195512056350708,
      "learning_rate": 1e-05,
      "loss": 0.5499,
      "mean_token_accuracy": 0.8309155106544495,
      "num_tokens": 136177024.0,
      "step": 854
    },
    {
      "epoch": 0.4348931841302136,
      "grad_norm": 1.1714608669281006,
      "learning_rate": 1e-05,
      "loss": 0.523,
      "mean_token_accuracy": 0.8379561305046082,
      "num_tokens": 136325905.0,
      "step": 855
    },
    {
      "epoch": 0.43540183112919634,
      "grad_norm": 1.1079695224761963,
      "learning_rate": 1e-05,
      "loss": 0.5771,
      "mean_token_accuracy": 0.8236668109893799,
      "num_tokens": 136482739.0,
      "step": 856
    },
    {
      "epoch": 0.43591047812817907,
      "grad_norm": 1.145799160003662,
      "learning_rate": 1e-05,
      "loss": 0.5307,
      "mean_token_accuracy": 0.8371036052703857,
      "num_tokens": 136652674.0,
      "step": 857
    },
    {
      "epoch": 0.43641912512716174,
      "grad_norm": 1.2814421653747559,
      "learning_rate": 1e-05,
      "loss": 0.5385,
      "mean_token_accuracy": 0.8340997695922852,
      "num_tokens": 136822688.0,
      "step": 858
    },
    {
      "epoch": 0.4369277721261445,
      "grad_norm": 1.0849554538726807,
      "learning_rate": 1e-05,
      "loss": 0.5552,
      "mean_token_accuracy": 0.8300553560256958,
      "num_tokens": 136980565.0,
      "step": 859
    },
    {
      "epoch": 0.43743641912512715,
      "grad_norm": 1.0674481391906738,
      "learning_rate": 1e-05,
      "loss": 0.5122,
      "mean_token_accuracy": 0.8417307138442993,
      "num_tokens": 137137295.0,
      "step": 860
    },
    {
      "epoch": 0.4379450661241099,
      "grad_norm": 1.2410690784454346,
      "learning_rate": 1e-05,
      "loss": 0.5406,
      "mean_token_accuracy": 0.83237624168396,
      "num_tokens": 137301870.0,
      "step": 861
    },
    {
      "epoch": 0.43845371312309256,
      "grad_norm": 1.118271827697754,
      "learning_rate": 1e-05,
      "loss": 0.5551,
      "mean_token_accuracy": 0.830010175704956,
      "num_tokens": 137449013.0,
      "step": 862
    },
    {
      "epoch": 0.4389623601220753,
      "grad_norm": 1.2069499492645264,
      "learning_rate": 1e-05,
      "loss": 0.5572,
      "mean_token_accuracy": 0.8290837407112122,
      "num_tokens": 137604020.0,
      "step": 863
    },
    {
      "epoch": 0.43947100712105797,
      "grad_norm": 1.1198499202728271,
      "learning_rate": 1e-05,
      "loss": 0.537,
      "mean_token_accuracy": 0.8342312574386597,
      "num_tokens": 137767680.0,
      "step": 864
    },
    {
      "epoch": 0.4399796541200407,
      "grad_norm": 1.056345820426941,
      "learning_rate": 1e-05,
      "loss": 0.5484,
      "mean_token_accuracy": 0.8331097960472107,
      "num_tokens": 137912184.0,
      "step": 865
    },
    {
      "epoch": 0.4404883011190234,
      "grad_norm": 1.2465364933013916,
      "learning_rate": 1e-05,
      "loss": 0.5658,
      "mean_token_accuracy": 0.8300220966339111,
      "num_tokens": 138066689.0,
      "step": 866
    },
    {
      "epoch": 0.4409969481180061,
      "grad_norm": 1.1372525691986084,
      "learning_rate": 1e-05,
      "loss": 0.5661,
      "mean_token_accuracy": 0.8273263573646545,
      "num_tokens": 138235428.0,
      "step": 867
    },
    {
      "epoch": 0.44150559511698884,
      "grad_norm": 1.4288570880889893,
      "learning_rate": 1e-05,
      "loss": 0.6132,
      "mean_token_accuracy": 0.8172624111175537,
      "num_tokens": 138390463.0,
      "step": 868
    },
    {
      "epoch": 0.4420142421159715,
      "grad_norm": 1.215345025062561,
      "learning_rate": 1e-05,
      "loss": 0.5718,
      "mean_token_accuracy": 0.8258681297302246,
      "num_tokens": 138552942.0,
      "step": 869
    },
    {
      "epoch": 0.44252288911495424,
      "grad_norm": 1.1598032712936401,
      "learning_rate": 1e-05,
      "loss": 0.5485,
      "mean_token_accuracy": 0.8313640356063843,
      "num_tokens": 138712726.0,
      "step": 870
    },
    {
      "epoch": 0.4430315361139369,
      "grad_norm": 1.1277799606323242,
      "learning_rate": 1e-05,
      "loss": 0.5425,
      "mean_token_accuracy": 0.8344129323959351,
      "num_tokens": 138880829.0,
      "step": 871
    },
    {
      "epoch": 0.44354018311291965,
      "grad_norm": 1.2340621948242188,
      "learning_rate": 1e-05,
      "loss": 0.564,
      "mean_token_accuracy": 0.826789140701294,
      "num_tokens": 139056494.0,
      "step": 872
    },
    {
      "epoch": 0.4440488301119023,
      "grad_norm": 1.2002519369125366,
      "learning_rate": 1e-05,
      "loss": 0.6162,
      "mean_token_accuracy": 0.8133549690246582,
      "num_tokens": 139216913.0,
      "step": 873
    },
    {
      "epoch": 0.44455747711088506,
      "grad_norm": 1.2109215259552002,
      "learning_rate": 1e-05,
      "loss": 0.5111,
      "mean_token_accuracy": 0.8412950038909912,
      "num_tokens": 139362971.0,
      "step": 874
    },
    {
      "epoch": 0.44506612410986773,
      "grad_norm": 1.1629414558410645,
      "learning_rate": 1e-05,
      "loss": 0.5791,
      "mean_token_accuracy": 0.8220655918121338,
      "num_tokens": 139513998.0,
      "step": 875
    },
    {
      "epoch": 0.44557477110885046,
      "grad_norm": 1.2482560873031616,
      "learning_rate": 1e-05,
      "loss": 0.5168,
      "mean_token_accuracy": 0.8395774960517883,
      "num_tokens": 139661330.0,
      "step": 876
    },
    {
      "epoch": 0.44608341810783314,
      "grad_norm": 1.1493937969207764,
      "learning_rate": 1e-05,
      "loss": 0.5436,
      "mean_token_accuracy": 0.8332647681236267,
      "num_tokens": 139830108.0,
      "step": 877
    },
    {
      "epoch": 0.44659206510681587,
      "grad_norm": 1.1169462203979492,
      "learning_rate": 1e-05,
      "loss": 0.5895,
      "mean_token_accuracy": 0.8215635418891907,
      "num_tokens": 139996848.0,
      "step": 878
    },
    {
      "epoch": 0.4471007121057986,
      "grad_norm": 1.1568292379379272,
      "learning_rate": 1e-05,
      "loss": 0.5616,
      "mean_token_accuracy": 0.8280988335609436,
      "num_tokens": 140162511.0,
      "step": 879
    },
    {
      "epoch": 0.4476093591047813,
      "grad_norm": 1.1624438762664795,
      "learning_rate": 1e-05,
      "loss": 0.5412,
      "mean_token_accuracy": 0.8338565826416016,
      "num_tokens": 140315404.0,
      "step": 880
    },
    {
      "epoch": 0.448118006103764,
      "grad_norm": 1.1456266641616821,
      "learning_rate": 1e-05,
      "loss": 0.5844,
      "mean_token_accuracy": 0.8213433623313904,
      "num_tokens": 140482165.0,
      "step": 881
    },
    {
      "epoch": 0.4486266531027467,
      "grad_norm": 1.1150074005126953,
      "learning_rate": 1e-05,
      "loss": 0.5454,
      "mean_token_accuracy": 0.8319356441497803,
      "num_tokens": 140632686.0,
      "step": 882
    },
    {
      "epoch": 0.4491353001017294,
      "grad_norm": 1.145408272743225,
      "learning_rate": 1e-05,
      "loss": 0.5557,
      "mean_token_accuracy": 0.8295482397079468,
      "num_tokens": 140795616.0,
      "step": 883
    },
    {
      "epoch": 0.4496439471007121,
      "grad_norm": 1.180821180343628,
      "learning_rate": 1e-05,
      "loss": 0.5282,
      "mean_token_accuracy": 0.8360565900802612,
      "num_tokens": 140945277.0,
      "step": 884
    },
    {
      "epoch": 0.4501525940996948,
      "grad_norm": 1.1291463375091553,
      "learning_rate": 1e-05,
      "loss": 0.5532,
      "mean_token_accuracy": 0.8307245969772339,
      "num_tokens": 141105845.0,
      "step": 885
    },
    {
      "epoch": 0.4506612410986775,
      "grad_norm": 1.061436653137207,
      "learning_rate": 1e-05,
      "loss": 0.5555,
      "mean_token_accuracy": 0.829817533493042,
      "num_tokens": 141261262.0,
      "step": 886
    },
    {
      "epoch": 0.45116988809766023,
      "grad_norm": 1.2997404336929321,
      "learning_rate": 1e-05,
      "loss": 0.5312,
      "mean_token_accuracy": 0.8381103277206421,
      "num_tokens": 141419743.0,
      "step": 887
    },
    {
      "epoch": 0.4516785350966429,
      "grad_norm": 1.1619393825531006,
      "learning_rate": 1e-05,
      "loss": 0.5665,
      "mean_token_accuracy": 0.8256223201751709,
      "num_tokens": 141584264.0,
      "step": 888
    },
    {
      "epoch": 0.45218718209562564,
      "grad_norm": 1.069907784461975,
      "learning_rate": 1e-05,
      "loss": 0.5579,
      "mean_token_accuracy": 0.8292261362075806,
      "num_tokens": 141744206.0,
      "step": 889
    },
    {
      "epoch": 0.4526958290946083,
      "grad_norm": 1.191235065460205,
      "learning_rate": 1e-05,
      "loss": 0.5037,
      "mean_token_accuracy": 0.842210590839386,
      "num_tokens": 141902690.0,
      "step": 890
    },
    {
      "epoch": 0.45320447609359105,
      "grad_norm": 1.1242882013320923,
      "learning_rate": 1e-05,
      "loss": 0.5517,
      "mean_token_accuracy": 0.8317337036132812,
      "num_tokens": 142051810.0,
      "step": 891
    },
    {
      "epoch": 0.4537131230925738,
      "grad_norm": 1.2400939464569092,
      "learning_rate": 1e-05,
      "loss": 0.5613,
      "mean_token_accuracy": 0.8265682458877563,
      "num_tokens": 142212848.0,
      "step": 892
    },
    {
      "epoch": 0.45422177009155645,
      "grad_norm": 1.1740095615386963,
      "learning_rate": 1e-05,
      "loss": 0.5759,
      "mean_token_accuracy": 0.8247115612030029,
      "num_tokens": 142378279.0,
      "step": 893
    },
    {
      "epoch": 0.4547304170905392,
      "grad_norm": 1.1696137189865112,
      "learning_rate": 1e-05,
      "loss": 0.5672,
      "mean_token_accuracy": 0.8267513513565063,
      "num_tokens": 142533419.0,
      "step": 894
    },
    {
      "epoch": 0.45523906408952186,
      "grad_norm": 1.1340491771697998,
      "learning_rate": 1e-05,
      "loss": 0.5554,
      "mean_token_accuracy": 0.8306752443313599,
      "num_tokens": 142681407.0,
      "step": 895
    },
    {
      "epoch": 0.4557477110885046,
      "grad_norm": 1.1108263731002808,
      "learning_rate": 1e-05,
      "loss": 0.5576,
      "mean_token_accuracy": 0.8294530510902405,
      "num_tokens": 142834121.0,
      "step": 896
    },
    {
      "epoch": 0.45625635808748727,
      "grad_norm": 1.1582520008087158,
      "learning_rate": 1e-05,
      "loss": 0.552,
      "mean_token_accuracy": 0.8302773237228394,
      "num_tokens": 142995941.0,
      "step": 897
    },
    {
      "epoch": 0.45676500508647,
      "grad_norm": 1.0763006210327148,
      "learning_rate": 1e-05,
      "loss": 0.5422,
      "mean_token_accuracy": 0.8331145644187927,
      "num_tokens": 143165685.0,
      "step": 898
    },
    {
      "epoch": 0.4572736520854527,
      "grad_norm": 1.2574509382247925,
      "learning_rate": 1e-05,
      "loss": 0.5354,
      "mean_token_accuracy": 0.835533618927002,
      "num_tokens": 143324501.0,
      "step": 899
    },
    {
      "epoch": 0.4577822990844354,
      "grad_norm": 1.1457747220993042,
      "learning_rate": 1e-05,
      "loss": 0.5445,
      "mean_token_accuracy": 0.8330206274986267,
      "num_tokens": 143485496.0,
      "step": 900
    },
    {
      "epoch": 0.4582909460834181,
      "grad_norm": 1.1134295463562012,
      "learning_rate": 1e-05,
      "loss": 0.5779,
      "mean_token_accuracy": 0.8217025399208069,
      "num_tokens": 143640935.0,
      "step": 901
    },
    {
      "epoch": 0.4587995930824008,
      "grad_norm": 1.3248838186264038,
      "learning_rate": 1e-05,
      "loss": 0.5638,
      "mean_token_accuracy": 0.8270464539527893,
      "num_tokens": 143805905.0,
      "step": 902
    },
    {
      "epoch": 0.45930824008138355,
      "grad_norm": 1.1526035070419312,
      "learning_rate": 1e-05,
      "loss": 0.5757,
      "mean_token_accuracy": 0.8245916366577148,
      "num_tokens": 143960760.0,
      "step": 903
    },
    {
      "epoch": 0.4598168870803662,
      "grad_norm": 1.0641709566116333,
      "learning_rate": 1e-05,
      "loss": 0.5105,
      "mean_token_accuracy": 0.8424915671348572,
      "num_tokens": 144107657.0,
      "step": 904
    },
    {
      "epoch": 0.46032553407934895,
      "grad_norm": 1.5470635890960693,
      "learning_rate": 1e-05,
      "loss": 0.5587,
      "mean_token_accuracy": 0.8290267586708069,
      "num_tokens": 144261684.0,
      "step": 905
    },
    {
      "epoch": 0.46083418107833163,
      "grad_norm": 1.2497509717941284,
      "learning_rate": 1e-05,
      "loss": 0.571,
      "mean_token_accuracy": 0.8268980979919434,
      "num_tokens": 144426788.0,
      "step": 906
    },
    {
      "epoch": 0.46134282807731436,
      "grad_norm": 1.128784418106079,
      "learning_rate": 1e-05,
      "loss": 0.5252,
      "mean_token_accuracy": 0.8391796946525574,
      "num_tokens": 144588094.0,
      "step": 907
    },
    {
      "epoch": 0.46185147507629704,
      "grad_norm": 1.4454443454742432,
      "learning_rate": 1e-05,
      "loss": 0.5486,
      "mean_token_accuracy": 0.8304920196533203,
      "num_tokens": 144742302.0,
      "step": 908
    },
    {
      "epoch": 0.46236012207527977,
      "grad_norm": 1.1578365564346313,
      "learning_rate": 1e-05,
      "loss": 0.5661,
      "mean_token_accuracy": 0.8252307176589966,
      "num_tokens": 144889815.0,
      "step": 909
    },
    {
      "epoch": 0.46286876907426244,
      "grad_norm": 1.193178653717041,
      "learning_rate": 1e-05,
      "loss": 0.5428,
      "mean_token_accuracy": 0.8328741788864136,
      "num_tokens": 145041480.0,
      "step": 910
    },
    {
      "epoch": 0.4633774160732452,
      "grad_norm": 1.1511712074279785,
      "learning_rate": 1e-05,
      "loss": 0.532,
      "mean_token_accuracy": 0.8380119800567627,
      "num_tokens": 145207314.0,
      "step": 911
    },
    {
      "epoch": 0.46388606307222785,
      "grad_norm": 1.1427394151687622,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.8285709023475647,
      "num_tokens": 145365776.0,
      "step": 912
    },
    {
      "epoch": 0.4643947100712106,
      "grad_norm": 1.1380131244659424,
      "learning_rate": 1e-05,
      "loss": 0.5531,
      "mean_token_accuracy": 0.8311206102371216,
      "num_tokens": 145530569.0,
      "step": 913
    },
    {
      "epoch": 0.4649033570701933,
      "grad_norm": 1.1422157287597656,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8367600440979004,
      "num_tokens": 145686829.0,
      "step": 914
    },
    {
      "epoch": 0.465412004069176,
      "grad_norm": 1.2226670980453491,
      "learning_rate": 1e-05,
      "loss": 0.5495,
      "mean_token_accuracy": 0.82952880859375,
      "num_tokens": 145853471.0,
      "step": 915
    },
    {
      "epoch": 0.4659206510681587,
      "grad_norm": 1.1117241382598877,
      "learning_rate": 1e-05,
      "loss": 0.5677,
      "mean_token_accuracy": 0.8247131109237671,
      "num_tokens": 146031134.0,
      "step": 916
    },
    {
      "epoch": 0.4664292980671414,
      "grad_norm": 1.225894570350647,
      "learning_rate": 1e-05,
      "loss": 0.5986,
      "mean_token_accuracy": 0.8198223114013672,
      "num_tokens": 146186384.0,
      "step": 917
    },
    {
      "epoch": 0.4669379450661241,
      "grad_norm": 1.0776770114898682,
      "learning_rate": 1e-05,
      "loss": 0.5679,
      "mean_token_accuracy": 0.8272132277488708,
      "num_tokens": 146342004.0,
      "step": 918
    },
    {
      "epoch": 0.4674465920651068,
      "grad_norm": 1.0807486772537231,
      "learning_rate": 1e-05,
      "loss": 0.5441,
      "mean_token_accuracy": 0.8315936923027039,
      "num_tokens": 146496365.0,
      "step": 919
    },
    {
      "epoch": 0.46795523906408953,
      "grad_norm": 1.1356221437454224,
      "learning_rate": 1e-05,
      "loss": 0.5655,
      "mean_token_accuracy": 0.8282874822616577,
      "num_tokens": 146660600.0,
      "step": 920
    },
    {
      "epoch": 0.4684638860630722,
      "grad_norm": 1.1465460062026978,
      "learning_rate": 1e-05,
      "loss": 0.6184,
      "mean_token_accuracy": 0.8121355772018433,
      "num_tokens": 146823441.0,
      "step": 921
    },
    {
      "epoch": 0.46897253306205494,
      "grad_norm": 1.2211737632751465,
      "learning_rate": 1e-05,
      "loss": 0.5871,
      "mean_token_accuracy": 0.8233176469802856,
      "num_tokens": 146980645.0,
      "step": 922
    },
    {
      "epoch": 0.4694811800610376,
      "grad_norm": 1.1721731424331665,
      "learning_rate": 1e-05,
      "loss": 0.58,
      "mean_token_accuracy": 0.8220207095146179,
      "num_tokens": 147140337.0,
      "step": 923
    },
    {
      "epoch": 0.46998982706002035,
      "grad_norm": 1.1544034481048584,
      "learning_rate": 1e-05,
      "loss": 0.5585,
      "mean_token_accuracy": 0.8287423849105835,
      "num_tokens": 147298752.0,
      "step": 924
    },
    {
      "epoch": 0.470498474059003,
      "grad_norm": 1.0708504915237427,
      "learning_rate": 1e-05,
      "loss": 0.5061,
      "mean_token_accuracy": 0.8421027660369873,
      "num_tokens": 147452293.0,
      "step": 925
    },
    {
      "epoch": 0.47100712105798576,
      "grad_norm": 1.4411351680755615,
      "learning_rate": 1e-05,
      "loss": 0.5483,
      "mean_token_accuracy": 0.8321117162704468,
      "num_tokens": 147606035.0,
      "step": 926
    },
    {
      "epoch": 0.4715157680569685,
      "grad_norm": 1.2126202583312988,
      "learning_rate": 1e-05,
      "loss": 0.52,
      "mean_token_accuracy": 0.8392668962478638,
      "num_tokens": 147740900.0,
      "step": 927
    },
    {
      "epoch": 0.47202441505595116,
      "grad_norm": 1.2801289558410645,
      "learning_rate": 1e-05,
      "loss": 0.5198,
      "mean_token_accuracy": 0.8385640978813171,
      "num_tokens": 147900796.0,
      "step": 928
    },
    {
      "epoch": 0.4725330620549339,
      "grad_norm": 1.2138499021530151,
      "learning_rate": 1e-05,
      "loss": 0.5677,
      "mean_token_accuracy": 0.8254543542861938,
      "num_tokens": 148048395.0,
      "step": 929
    },
    {
      "epoch": 0.47304170905391657,
      "grad_norm": 1.1250996589660645,
      "learning_rate": 1e-05,
      "loss": 0.4999,
      "mean_token_accuracy": 0.8460116386413574,
      "num_tokens": 148213700.0,
      "step": 930
    },
    {
      "epoch": 0.4735503560528993,
      "grad_norm": 1.2026902437210083,
      "learning_rate": 1e-05,
      "loss": 0.5484,
      "mean_token_accuracy": 0.8299406170845032,
      "num_tokens": 148360212.0,
      "step": 931
    },
    {
      "epoch": 0.474059003051882,
      "grad_norm": 1.1241765022277832,
      "learning_rate": 1e-05,
      "loss": 0.5436,
      "mean_token_accuracy": 0.8328520655632019,
      "num_tokens": 148519502.0,
      "step": 932
    },
    {
      "epoch": 0.4745676500508647,
      "grad_norm": 1.1928129196166992,
      "learning_rate": 1e-05,
      "loss": 0.5709,
      "mean_token_accuracy": 0.8253394365310669,
      "num_tokens": 148685441.0,
      "step": 933
    },
    {
      "epoch": 0.4750762970498474,
      "grad_norm": 1.168749451637268,
      "learning_rate": 1e-05,
      "loss": 0.5677,
      "mean_token_accuracy": 0.8280705809593201,
      "num_tokens": 148837050.0,
      "step": 934
    },
    {
      "epoch": 0.4755849440488301,
      "grad_norm": 1.1415495872497559,
      "learning_rate": 1e-05,
      "loss": 0.5312,
      "mean_token_accuracy": 0.8370417356491089,
      "num_tokens": 149001259.0,
      "step": 935
    },
    {
      "epoch": 0.4760935910478128,
      "grad_norm": 1.066994309425354,
      "learning_rate": 1e-05,
      "loss": 0.5791,
      "mean_token_accuracy": 0.8250643610954285,
      "num_tokens": 149158085.0,
      "step": 936
    },
    {
      "epoch": 0.4766022380467955,
      "grad_norm": 1.1611919403076172,
      "learning_rate": 1e-05,
      "loss": 0.5181,
      "mean_token_accuracy": 0.8408796787261963,
      "num_tokens": 149306579.0,
      "step": 937
    },
    {
      "epoch": 0.47711088504577825,
      "grad_norm": 1.1194415092468262,
      "learning_rate": 1e-05,
      "loss": 0.604,
      "mean_token_accuracy": 0.8182238340377808,
      "num_tokens": 149470715.0,
      "step": 938
    },
    {
      "epoch": 0.47761953204476093,
      "grad_norm": 1.149794578552246,
      "learning_rate": 1e-05,
      "loss": 0.5573,
      "mean_token_accuracy": 0.8303046226501465,
      "num_tokens": 149626418.0,
      "step": 939
    },
    {
      "epoch": 0.47812817904374366,
      "grad_norm": 1.0877587795257568,
      "learning_rate": 1e-05,
      "loss": 0.5724,
      "mean_token_accuracy": 0.8244882822036743,
      "num_tokens": 149779243.0,
      "step": 940
    },
    {
      "epoch": 0.47863682604272634,
      "grad_norm": 1.0702544450759888,
      "learning_rate": 1e-05,
      "loss": 0.5454,
      "mean_token_accuracy": 0.832623302936554,
      "num_tokens": 149928437.0,
      "step": 941
    },
    {
      "epoch": 0.47914547304170907,
      "grad_norm": 1.1327913999557495,
      "learning_rate": 1e-05,
      "loss": 0.6013,
      "mean_token_accuracy": 0.8169209957122803,
      "num_tokens": 150086890.0,
      "step": 942
    },
    {
      "epoch": 0.47965412004069174,
      "grad_norm": 0.9984365105628967,
      "learning_rate": 1e-05,
      "loss": 0.5139,
      "mean_token_accuracy": 0.8412238359451294,
      "num_tokens": 150246489.0,
      "step": 943
    },
    {
      "epoch": 0.4801627670396745,
      "grad_norm": 0.9913185834884644,
      "learning_rate": 1e-05,
      "loss": 0.5422,
      "mean_token_accuracy": 0.8335740566253662,
      "num_tokens": 150414091.0,
      "step": 944
    },
    {
      "epoch": 0.48067141403865715,
      "grad_norm": 1.057745337486267,
      "learning_rate": 1e-05,
      "loss": 0.5625,
      "mean_token_accuracy": 0.828140139579773,
      "num_tokens": 150581538.0,
      "step": 945
    },
    {
      "epoch": 0.4811800610376399,
      "grad_norm": 1.0825947523117065,
      "learning_rate": 1e-05,
      "loss": 0.5693,
      "mean_token_accuracy": 0.8258931636810303,
      "num_tokens": 150749759.0,
      "step": 946
    },
    {
      "epoch": 0.48168870803662256,
      "grad_norm": 1.0107381343841553,
      "learning_rate": 1e-05,
      "loss": 0.5458,
      "mean_token_accuracy": 0.8325010538101196,
      "num_tokens": 150911352.0,
      "step": 947
    },
    {
      "epoch": 0.4821973550356053,
      "grad_norm": 1.1812866926193237,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.8303778171539307,
      "num_tokens": 151078517.0,
      "step": 948
    },
    {
      "epoch": 0.482706002034588,
      "grad_norm": 1.0815045833587646,
      "learning_rate": 1e-05,
      "loss": 0.5745,
      "mean_token_accuracy": 0.8253252506256104,
      "num_tokens": 151244643.0,
      "step": 949
    },
    {
      "epoch": 0.4832146490335707,
      "grad_norm": 1.175293207168579,
      "learning_rate": 1e-05,
      "loss": 0.5307,
      "mean_token_accuracy": 0.8352696895599365,
      "num_tokens": 151404760.0,
      "step": 950
    },
    {
      "epoch": 0.48372329603255343,
      "grad_norm": 1.2163329124450684,
      "learning_rate": 1e-05,
      "loss": 0.5421,
      "mean_token_accuracy": 0.8340187072753906,
      "num_tokens": 151555236.0,
      "step": 951
    },
    {
      "epoch": 0.4842319430315361,
      "grad_norm": 1.1777931451797485,
      "learning_rate": 1e-05,
      "loss": 0.5089,
      "mean_token_accuracy": 0.841942310333252,
      "num_tokens": 151714374.0,
      "step": 952
    },
    {
      "epoch": 0.48474059003051884,
      "grad_norm": 1.1555335521697998,
      "learning_rate": 1e-05,
      "loss": 0.5672,
      "mean_token_accuracy": 0.8266357183456421,
      "num_tokens": 151873694.0,
      "step": 953
    },
    {
      "epoch": 0.4852492370295015,
      "grad_norm": 1.3585796356201172,
      "learning_rate": 1e-05,
      "loss": 0.5114,
      "mean_token_accuracy": 0.8436131477355957,
      "num_tokens": 152034628.0,
      "step": 954
    },
    {
      "epoch": 0.48575788402848424,
      "grad_norm": 1.2470099925994873,
      "learning_rate": 1e-05,
      "loss": 0.5767,
      "mean_token_accuracy": 0.8253637552261353,
      "num_tokens": 152182009.0,
      "step": 955
    },
    {
      "epoch": 0.4862665310274669,
      "grad_norm": 1.1471227407455444,
      "learning_rate": 1e-05,
      "loss": 0.5231,
      "mean_token_accuracy": 0.8380463719367981,
      "num_tokens": 152336970.0,
      "step": 956
    },
    {
      "epoch": 0.48677517802644965,
      "grad_norm": 1.0701148509979248,
      "learning_rate": 1e-05,
      "loss": 0.567,
      "mean_token_accuracy": 0.8259918689727783,
      "num_tokens": 152512035.0,
      "step": 957
    },
    {
      "epoch": 0.4872838250254323,
      "grad_norm": 1.0370410680770874,
      "learning_rate": 1e-05,
      "loss": 0.5557,
      "mean_token_accuracy": 0.8298166990280151,
      "num_tokens": 152684663.0,
      "step": 958
    },
    {
      "epoch": 0.48779247202441506,
      "grad_norm": 1.0619418621063232,
      "learning_rate": 1e-05,
      "loss": 0.5838,
      "mean_token_accuracy": 0.8224383592605591,
      "num_tokens": 152850154.0,
      "step": 959
    },
    {
      "epoch": 0.4883011190233978,
      "grad_norm": 1.0469465255737305,
      "learning_rate": 1e-05,
      "loss": 0.5535,
      "mean_token_accuracy": 0.8312027454376221,
      "num_tokens": 153010247.0,
      "step": 960
    },
    {
      "epoch": 0.48880976602238047,
      "grad_norm": 1.0778266191482544,
      "learning_rate": 1e-05,
      "loss": 0.5189,
      "mean_token_accuracy": 0.8391969203948975,
      "num_tokens": 153174299.0,
      "step": 961
    },
    {
      "epoch": 0.4893184130213632,
      "grad_norm": 2.199824810028076,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.835580050945282,
      "num_tokens": 153337306.0,
      "step": 962
    },
    {
      "epoch": 0.4898270600203459,
      "grad_norm": 1.198591947555542,
      "learning_rate": 1e-05,
      "loss": 0.562,
      "mean_token_accuracy": 0.8278141617774963,
      "num_tokens": 153492940.0,
      "step": 963
    },
    {
      "epoch": 0.4903357070193286,
      "grad_norm": 1.6478595733642578,
      "learning_rate": 1e-05,
      "loss": 0.5354,
      "mean_token_accuracy": 0.8336241245269775,
      "num_tokens": 153652759.0,
      "step": 964
    },
    {
      "epoch": 0.4908443540183113,
      "grad_norm": 1.2517733573913574,
      "learning_rate": 1e-05,
      "loss": 0.5255,
      "mean_token_accuracy": 0.837712287902832,
      "num_tokens": 153818682.0,
      "step": 965
    },
    {
      "epoch": 0.491353001017294,
      "grad_norm": 1.17851984500885,
      "learning_rate": 1e-05,
      "loss": 0.538,
      "mean_token_accuracy": 0.8355500102043152,
      "num_tokens": 153982904.0,
      "step": 966
    },
    {
      "epoch": 0.4918616480162767,
      "grad_norm": 1.2655420303344727,
      "learning_rate": 1e-05,
      "loss": 0.5469,
      "mean_token_accuracy": 0.8304451704025269,
      "num_tokens": 154142701.0,
      "step": 967
    },
    {
      "epoch": 0.4923702950152594,
      "grad_norm": 1.3124351501464844,
      "learning_rate": 1e-05,
      "loss": 0.5952,
      "mean_token_accuracy": 0.8196591138839722,
      "num_tokens": 154318454.0,
      "step": 968
    },
    {
      "epoch": 0.4928789420142421,
      "grad_norm": 1.16322922706604,
      "learning_rate": 1e-05,
      "loss": 0.5998,
      "mean_token_accuracy": 0.8195618987083435,
      "num_tokens": 154485998.0,
      "step": 969
    },
    {
      "epoch": 0.4933875890132248,
      "grad_norm": 1.2704930305480957,
      "learning_rate": 1e-05,
      "loss": 0.5492,
      "mean_token_accuracy": 0.8322117328643799,
      "num_tokens": 154646655.0,
      "step": 970
    },
    {
      "epoch": 0.4938962360122075,
      "grad_norm": 1.1108856201171875,
      "learning_rate": 1e-05,
      "loss": 0.5453,
      "mean_token_accuracy": 0.8348992466926575,
      "num_tokens": 154807349.0,
      "step": 971
    },
    {
      "epoch": 0.49440488301119023,
      "grad_norm": 1.1756478548049927,
      "learning_rate": 1e-05,
      "loss": 0.5526,
      "mean_token_accuracy": 0.829748272895813,
      "num_tokens": 154961491.0,
      "step": 972
    },
    {
      "epoch": 0.49491353001017296,
      "grad_norm": 1.2276084423065186,
      "learning_rate": 1e-05,
      "loss": 0.5433,
      "mean_token_accuracy": 0.8331449031829834,
      "num_tokens": 155114869.0,
      "step": 973
    },
    {
      "epoch": 0.49542217700915564,
      "grad_norm": 1.1128617525100708,
      "learning_rate": 1e-05,
      "loss": 0.5419,
      "mean_token_accuracy": 0.8313051462173462,
      "num_tokens": 155273562.0,
      "step": 974
    },
    {
      "epoch": 0.49593082400813837,
      "grad_norm": 1.2433724403381348,
      "learning_rate": 1e-05,
      "loss": 0.483,
      "mean_token_accuracy": 0.84830641746521,
      "num_tokens": 155431483.0,
      "step": 975
    },
    {
      "epoch": 0.49643947100712105,
      "grad_norm": 1.2141236066818237,
      "learning_rate": 1e-05,
      "loss": 0.5588,
      "mean_token_accuracy": 0.8308509588241577,
      "num_tokens": 155577059.0,
      "step": 976
    },
    {
      "epoch": 0.4969481180061038,
      "grad_norm": 1.0577449798583984,
      "learning_rate": 1e-05,
      "loss": 0.5757,
      "mean_token_accuracy": 0.8266185522079468,
      "num_tokens": 155734946.0,
      "step": 977
    },
    {
      "epoch": 0.49745676500508645,
      "grad_norm": 1.3903453350067139,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.8309212923049927,
      "num_tokens": 155896969.0,
      "step": 978
    },
    {
      "epoch": 0.4979654120040692,
      "grad_norm": 1.0999048948287964,
      "learning_rate": 1e-05,
      "loss": 0.5681,
      "mean_token_accuracy": 0.827406644821167,
      "num_tokens": 156060582.0,
      "step": 979
    },
    {
      "epoch": 0.49847405900305186,
      "grad_norm": 1.138027310371399,
      "learning_rate": 1e-05,
      "loss": 0.5366,
      "mean_token_accuracy": 0.8354133367538452,
      "num_tokens": 156221661.0,
      "step": 980
    },
    {
      "epoch": 0.4989827060020346,
      "grad_norm": 1.1465777158737183,
      "learning_rate": 1e-05,
      "loss": 0.5373,
      "mean_token_accuracy": 0.8339407444000244,
      "num_tokens": 156378238.0,
      "step": 981
    },
    {
      "epoch": 0.49949135300101727,
      "grad_norm": 1.1058530807495117,
      "learning_rate": 1e-05,
      "loss": 0.5856,
      "mean_token_accuracy": 0.820149838924408,
      "num_tokens": 156538559.0,
      "step": 982
    },
    {
      "epoch": 0.5,
      "grad_norm": 1.1324735879898071,
      "learning_rate": 1e-05,
      "loss": 0.5597,
      "mean_token_accuracy": 0.8275347948074341,
      "num_tokens": 156700036.0,
      "step": 983
    },
    {
      "epoch": 0.5005086469989827,
      "grad_norm": 1.0057607889175415,
      "learning_rate": 1e-05,
      "loss": 0.5674,
      "mean_token_accuracy": 0.8275812864303589,
      "num_tokens": 156879118.0,
      "step": 984
    },
    {
      "epoch": 0.5010172939979655,
      "grad_norm": 1.0707874298095703,
      "learning_rate": 1e-05,
      "loss": 0.5083,
      "mean_token_accuracy": 0.8437913656234741,
      "num_tokens": 157043565.0,
      "step": 985
    },
    {
      "epoch": 0.5015259409969481,
      "grad_norm": 1.1633177995681763,
      "learning_rate": 1e-05,
      "loss": 0.5465,
      "mean_token_accuracy": 0.8311898708343506,
      "num_tokens": 157198178.0,
      "step": 986
    },
    {
      "epoch": 0.5020345879959308,
      "grad_norm": 1.0955051183700562,
      "learning_rate": 1e-05,
      "loss": 0.5503,
      "mean_token_accuracy": 0.8318169116973877,
      "num_tokens": 157349621.0,
      "step": 987
    },
    {
      "epoch": 0.5025432349949135,
      "grad_norm": 1.1708135604858398,
      "learning_rate": 1e-05,
      "loss": 0.5367,
      "mean_token_accuracy": 0.835382878780365,
      "num_tokens": 157519093.0,
      "step": 988
    },
    {
      "epoch": 0.5030518819938963,
      "grad_norm": 1.075040340423584,
      "learning_rate": 1e-05,
      "loss": 0.5385,
      "mean_token_accuracy": 0.8336274027824402,
      "num_tokens": 157671835.0,
      "step": 989
    },
    {
      "epoch": 0.503560528992879,
      "grad_norm": 1.1236952543258667,
      "learning_rate": 1e-05,
      "loss": 0.5372,
      "mean_token_accuracy": 0.8344086408615112,
      "num_tokens": 157826508.0,
      "step": 990
    },
    {
      "epoch": 0.5040691759918616,
      "grad_norm": 1.1851485967636108,
      "learning_rate": 1e-05,
      "loss": 0.5642,
      "mean_token_accuracy": 0.8271092772483826,
      "num_tokens": 157983854.0,
      "step": 991
    },
    {
      "epoch": 0.5045778229908443,
      "grad_norm": 1.127675175666809,
      "learning_rate": 1e-05,
      "loss": 0.5536,
      "mean_token_accuracy": 0.8302778601646423,
      "num_tokens": 158153856.0,
      "step": 992
    },
    {
      "epoch": 0.5050864699898271,
      "grad_norm": 1.0169093608856201,
      "learning_rate": 1e-05,
      "loss": 0.5267,
      "mean_token_accuracy": 0.8384323120117188,
      "num_tokens": 158311723.0,
      "step": 993
    },
    {
      "epoch": 0.5055951169888098,
      "grad_norm": 1.11962890625,
      "learning_rate": 1e-05,
      "loss": 0.5822,
      "mean_token_accuracy": 0.825655460357666,
      "num_tokens": 158468860.0,
      "step": 994
    },
    {
      "epoch": 0.5061037639877924,
      "grad_norm": 1.0792301893234253,
      "learning_rate": 1e-05,
      "loss": 0.5576,
      "mean_token_accuracy": 0.8288288712501526,
      "num_tokens": 158623811.0,
      "step": 995
    },
    {
      "epoch": 0.5066124109867752,
      "grad_norm": 1.1109192371368408,
      "learning_rate": 1e-05,
      "loss": 0.5424,
      "mean_token_accuracy": 0.8330716490745544,
      "num_tokens": 158789067.0,
      "step": 996
    },
    {
      "epoch": 0.5071210579857579,
      "grad_norm": 1.1863211393356323,
      "learning_rate": 1e-05,
      "loss": 0.5679,
      "mean_token_accuracy": 0.8257003426551819,
      "num_tokens": 158950421.0,
      "step": 997
    },
    {
      "epoch": 0.5076297049847406,
      "grad_norm": 1.1955516338348389,
      "learning_rate": 1e-05,
      "loss": 0.5381,
      "mean_token_accuracy": 0.8345174789428711,
      "num_tokens": 159115006.0,
      "step": 998
    },
    {
      "epoch": 0.5081383519837233,
      "grad_norm": 1.2774521112442017,
      "learning_rate": 1e-05,
      "loss": 0.5341,
      "mean_token_accuracy": 0.8339723348617554,
      "num_tokens": 159267147.0,
      "step": 999
    },
    {
      "epoch": 0.508646998982706,
      "grad_norm": 1.023662805557251,
      "learning_rate": 1e-05,
      "loss": 0.5263,
      "mean_token_accuracy": 0.8373016119003296,
      "num_tokens": 159430845.0,
      "step": 1000
    },
    {
      "epoch": 0.5091556459816887,
      "grad_norm": 1.1863523721694946,
      "learning_rate": 1e-05,
      "loss": 0.5531,
      "mean_token_accuracy": 0.8315171003341675,
      "num_tokens": 159592673.0,
      "step": 1001
    },
    {
      "epoch": 0.5096642929806714,
      "grad_norm": 1.152750849723816,
      "learning_rate": 1e-05,
      "loss": 0.5563,
      "mean_token_accuracy": 0.8282335996627808,
      "num_tokens": 159766028.0,
      "step": 1002
    },
    {
      "epoch": 0.5101729399796541,
      "grad_norm": 1.1441240310668945,
      "learning_rate": 1e-05,
      "loss": 0.5446,
      "mean_token_accuracy": 0.8327241539955139,
      "num_tokens": 159928990.0,
      "step": 1003
    },
    {
      "epoch": 0.5106815869786369,
      "grad_norm": 1.1187294721603394,
      "learning_rate": 1e-05,
      "loss": 0.5088,
      "mean_token_accuracy": 0.8430362939834595,
      "num_tokens": 160093084.0,
      "step": 1004
    },
    {
      "epoch": 0.5111902339776195,
      "grad_norm": 1.1033687591552734,
      "learning_rate": 1e-05,
      "loss": 0.5685,
      "mean_token_accuracy": 0.8255316019058228,
      "num_tokens": 160253282.0,
      "step": 1005
    },
    {
      "epoch": 0.5116988809766022,
      "grad_norm": 1.1011110544204712,
      "learning_rate": 1e-05,
      "loss": 0.5266,
      "mean_token_accuracy": 0.839083731174469,
      "num_tokens": 160414551.0,
      "step": 1006
    },
    {
      "epoch": 0.512207527975585,
      "grad_norm": 1.2247616052627563,
      "learning_rate": 1e-05,
      "loss": 0.5724,
      "mean_token_accuracy": 0.8251423239707947,
      "num_tokens": 160566903.0,
      "step": 1007
    },
    {
      "epoch": 0.5127161749745677,
      "grad_norm": 0.9921314716339111,
      "learning_rate": 1e-05,
      "loss": 0.5224,
      "mean_token_accuracy": 0.8400283455848694,
      "num_tokens": 160732492.0,
      "step": 1008
    },
    {
      "epoch": 0.5132248219735503,
      "grad_norm": 1.0750997066497803,
      "learning_rate": 1e-05,
      "loss": 0.5712,
      "mean_token_accuracy": 0.825303852558136,
      "num_tokens": 160901405.0,
      "step": 1009
    },
    {
      "epoch": 0.513733468972533,
      "grad_norm": 1.1010903120040894,
      "learning_rate": 1e-05,
      "loss": 0.5304,
      "mean_token_accuracy": 0.8378999829292297,
      "num_tokens": 161063180.0,
      "step": 1010
    },
    {
      "epoch": 0.5142421159715158,
      "grad_norm": 1.110854983329773,
      "learning_rate": 1e-05,
      "loss": 0.5633,
      "mean_token_accuracy": 0.8287266492843628,
      "num_tokens": 161224667.0,
      "step": 1011
    },
    {
      "epoch": 0.5147507629704985,
      "grad_norm": 1.1269981861114502,
      "learning_rate": 1e-05,
      "loss": 0.5547,
      "mean_token_accuracy": 0.8304728865623474,
      "num_tokens": 161375840.0,
      "step": 1012
    },
    {
      "epoch": 0.5152594099694812,
      "grad_norm": 1.1674067974090576,
      "learning_rate": 1e-05,
      "loss": 0.5615,
      "mean_token_accuracy": 0.8291001319885254,
      "num_tokens": 161539381.0,
      "step": 1013
    },
    {
      "epoch": 0.5157680569684638,
      "grad_norm": 1.091300129890442,
      "learning_rate": 1e-05,
      "loss": 0.5591,
      "mean_token_accuracy": 0.8284760117530823,
      "num_tokens": 161712451.0,
      "step": 1014
    },
    {
      "epoch": 0.5162767039674466,
      "grad_norm": 1.140444278717041,
      "learning_rate": 1e-05,
      "loss": 0.559,
      "mean_token_accuracy": 0.8282883167266846,
      "num_tokens": 161891850.0,
      "step": 1015
    },
    {
      "epoch": 0.5167853509664293,
      "grad_norm": 1.1514443159103394,
      "learning_rate": 1e-05,
      "loss": 0.5263,
      "mean_token_accuracy": 0.8369109034538269,
      "num_tokens": 162053742.0,
      "step": 1016
    },
    {
      "epoch": 0.517293997965412,
      "grad_norm": 1.0525281429290771,
      "learning_rate": 1e-05,
      "loss": 0.5303,
      "mean_token_accuracy": 0.8366672992706299,
      "num_tokens": 162209650.0,
      "step": 1017
    },
    {
      "epoch": 0.5178026449643948,
      "grad_norm": 1.1147992610931396,
      "learning_rate": 1e-05,
      "loss": 0.58,
      "mean_token_accuracy": 0.8237215280532837,
      "num_tokens": 162383502.0,
      "step": 1018
    },
    {
      "epoch": 0.5183112919633774,
      "grad_norm": 1.1428780555725098,
      "learning_rate": 1e-05,
      "loss": 0.5154,
      "mean_token_accuracy": 0.8408278822898865,
      "num_tokens": 162544656.0,
      "step": 1019
    },
    {
      "epoch": 0.5188199389623601,
      "grad_norm": 1.012333631515503,
      "learning_rate": 1e-05,
      "loss": 0.556,
      "mean_token_accuracy": 0.8279523849487305,
      "num_tokens": 162721800.0,
      "step": 1020
    },
    {
      "epoch": 0.5193285859613428,
      "grad_norm": 1.2154921293258667,
      "learning_rate": 1e-05,
      "loss": 0.5456,
      "mean_token_accuracy": 0.8320083618164062,
      "num_tokens": 162876310.0,
      "step": 1021
    },
    {
      "epoch": 0.5198372329603256,
      "grad_norm": 1.0225021839141846,
      "learning_rate": 1e-05,
      "loss": 0.5454,
      "mean_token_accuracy": 0.8321446776390076,
      "num_tokens": 163040860.0,
      "step": 1022
    },
    {
      "epoch": 0.5203458799593083,
      "grad_norm": 1.1398831605911255,
      "learning_rate": 1e-05,
      "loss": 0.5557,
      "mean_token_accuracy": 0.8298647999763489,
      "num_tokens": 163207134.0,
      "step": 1023
    },
    {
      "epoch": 0.5208545269582909,
      "grad_norm": 1.1399364471435547,
      "learning_rate": 1e-05,
      "loss": 0.5316,
      "mean_token_accuracy": 0.8351213335990906,
      "num_tokens": 163354092.0,
      "step": 1024
    },
    {
      "epoch": 0.5213631739572736,
      "grad_norm": 1.1851685047149658,
      "learning_rate": 1e-05,
      "loss": 0.5525,
      "mean_token_accuracy": 0.8305678367614746,
      "num_tokens": 163506047.0,
      "step": 1025
    },
    {
      "epoch": 0.5218718209562564,
      "grad_norm": 1.1319615840911865,
      "learning_rate": 1e-05,
      "loss": 0.5456,
      "mean_token_accuracy": 0.832294762134552,
      "num_tokens": 163677302.0,
      "step": 1026
    },
    {
      "epoch": 0.5223804679552391,
      "grad_norm": 1.1624354124069214,
      "learning_rate": 1e-05,
      "loss": 0.5315,
      "mean_token_accuracy": 0.8342126607894897,
      "num_tokens": 163838092.0,
      "step": 1027
    },
    {
      "epoch": 0.5228891149542217,
      "grad_norm": 1.1127524375915527,
      "learning_rate": 1e-05,
      "loss": 0.5237,
      "mean_token_accuracy": 0.8383346796035767,
      "num_tokens": 163993198.0,
      "step": 1028
    },
    {
      "epoch": 0.5233977619532044,
      "grad_norm": 1.3149102926254272,
      "learning_rate": 1e-05,
      "loss": 0.5819,
      "mean_token_accuracy": 0.8223558068275452,
      "num_tokens": 164159543.0,
      "step": 1029
    },
    {
      "epoch": 0.5239064089521872,
      "grad_norm": 1.1892825365066528,
      "learning_rate": 1e-05,
      "loss": 0.5362,
      "mean_token_accuracy": 0.8346065282821655,
      "num_tokens": 164325603.0,
      "step": 1030
    },
    {
      "epoch": 0.5244150559511699,
      "grad_norm": 1.3377810716629028,
      "learning_rate": 1e-05,
      "loss": 0.5318,
      "mean_token_accuracy": 0.8361243009567261,
      "num_tokens": 164481267.0,
      "step": 1031
    },
    {
      "epoch": 0.5249237029501526,
      "grad_norm": 1.2706295251846313,
      "learning_rate": 1e-05,
      "loss": 0.5183,
      "mean_token_accuracy": 0.8390676975250244,
      "num_tokens": 164634955.0,
      "step": 1032
    },
    {
      "epoch": 0.5254323499491353,
      "grad_norm": 1.0936428308486938,
      "learning_rate": 1e-05,
      "loss": 0.5247,
      "mean_token_accuracy": 0.8365492820739746,
      "num_tokens": 164788873.0,
      "step": 1033
    },
    {
      "epoch": 0.525940996948118,
      "grad_norm": 1.2996271848678589,
      "learning_rate": 1e-05,
      "loss": 0.5261,
      "mean_token_accuracy": 0.8357228636741638,
      "num_tokens": 164949736.0,
      "step": 1034
    },
    {
      "epoch": 0.5264496439471007,
      "grad_norm": 1.0801299810409546,
      "learning_rate": 1e-05,
      "loss": 0.5636,
      "mean_token_accuracy": 0.826639711856842,
      "num_tokens": 165111991.0,
      "step": 1035
    },
    {
      "epoch": 0.5269582909460834,
      "grad_norm": 1.163053274154663,
      "learning_rate": 1e-05,
      "loss": 0.5301,
      "mean_token_accuracy": 0.837220311164856,
      "num_tokens": 165268681.0,
      "step": 1036
    },
    {
      "epoch": 0.5274669379450662,
      "grad_norm": 1.0252866744995117,
      "learning_rate": 1e-05,
      "loss": 0.514,
      "mean_token_accuracy": 0.841158390045166,
      "num_tokens": 165439160.0,
      "step": 1037
    },
    {
      "epoch": 0.5279755849440488,
      "grad_norm": 1.0195773839950562,
      "learning_rate": 1e-05,
      "loss": 0.5127,
      "mean_token_accuracy": 0.8401865363121033,
      "num_tokens": 165611395.0,
      "step": 1038
    },
    {
      "epoch": 0.5284842319430315,
      "grad_norm": 1.4410414695739746,
      "learning_rate": 1e-05,
      "loss": 0.5624,
      "mean_token_accuracy": 0.8270487189292908,
      "num_tokens": 165770065.0,
      "step": 1039
    },
    {
      "epoch": 0.5289928789420142,
      "grad_norm": 1.1248342990875244,
      "learning_rate": 1e-05,
      "loss": 0.5825,
      "mean_token_accuracy": 0.8211767673492432,
      "num_tokens": 165930541.0,
      "step": 1040
    },
    {
      "epoch": 0.529501525940997,
      "grad_norm": 1.0635501146316528,
      "learning_rate": 1e-05,
      "loss": 0.5391,
      "mean_token_accuracy": 0.8336089849472046,
      "num_tokens": 166095798.0,
      "step": 1041
    },
    {
      "epoch": 0.5300101729399797,
      "grad_norm": 1.1478872299194336,
      "learning_rate": 1e-05,
      "loss": 0.5411,
      "mean_token_accuracy": 0.8351947665214539,
      "num_tokens": 166253912.0,
      "step": 1042
    },
    {
      "epoch": 0.5305188199389623,
      "grad_norm": 1.0534822940826416,
      "learning_rate": 1e-05,
      "loss": 0.5214,
      "mean_token_accuracy": 0.8404939770698547,
      "num_tokens": 166414865.0,
      "step": 1043
    },
    {
      "epoch": 0.5310274669379451,
      "grad_norm": 1.2754970788955688,
      "learning_rate": 1e-05,
      "loss": 0.5586,
      "mean_token_accuracy": 0.8279399871826172,
      "num_tokens": 166570026.0,
      "step": 1044
    },
    {
      "epoch": 0.5315361139369278,
      "grad_norm": 1.091422438621521,
      "learning_rate": 1e-05,
      "loss": 0.5383,
      "mean_token_accuracy": 0.8338953852653503,
      "num_tokens": 166723092.0,
      "step": 1045
    },
    {
      "epoch": 0.5320447609359105,
      "grad_norm": 1.176932454109192,
      "learning_rate": 1e-05,
      "loss": 0.5556,
      "mean_token_accuracy": 0.8305802345275879,
      "num_tokens": 166891414.0,
      "step": 1046
    },
    {
      "epoch": 0.5325534079348931,
      "grad_norm": 1.0924545526504517,
      "learning_rate": 1e-05,
      "loss": 0.5757,
      "mean_token_accuracy": 0.8248872756958008,
      "num_tokens": 167043099.0,
      "step": 1047
    },
    {
      "epoch": 0.5330620549338759,
      "grad_norm": 1.1945780515670776,
      "learning_rate": 1e-05,
      "loss": 0.5437,
      "mean_token_accuracy": 0.833451509475708,
      "num_tokens": 167199448.0,
      "step": 1048
    },
    {
      "epoch": 0.5335707019328586,
      "grad_norm": 1.0573724508285522,
      "learning_rate": 1e-05,
      "loss": 0.5551,
      "mean_token_accuracy": 0.8307552337646484,
      "num_tokens": 167356786.0,
      "step": 1049
    },
    {
      "epoch": 0.5340793489318413,
      "grad_norm": 1.07781982421875,
      "learning_rate": 1e-05,
      "loss": 0.5367,
      "mean_token_accuracy": 0.8337694406509399,
      "num_tokens": 167508836.0,
      "step": 1050
    },
    {
      "epoch": 0.534587995930824,
      "grad_norm": 1.102200984954834,
      "learning_rate": 1e-05,
      "loss": 0.5294,
      "mean_token_accuracy": 0.836653470993042,
      "num_tokens": 167670690.0,
      "step": 1051
    },
    {
      "epoch": 0.5350966429298067,
      "grad_norm": 1.1723092794418335,
      "learning_rate": 1e-05,
      "loss": 0.5686,
      "mean_token_accuracy": 0.8260093331336975,
      "num_tokens": 167821540.0,
      "step": 1052
    },
    {
      "epoch": 0.5356052899287894,
      "grad_norm": 1.1371201276779175,
      "learning_rate": 1e-05,
      "loss": 0.5225,
      "mean_token_accuracy": 0.8382167220115662,
      "num_tokens": 167998236.0,
      "step": 1053
    },
    {
      "epoch": 0.5361139369277721,
      "grad_norm": 1.0949405431747437,
      "learning_rate": 1e-05,
      "loss": 0.5429,
      "mean_token_accuracy": 0.8328778743743896,
      "num_tokens": 168163810.0,
      "step": 1054
    },
    {
      "epoch": 0.5366225839267549,
      "grad_norm": 1.0404272079467773,
      "learning_rate": 1e-05,
      "loss": 0.5113,
      "mean_token_accuracy": 0.8400763273239136,
      "num_tokens": 168312526.0,
      "step": 1055
    },
    {
      "epoch": 0.5371312309257376,
      "grad_norm": 1.4301345348358154,
      "learning_rate": 1e-05,
      "loss": 0.5279,
      "mean_token_accuracy": 0.8389478921890259,
      "num_tokens": 168465144.0,
      "step": 1056
    },
    {
      "epoch": 0.5376398779247202,
      "grad_norm": 1.148573875427246,
      "learning_rate": 1e-05,
      "loss": 0.5309,
      "mean_token_accuracy": 0.8361720442771912,
      "num_tokens": 168621602.0,
      "step": 1057
    },
    {
      "epoch": 0.5381485249237029,
      "grad_norm": 1.119145393371582,
      "learning_rate": 1e-05,
      "loss": 0.5687,
      "mean_token_accuracy": 0.8250982165336609,
      "num_tokens": 168772296.0,
      "step": 1058
    },
    {
      "epoch": 0.5386571719226857,
      "grad_norm": 1.237609624862671,
      "learning_rate": 1e-05,
      "loss": 0.5307,
      "mean_token_accuracy": 0.835430383682251,
      "num_tokens": 168930909.0,
      "step": 1059
    },
    {
      "epoch": 0.5391658189216684,
      "grad_norm": 1.0655325651168823,
      "learning_rate": 1e-05,
      "loss": 0.5587,
      "mean_token_accuracy": 0.828170895576477,
      "num_tokens": 169084941.0,
      "step": 1060
    },
    {
      "epoch": 0.539674465920651,
      "grad_norm": 1.2636488676071167,
      "learning_rate": 1e-05,
      "loss": 0.5254,
      "mean_token_accuracy": 0.8372913599014282,
      "num_tokens": 169235523.0,
      "step": 1061
    },
    {
      "epoch": 0.5401831129196337,
      "grad_norm": 1.2027055025100708,
      "learning_rate": 1e-05,
      "loss": 0.526,
      "mean_token_accuracy": 0.8375190496444702,
      "num_tokens": 169398931.0,
      "step": 1062
    },
    {
      "epoch": 0.5406917599186165,
      "grad_norm": 1.132272720336914,
      "learning_rate": 1e-05,
      "loss": 0.5217,
      "mean_token_accuracy": 0.8366562724113464,
      "num_tokens": 169559733.0,
      "step": 1063
    },
    {
      "epoch": 0.5412004069175992,
      "grad_norm": 1.028488039970398,
      "learning_rate": 1e-05,
      "loss": 0.5385,
      "mean_token_accuracy": 0.8357600569725037,
      "num_tokens": 169713890.0,
      "step": 1064
    },
    {
      "epoch": 0.5417090539165819,
      "grad_norm": 1.1254281997680664,
      "learning_rate": 1e-05,
      "loss": 0.5288,
      "mean_token_accuracy": 0.8391003608703613,
      "num_tokens": 169872319.0,
      "step": 1065
    },
    {
      "epoch": 0.5422177009155646,
      "grad_norm": 1.1480300426483154,
      "learning_rate": 1e-05,
      "loss": 0.5769,
      "mean_token_accuracy": 0.8241935968399048,
      "num_tokens": 170042021.0,
      "step": 1066
    },
    {
      "epoch": 0.5427263479145473,
      "grad_norm": 1.2215930223464966,
      "learning_rate": 1e-05,
      "loss": 0.5465,
      "mean_token_accuracy": 0.8304744362831116,
      "num_tokens": 170190297.0,
      "step": 1067
    },
    {
      "epoch": 0.54323499491353,
      "grad_norm": 1.1113042831420898,
      "learning_rate": 1e-05,
      "loss": 0.5116,
      "mean_token_accuracy": 0.8419883251190186,
      "num_tokens": 170344681.0,
      "step": 1068
    },
    {
      "epoch": 0.5437436419125127,
      "grad_norm": 1.1927094459533691,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.8285712003707886,
      "num_tokens": 170503123.0,
      "step": 1069
    },
    {
      "epoch": 0.5442522889114955,
      "grad_norm": 1.1104354858398438,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.8371394872665405,
      "num_tokens": 170662002.0,
      "step": 1070
    },
    {
      "epoch": 0.5447609359104781,
      "grad_norm": 1.0148671865463257,
      "learning_rate": 1e-05,
      "loss": 0.5888,
      "mean_token_accuracy": 0.8197829723358154,
      "num_tokens": 170819743.0,
      "step": 1071
    },
    {
      "epoch": 0.5452695829094608,
      "grad_norm": 1.0770317316055298,
      "learning_rate": 1e-05,
      "loss": 0.5127,
      "mean_token_accuracy": 0.8423032760620117,
      "num_tokens": 170965381.0,
      "step": 1072
    },
    {
      "epoch": 0.5457782299084435,
      "grad_norm": 1.5431606769561768,
      "learning_rate": 1e-05,
      "loss": 0.5538,
      "mean_token_accuracy": 0.8318929076194763,
      "num_tokens": 171118434.0,
      "step": 1073
    },
    {
      "epoch": 0.5462868769074263,
      "grad_norm": 1.0981653928756714,
      "learning_rate": 1e-05,
      "loss": 0.5933,
      "mean_token_accuracy": 0.8221914768218994,
      "num_tokens": 171297534.0,
      "step": 1074
    },
    {
      "epoch": 0.546795523906409,
      "grad_norm": 1.0307037830352783,
      "learning_rate": 1e-05,
      "loss": 0.527,
      "mean_token_accuracy": 0.8369515538215637,
      "num_tokens": 171454468.0,
      "step": 1075
    },
    {
      "epoch": 0.5473041709053916,
      "grad_norm": 1.146727442741394,
      "learning_rate": 1e-05,
      "loss": 0.558,
      "mean_token_accuracy": 0.8274430632591248,
      "num_tokens": 171615049.0,
      "step": 1076
    },
    {
      "epoch": 0.5478128179043744,
      "grad_norm": 1.060461401939392,
      "learning_rate": 1e-05,
      "loss": 0.5265,
      "mean_token_accuracy": 0.8362975120544434,
      "num_tokens": 171777974.0,
      "step": 1077
    },
    {
      "epoch": 0.5483214649033571,
      "grad_norm": 1.0394412279129028,
      "learning_rate": 1e-05,
      "loss": 0.5253,
      "mean_token_accuracy": 0.8367418050765991,
      "num_tokens": 171931276.0,
      "step": 1078
    },
    {
      "epoch": 0.5488301119023398,
      "grad_norm": 1.0544158220291138,
      "learning_rate": 1e-05,
      "loss": 0.5323,
      "mean_token_accuracy": 0.8363548517227173,
      "num_tokens": 172092986.0,
      "step": 1079
    },
    {
      "epoch": 0.5493387589013224,
      "grad_norm": 1.0615379810333252,
      "learning_rate": 1e-05,
      "loss": 0.5448,
      "mean_token_accuracy": 0.8330628871917725,
      "num_tokens": 172259334.0,
      "step": 1080
    },
    {
      "epoch": 0.5498474059003052,
      "grad_norm": 1.0669916868209839,
      "learning_rate": 1e-05,
      "loss": 0.5323,
      "mean_token_accuracy": 0.8377177715301514,
      "num_tokens": 172432734.0,
      "step": 1081
    },
    {
      "epoch": 0.5503560528992879,
      "grad_norm": 1.0713269710540771,
      "learning_rate": 1e-05,
      "loss": 0.5222,
      "mean_token_accuracy": 0.8381586074829102,
      "num_tokens": 172580086.0,
      "step": 1082
    },
    {
      "epoch": 0.5508646998982706,
      "grad_norm": 1.1586099863052368,
      "learning_rate": 1e-05,
      "loss": 0.5418,
      "mean_token_accuracy": 0.8332674503326416,
      "num_tokens": 172733854.0,
      "step": 1083
    },
    {
      "epoch": 0.5513733468972533,
      "grad_norm": 1.0385310649871826,
      "learning_rate": 1e-05,
      "loss": 0.5574,
      "mean_token_accuracy": 0.8295480012893677,
      "num_tokens": 172895819.0,
      "step": 1084
    },
    {
      "epoch": 0.551881993896236,
      "grad_norm": 1.0283936262130737,
      "learning_rate": 1e-05,
      "loss": 0.5268,
      "mean_token_accuracy": 0.8370460867881775,
      "num_tokens": 173057284.0,
      "step": 1085
    },
    {
      "epoch": 0.5523906408952187,
      "grad_norm": 1.065392255783081,
      "learning_rate": 1e-05,
      "loss": 0.5244,
      "mean_token_accuracy": 0.8387396931648254,
      "num_tokens": 173207899.0,
      "step": 1086
    },
    {
      "epoch": 0.5528992878942014,
      "grad_norm": 1.185757040977478,
      "learning_rate": 1e-05,
      "loss": 0.5484,
      "mean_token_accuracy": 0.8312965631484985,
      "num_tokens": 173378839.0,
      "step": 1087
    },
    {
      "epoch": 0.5534079348931842,
      "grad_norm": 1.1472268104553223,
      "learning_rate": 1e-05,
      "loss": 0.5372,
      "mean_token_accuracy": 0.8326317667961121,
      "num_tokens": 173547290.0,
      "step": 1088
    },
    {
      "epoch": 0.5539165818921669,
      "grad_norm": 1.1030000448226929,
      "learning_rate": 1e-05,
      "loss": 0.5495,
      "mean_token_accuracy": 0.8318257331848145,
      "num_tokens": 173706562.0,
      "step": 1089
    },
    {
      "epoch": 0.5544252288911495,
      "grad_norm": 1.065309762954712,
      "learning_rate": 1e-05,
      "loss": 0.5729,
      "mean_token_accuracy": 0.8261637687683105,
      "num_tokens": 173872148.0,
      "step": 1090
    },
    {
      "epoch": 0.5549338758901322,
      "grad_norm": 1.1334059238433838,
      "learning_rate": 1e-05,
      "loss": 0.5297,
      "mean_token_accuracy": 0.8351319432258606,
      "num_tokens": 174021510.0,
      "step": 1091
    },
    {
      "epoch": 0.555442522889115,
      "grad_norm": 1.1944429874420166,
      "learning_rate": 1e-05,
      "loss": 0.5754,
      "mean_token_accuracy": 0.8241192102432251,
      "num_tokens": 174174594.0,
      "step": 1092
    },
    {
      "epoch": 0.5559511698880977,
      "grad_norm": 1.242767095565796,
      "learning_rate": 1e-05,
      "loss": 0.5618,
      "mean_token_accuracy": 0.8247935771942139,
      "num_tokens": 174327672.0,
      "step": 1093
    },
    {
      "epoch": 0.5564598168870803,
      "grad_norm": 1.1942962408065796,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8370338678359985,
      "num_tokens": 174483118.0,
      "step": 1094
    },
    {
      "epoch": 0.556968463886063,
      "grad_norm": 1.097080111503601,
      "learning_rate": 1e-05,
      "loss": 0.5307,
      "mean_token_accuracy": 0.8347339630126953,
      "num_tokens": 174644489.0,
      "step": 1095
    },
    {
      "epoch": 0.5574771108850458,
      "grad_norm": 1.1676805019378662,
      "learning_rate": 1e-05,
      "loss": 0.5421,
      "mean_token_accuracy": 0.8333439826965332,
      "num_tokens": 174798555.0,
      "step": 1096
    },
    {
      "epoch": 0.5579857578840285,
      "grad_norm": 1.094206690788269,
      "learning_rate": 1e-05,
      "loss": 0.535,
      "mean_token_accuracy": 0.8350372910499573,
      "num_tokens": 174955732.0,
      "step": 1097
    },
    {
      "epoch": 0.5584944048830112,
      "grad_norm": 0.9773333668708801,
      "learning_rate": 1e-05,
      "loss": 0.5332,
      "mean_token_accuracy": 0.8351495862007141,
      "num_tokens": 175130207.0,
      "step": 1098
    },
    {
      "epoch": 0.559003051881994,
      "grad_norm": 1.1231311559677124,
      "learning_rate": 1e-05,
      "loss": 0.5168,
      "mean_token_accuracy": 0.8395377397537231,
      "num_tokens": 175283656.0,
      "step": 1099
    },
    {
      "epoch": 0.5595116988809766,
      "grad_norm": 1.065348744392395,
      "learning_rate": 1e-05,
      "loss": 0.5255,
      "mean_token_accuracy": 0.8385817408561707,
      "num_tokens": 175445505.0,
      "step": 1100
    },
    {
      "epoch": 0.5600203458799593,
      "grad_norm": 1.126814603805542,
      "learning_rate": 1e-05,
      "loss": 0.5341,
      "mean_token_accuracy": 0.8354888558387756,
      "num_tokens": 175600838.0,
      "step": 1101
    },
    {
      "epoch": 0.560528992878942,
      "grad_norm": 1.0460002422332764,
      "learning_rate": 1e-05,
      "loss": 0.5612,
      "mean_token_accuracy": 0.827361524105072,
      "num_tokens": 175767713.0,
      "step": 1102
    },
    {
      "epoch": 0.5610376398779248,
      "grad_norm": 1.156819462776184,
      "learning_rate": 1e-05,
      "loss": 0.5619,
      "mean_token_accuracy": 0.8260256052017212,
      "num_tokens": 175927627.0,
      "step": 1103
    },
    {
      "epoch": 0.5615462868769074,
      "grad_norm": 1.053056001663208,
      "learning_rate": 1e-05,
      "loss": 0.5522,
      "mean_token_accuracy": 0.8309029340744019,
      "num_tokens": 176094548.0,
      "step": 1104
    },
    {
      "epoch": 0.5620549338758901,
      "grad_norm": 1.022141933441162,
      "learning_rate": 1e-05,
      "loss": 0.5339,
      "mean_token_accuracy": 0.8342273235321045,
      "num_tokens": 176249226.0,
      "step": 1105
    },
    {
      "epoch": 0.5625635808748728,
      "grad_norm": 1.0103915929794312,
      "learning_rate": 1e-05,
      "loss": 0.5394,
      "mean_token_accuracy": 0.8349782824516296,
      "num_tokens": 176419770.0,
      "step": 1106
    },
    {
      "epoch": 0.5630722278738556,
      "grad_norm": 1.1695556640625,
      "learning_rate": 1e-05,
      "loss": 0.5223,
      "mean_token_accuracy": 0.8384734988212585,
      "num_tokens": 176579372.0,
      "step": 1107
    },
    {
      "epoch": 0.5635808748728383,
      "grad_norm": 1.1732338666915894,
      "learning_rate": 1e-05,
      "loss": 0.5587,
      "mean_token_accuracy": 0.8287216424942017,
      "num_tokens": 176731375.0,
      "step": 1108
    },
    {
      "epoch": 0.5640895218718209,
      "grad_norm": 1.058566689491272,
      "learning_rate": 1e-05,
      "loss": 0.5262,
      "mean_token_accuracy": 0.8381431698799133,
      "num_tokens": 176882380.0,
      "step": 1109
    },
    {
      "epoch": 0.5645981688708036,
      "grad_norm": 1.0918478965759277,
      "learning_rate": 1e-05,
      "loss": 0.5236,
      "mean_token_accuracy": 0.8384039402008057,
      "num_tokens": 177037040.0,
      "step": 1110
    },
    {
      "epoch": 0.5651068158697864,
      "grad_norm": 1.040511965751648,
      "learning_rate": 1e-05,
      "loss": 0.563,
      "mean_token_accuracy": 0.8262896537780762,
      "num_tokens": 177193855.0,
      "step": 1111
    },
    {
      "epoch": 0.5656154628687691,
      "grad_norm": 1.1200957298278809,
      "learning_rate": 1e-05,
      "loss": 0.5427,
      "mean_token_accuracy": 0.8336125016212463,
      "num_tokens": 177353869.0,
      "step": 1112
    },
    {
      "epoch": 0.5661241098677517,
      "grad_norm": 1.0688713788986206,
      "learning_rate": 1e-05,
      "loss": 0.5307,
      "mean_token_accuracy": 0.8350938558578491,
      "num_tokens": 177503862.0,
      "step": 1113
    },
    {
      "epoch": 0.5666327568667345,
      "grad_norm": 1.0373692512512207,
      "learning_rate": 1e-05,
      "loss": 0.5115,
      "mean_token_accuracy": 0.8404502272605896,
      "num_tokens": 177663239.0,
      "step": 1114
    },
    {
      "epoch": 0.5671414038657172,
      "grad_norm": 1.0870791673660278,
      "learning_rate": 1e-05,
      "loss": 0.5215,
      "mean_token_accuracy": 0.8382793068885803,
      "num_tokens": 177822738.0,
      "step": 1115
    },
    {
      "epoch": 0.5676500508646999,
      "grad_norm": 1.2546334266662598,
      "learning_rate": 1e-05,
      "loss": 0.5379,
      "mean_token_accuracy": 0.8341033458709717,
      "num_tokens": 177989713.0,
      "step": 1116
    },
    {
      "epoch": 0.5681586978636826,
      "grad_norm": 1.1117446422576904,
      "learning_rate": 1e-05,
      "loss": 0.5515,
      "mean_token_accuracy": 0.8312805891036987,
      "num_tokens": 178140529.0,
      "step": 1117
    },
    {
      "epoch": 0.5686673448626653,
      "grad_norm": 1.0564945936203003,
      "learning_rate": 1e-05,
      "loss": 0.5421,
      "mean_token_accuracy": 0.8333523273468018,
      "num_tokens": 178311252.0,
      "step": 1118
    },
    {
      "epoch": 0.569175991861648,
      "grad_norm": 1.1679905652999878,
      "learning_rate": 1e-05,
      "loss": 0.5638,
      "mean_token_accuracy": 0.8269079923629761,
      "num_tokens": 178475164.0,
      "step": 1119
    },
    {
      "epoch": 0.5696846388606307,
      "grad_norm": 1.115697979927063,
      "learning_rate": 1e-05,
      "loss": 0.5685,
      "mean_token_accuracy": 0.8261163830757141,
      "num_tokens": 178635634.0,
      "step": 1120
    },
    {
      "epoch": 0.5701932858596134,
      "grad_norm": 1.1107937097549438,
      "learning_rate": 1e-05,
      "loss": 0.5671,
      "mean_token_accuracy": 0.8257395625114441,
      "num_tokens": 178786348.0,
      "step": 1121
    },
    {
      "epoch": 0.5707019328585962,
      "grad_norm": 1.0822064876556396,
      "learning_rate": 1e-05,
      "loss": 0.5457,
      "mean_token_accuracy": 0.8325570225715637,
      "num_tokens": 178948373.0,
      "step": 1122
    },
    {
      "epoch": 0.5712105798575788,
      "grad_norm": 1.142685055732727,
      "learning_rate": 1e-05,
      "loss": 0.509,
      "mean_token_accuracy": 0.8404792547225952,
      "num_tokens": 179104122.0,
      "step": 1123
    },
    {
      "epoch": 0.5717192268565615,
      "grad_norm": 1.0351077318191528,
      "learning_rate": 1e-05,
      "loss": 0.5318,
      "mean_token_accuracy": 0.8346115350723267,
      "num_tokens": 179265598.0,
      "step": 1124
    },
    {
      "epoch": 0.5722278738555443,
      "grad_norm": 1.0350379943847656,
      "learning_rate": 1e-05,
      "loss": 0.5521,
      "mean_token_accuracy": 0.8314869999885559,
      "num_tokens": 179417035.0,
      "step": 1125
    },
    {
      "epoch": 0.572736520854527,
      "grad_norm": 1.0398163795471191,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8392678499221802,
      "num_tokens": 179577100.0,
      "step": 1126
    },
    {
      "epoch": 0.5732451678535097,
      "grad_norm": 1.0036226511001587,
      "learning_rate": 1e-05,
      "loss": 0.5552,
      "mean_token_accuracy": 0.8290267586708069,
      "num_tokens": 179744420.0,
      "step": 1127
    },
    {
      "epoch": 0.5737538148524923,
      "grad_norm": 1.1067705154418945,
      "learning_rate": 1e-05,
      "loss": 0.54,
      "mean_token_accuracy": 0.8334242701530457,
      "num_tokens": 179892505.0,
      "step": 1128
    },
    {
      "epoch": 0.5742624618514751,
      "grad_norm": 1.1229734420776367,
      "learning_rate": 1e-05,
      "loss": 0.5621,
      "mean_token_accuracy": 0.8270100355148315,
      "num_tokens": 180051796.0,
      "step": 1129
    },
    {
      "epoch": 0.5747711088504578,
      "grad_norm": 1.0072256326675415,
      "learning_rate": 1e-05,
      "loss": 0.5334,
      "mean_token_accuracy": 0.8366422653198242,
      "num_tokens": 180223161.0,
      "step": 1130
    },
    {
      "epoch": 0.5752797558494405,
      "grad_norm": 1.038099765777588,
      "learning_rate": 1e-05,
      "loss": 0.5532,
      "mean_token_accuracy": 0.8312008380889893,
      "num_tokens": 180381038.0,
      "step": 1131
    },
    {
      "epoch": 0.5757884028484231,
      "grad_norm": 1.06266450881958,
      "learning_rate": 1e-05,
      "loss": 0.5434,
      "mean_token_accuracy": 0.8321468234062195,
      "num_tokens": 180529292.0,
      "step": 1132
    },
    {
      "epoch": 0.5762970498474059,
      "grad_norm": 1.0975128412246704,
      "learning_rate": 1e-05,
      "loss": 0.5221,
      "mean_token_accuracy": 0.8384957313537598,
      "num_tokens": 180673192.0,
      "step": 1133
    },
    {
      "epoch": 0.5768056968463886,
      "grad_norm": 1.063176155090332,
      "learning_rate": 1e-05,
      "loss": 0.5348,
      "mean_token_accuracy": 0.8349170684814453,
      "num_tokens": 180831990.0,
      "step": 1134
    },
    {
      "epoch": 0.5773143438453713,
      "grad_norm": 1.0480600595474243,
      "learning_rate": 1e-05,
      "loss": 0.5492,
      "mean_token_accuracy": 0.830919623374939,
      "num_tokens": 180994135.0,
      "step": 1135
    },
    {
      "epoch": 0.5778229908443541,
      "grad_norm": 1.0666347742080688,
      "learning_rate": 1e-05,
      "loss": 0.5516,
      "mean_token_accuracy": 0.8308299779891968,
      "num_tokens": 181158179.0,
      "step": 1136
    },
    {
      "epoch": 0.5783316378433367,
      "grad_norm": 1.0088244676589966,
      "learning_rate": 1e-05,
      "loss": 0.5192,
      "mean_token_accuracy": 0.8389890193939209,
      "num_tokens": 181316469.0,
      "step": 1137
    },
    {
      "epoch": 0.5788402848423194,
      "grad_norm": 1.0174601078033447,
      "learning_rate": 1e-05,
      "loss": 0.518,
      "mean_token_accuracy": 0.8383738994598389,
      "num_tokens": 181465876.0,
      "step": 1138
    },
    {
      "epoch": 0.5793489318413021,
      "grad_norm": 1.0556000471115112,
      "learning_rate": 1e-05,
      "loss": 0.5127,
      "mean_token_accuracy": 0.8405688405036926,
      "num_tokens": 181626190.0,
      "step": 1139
    },
    {
      "epoch": 0.5798575788402849,
      "grad_norm": 1.082321047782898,
      "learning_rate": 1e-05,
      "loss": 0.586,
      "mean_token_accuracy": 0.8204494714736938,
      "num_tokens": 181772486.0,
      "step": 1140
    },
    {
      "epoch": 0.5803662258392676,
      "grad_norm": 1.0585923194885254,
      "learning_rate": 1e-05,
      "loss": 0.5468,
      "mean_token_accuracy": 0.8332372903823853,
      "num_tokens": 181930828.0,
      "step": 1141
    },
    {
      "epoch": 0.5808748728382502,
      "grad_norm": 1.0120269060134888,
      "learning_rate": 1e-05,
      "loss": 0.5399,
      "mean_token_accuracy": 0.8326765298843384,
      "num_tokens": 182092553.0,
      "step": 1142
    },
    {
      "epoch": 0.5813835198372329,
      "grad_norm": 1.0479254722595215,
      "learning_rate": 1e-05,
      "loss": 0.5443,
      "mean_token_accuracy": 0.8317548036575317,
      "num_tokens": 182242889.0,
      "step": 1143
    },
    {
      "epoch": 0.5818921668362157,
      "grad_norm": 1.0533218383789062,
      "learning_rate": 1e-05,
      "loss": 0.5567,
      "mean_token_accuracy": 0.8292123079299927,
      "num_tokens": 182399236.0,
      "step": 1144
    },
    {
      "epoch": 0.5824008138351984,
      "grad_norm": 1.0289981365203857,
      "learning_rate": 1e-05,
      "loss": 0.564,
      "mean_token_accuracy": 0.827014684677124,
      "num_tokens": 182555200.0,
      "step": 1145
    },
    {
      "epoch": 0.582909460834181,
      "grad_norm": 1.0775690078735352,
      "learning_rate": 1e-05,
      "loss": 0.5442,
      "mean_token_accuracy": 0.8320527076721191,
      "num_tokens": 182711095.0,
      "step": 1146
    },
    {
      "epoch": 0.5834181078331638,
      "grad_norm": 1.0060006380081177,
      "learning_rate": 1e-05,
      "loss": 0.564,
      "mean_token_accuracy": 0.8269708752632141,
      "num_tokens": 182884563.0,
      "step": 1147
    },
    {
      "epoch": 0.5839267548321465,
      "grad_norm": 1.0512561798095703,
      "learning_rate": 1e-05,
      "loss": 0.5789,
      "mean_token_accuracy": 0.8244001865386963,
      "num_tokens": 183045279.0,
      "step": 1148
    },
    {
      "epoch": 0.5844354018311292,
      "grad_norm": 1.0781697034835815,
      "learning_rate": 1e-05,
      "loss": 0.5444,
      "mean_token_accuracy": 0.8334587216377258,
      "num_tokens": 183209431.0,
      "step": 1149
    },
    {
      "epoch": 0.5849440488301119,
      "grad_norm": 1.094925045967102,
      "learning_rate": 1e-05,
      "loss": 0.5202,
      "mean_token_accuracy": 0.8398051261901855,
      "num_tokens": 183377751.0,
      "step": 1150
    },
    {
      "epoch": 0.5854526958290946,
      "grad_norm": 1.044399619102478,
      "learning_rate": 1e-05,
      "loss": 0.5227,
      "mean_token_accuracy": 0.8384654521942139,
      "num_tokens": 183530201.0,
      "step": 1151
    },
    {
      "epoch": 0.5859613428280773,
      "grad_norm": 1.09066641330719,
      "learning_rate": 1e-05,
      "loss": 0.5473,
      "mean_token_accuracy": 0.8314449191093445,
      "num_tokens": 183676743.0,
      "step": 1152
    },
    {
      "epoch": 0.58646998982706,
      "grad_norm": 1.0624159574508667,
      "learning_rate": 1e-05,
      "loss": 0.5284,
      "mean_token_accuracy": 0.8383419513702393,
      "num_tokens": 183838597.0,
      "step": 1153
    },
    {
      "epoch": 0.5869786368260427,
      "grad_norm": 1.0591371059417725,
      "learning_rate": 1e-05,
      "loss": 0.5352,
      "mean_token_accuracy": 0.8353383541107178,
      "num_tokens": 184006491.0,
      "step": 1154
    },
    {
      "epoch": 0.5874872838250255,
      "grad_norm": 1.000448226928711,
      "learning_rate": 1e-05,
      "loss": 0.5334,
      "mean_token_accuracy": 0.8357807993888855,
      "num_tokens": 184170537.0,
      "step": 1155
    },
    {
      "epoch": 0.5879959308240081,
      "grad_norm": 1.1702582836151123,
      "learning_rate": 1e-05,
      "loss": 0.523,
      "mean_token_accuracy": 0.8372135162353516,
      "num_tokens": 184334536.0,
      "step": 1156
    },
    {
      "epoch": 0.5885045778229908,
      "grad_norm": 1.1041123867034912,
      "learning_rate": 1e-05,
      "loss": 0.5201,
      "mean_token_accuracy": 0.835742712020874,
      "num_tokens": 184489719.0,
      "step": 1157
    },
    {
      "epoch": 0.5890132248219736,
      "grad_norm": 1.1316978931427002,
      "learning_rate": 1e-05,
      "loss": 0.5582,
      "mean_token_accuracy": 0.8292644619941711,
      "num_tokens": 184644555.0,
      "step": 1158
    },
    {
      "epoch": 0.5895218718209563,
      "grad_norm": 1.07127046585083,
      "learning_rate": 1e-05,
      "loss": 0.5291,
      "mean_token_accuracy": 0.8366115093231201,
      "num_tokens": 184803771.0,
      "step": 1159
    },
    {
      "epoch": 0.590030518819939,
      "grad_norm": 1.0761888027191162,
      "learning_rate": 1e-05,
      "loss": 0.5475,
      "mean_token_accuracy": 0.8302804231643677,
      "num_tokens": 184956764.0,
      "step": 1160
    },
    {
      "epoch": 0.5905391658189216,
      "grad_norm": 1.1104953289031982,
      "learning_rate": 1e-05,
      "loss": 0.5195,
      "mean_token_accuracy": 0.8388170599937439,
      "num_tokens": 185122654.0,
      "step": 1161
    },
    {
      "epoch": 0.5910478128179044,
      "grad_norm": 1.101953387260437,
      "learning_rate": 1e-05,
      "loss": 0.5535,
      "mean_token_accuracy": 0.8308421969413757,
      "num_tokens": 185278874.0,
      "step": 1162
    },
    {
      "epoch": 0.5915564598168871,
      "grad_norm": 1.1842914819717407,
      "learning_rate": 1e-05,
      "loss": 0.5308,
      "mean_token_accuracy": 0.8350153565406799,
      "num_tokens": 185438919.0,
      "step": 1163
    },
    {
      "epoch": 0.5920651068158698,
      "grad_norm": 1.2792435884475708,
      "learning_rate": 1e-05,
      "loss": 0.5258,
      "mean_token_accuracy": 0.8371274471282959,
      "num_tokens": 185603559.0,
      "step": 1164
    },
    {
      "epoch": 0.5925737538148524,
      "grad_norm": 1.0808415412902832,
      "learning_rate": 1e-05,
      "loss": 0.5209,
      "mean_token_accuracy": 0.8390897512435913,
      "num_tokens": 185745740.0,
      "step": 1165
    },
    {
      "epoch": 0.5930824008138352,
      "grad_norm": 1.0706919431686401,
      "learning_rate": 1e-05,
      "loss": 0.5521,
      "mean_token_accuracy": 0.8303017616271973,
      "num_tokens": 185904158.0,
      "step": 1166
    },
    {
      "epoch": 0.5935910478128179,
      "grad_norm": 1.181048035621643,
      "learning_rate": 1e-05,
      "loss": 0.5384,
      "mean_token_accuracy": 0.8349406123161316,
      "num_tokens": 186062692.0,
      "step": 1167
    },
    {
      "epoch": 0.5940996948118006,
      "grad_norm": 1.0889734029769897,
      "learning_rate": 1e-05,
      "loss": 0.5162,
      "mean_token_accuracy": 0.8389154672622681,
      "num_tokens": 186208990.0,
      "step": 1168
    },
    {
      "epoch": 0.5946083418107834,
      "grad_norm": 1.1245179176330566,
      "learning_rate": 1e-05,
      "loss": 0.531,
      "mean_token_accuracy": 0.8353163003921509,
      "num_tokens": 186353260.0,
      "step": 1169
    },
    {
      "epoch": 0.595116988809766,
      "grad_norm": 1.2074040174484253,
      "learning_rate": 1e-05,
      "loss": 0.5836,
      "mean_token_accuracy": 0.821225643157959,
      "num_tokens": 186527391.0,
      "step": 1170
    },
    {
      "epoch": 0.5956256358087487,
      "grad_norm": 1.1722588539123535,
      "learning_rate": 1e-05,
      "loss": 0.5531,
      "mean_token_accuracy": 0.8324877023696899,
      "num_tokens": 186690975.0,
      "step": 1171
    },
    {
      "epoch": 0.5961342828077314,
      "grad_norm": 1.001542568206787,
      "learning_rate": 1e-05,
      "loss": 0.4979,
      "mean_token_accuracy": 0.8467972874641418,
      "num_tokens": 186851711.0,
      "step": 1172
    },
    {
      "epoch": 0.5966429298067142,
      "grad_norm": 1.0898191928863525,
      "learning_rate": 1e-05,
      "loss": 0.5337,
      "mean_token_accuracy": 0.8319215178489685,
      "num_tokens": 187016950.0,
      "step": 1173
    },
    {
      "epoch": 0.5971515768056969,
      "grad_norm": 1.1995619535446167,
      "learning_rate": 1e-05,
      "loss": 0.55,
      "mean_token_accuracy": 0.8314080238342285,
      "num_tokens": 187158410.0,
      "step": 1174
    },
    {
      "epoch": 0.5976602238046795,
      "grad_norm": 1.0415512323379517,
      "learning_rate": 1e-05,
      "loss": 0.5048,
      "mean_token_accuracy": 0.843203067779541,
      "num_tokens": 187319518.0,
      "step": 1175
    },
    {
      "epoch": 0.5981688708036622,
      "grad_norm": 1.0477491617202759,
      "learning_rate": 1e-05,
      "loss": 0.5288,
      "mean_token_accuracy": 0.8374745845794678,
      "num_tokens": 187492387.0,
      "step": 1176
    },
    {
      "epoch": 0.598677517802645,
      "grad_norm": 1.053478479385376,
      "learning_rate": 1e-05,
      "loss": 0.5652,
      "mean_token_accuracy": 0.8271386623382568,
      "num_tokens": 187660820.0,
      "step": 1177
    },
    {
      "epoch": 0.5991861648016277,
      "grad_norm": 1.1635273694992065,
      "learning_rate": 1e-05,
      "loss": 0.5464,
      "mean_token_accuracy": 0.8312414884567261,
      "num_tokens": 187815200.0,
      "step": 1178
    },
    {
      "epoch": 0.5996948118006104,
      "grad_norm": 1.155785083770752,
      "learning_rate": 1e-05,
      "loss": 0.5212,
      "mean_token_accuracy": 0.8382295370101929,
      "num_tokens": 187976720.0,
      "step": 1179
    },
    {
      "epoch": 0.6002034587995931,
      "grad_norm": 1.0066823959350586,
      "learning_rate": 1e-05,
      "loss": 0.5092,
      "mean_token_accuracy": 0.841545581817627,
      "num_tokens": 188141681.0,
      "step": 1180
    },
    {
      "epoch": 0.6007121057985758,
      "grad_norm": 1.1968673467636108,
      "learning_rate": 1e-05,
      "loss": 0.5348,
      "mean_token_accuracy": 0.8361446857452393,
      "num_tokens": 188306729.0,
      "step": 1181
    },
    {
      "epoch": 0.6012207527975585,
      "grad_norm": 1.034803867340088,
      "learning_rate": 1e-05,
      "loss": 0.5675,
      "mean_token_accuracy": 0.8266303539276123,
      "num_tokens": 188459816.0,
      "step": 1182
    },
    {
      "epoch": 0.6017293997965412,
      "grad_norm": 1.1092537641525269,
      "learning_rate": 1e-05,
      "loss": 0.5556,
      "mean_token_accuracy": 0.8273969292640686,
      "num_tokens": 188624442.0,
      "step": 1183
    },
    {
      "epoch": 0.602238046795524,
      "grad_norm": 1.0602318048477173,
      "learning_rate": 1e-05,
      "loss": 0.5564,
      "mean_token_accuracy": 0.8292548060417175,
      "num_tokens": 188793110.0,
      "step": 1184
    },
    {
      "epoch": 0.6027466937945066,
      "grad_norm": 1.0197434425354004,
      "learning_rate": 1e-05,
      "loss": 0.51,
      "mean_token_accuracy": 0.839547872543335,
      "num_tokens": 188959411.0,
      "step": 1185
    },
    {
      "epoch": 0.6032553407934893,
      "grad_norm": 1.0977286100387573,
      "learning_rate": 1e-05,
      "loss": 0.5326,
      "mean_token_accuracy": 0.8351724147796631,
      "num_tokens": 189123138.0,
      "step": 1186
    },
    {
      "epoch": 0.603763987792472,
      "grad_norm": 1.0537692308425903,
      "learning_rate": 1e-05,
      "loss": 0.529,
      "mean_token_accuracy": 0.836279034614563,
      "num_tokens": 189288712.0,
      "step": 1187
    },
    {
      "epoch": 0.6042726347914548,
      "grad_norm": 1.103000283241272,
      "learning_rate": 1e-05,
      "loss": 0.5238,
      "mean_token_accuracy": 0.838005781173706,
      "num_tokens": 189448908.0,
      "step": 1188
    },
    {
      "epoch": 0.6047812817904374,
      "grad_norm": 1.1983293294906616,
      "learning_rate": 1e-05,
      "loss": 0.5514,
      "mean_token_accuracy": 0.8291587233543396,
      "num_tokens": 189601380.0,
      "step": 1189
    },
    {
      "epoch": 0.6052899287894201,
      "grad_norm": 1.2270739078521729,
      "learning_rate": 1e-05,
      "loss": 0.5507,
      "mean_token_accuracy": 0.8289862871170044,
      "num_tokens": 189760563.0,
      "step": 1190
    },
    {
      "epoch": 0.6057985757884028,
      "grad_norm": 1.155634880065918,
      "learning_rate": 1e-05,
      "loss": 0.5417,
      "mean_token_accuracy": 0.8324066400527954,
      "num_tokens": 189919046.0,
      "step": 1191
    },
    {
      "epoch": 0.6063072227873856,
      "grad_norm": 1.123648762702942,
      "learning_rate": 1e-05,
      "loss": 0.547,
      "mean_token_accuracy": 0.8326415419578552,
      "num_tokens": 190067307.0,
      "step": 1192
    },
    {
      "epoch": 0.6068158697863683,
      "grad_norm": 1.0511082410812378,
      "learning_rate": 1e-05,
      "loss": 0.5268,
      "mean_token_accuracy": 0.8373522758483887,
      "num_tokens": 190216290.0,
      "step": 1193
    },
    {
      "epoch": 0.6073245167853509,
      "grad_norm": 1.1812682151794434,
      "learning_rate": 1e-05,
      "loss": 0.5315,
      "mean_token_accuracy": 0.834931492805481,
      "num_tokens": 190372698.0,
      "step": 1194
    },
    {
      "epoch": 0.6078331637843337,
      "grad_norm": 1.1167006492614746,
      "learning_rate": 1e-05,
      "loss": 0.5295,
      "mean_token_accuracy": 0.8358839154243469,
      "num_tokens": 190528564.0,
      "step": 1195
    },
    {
      "epoch": 0.6083418107833164,
      "grad_norm": 1.1498664617538452,
      "learning_rate": 1e-05,
      "loss": 0.559,
      "mean_token_accuracy": 0.8274659514427185,
      "num_tokens": 190679521.0,
      "step": 1196
    },
    {
      "epoch": 0.6088504577822991,
      "grad_norm": 1.1137135028839111,
      "learning_rate": 1e-05,
      "loss": 0.5435,
      "mean_token_accuracy": 0.8332474827766418,
      "num_tokens": 190826754.0,
      "step": 1197
    },
    {
      "epoch": 0.6093591047812817,
      "grad_norm": 1.2709776163101196,
      "learning_rate": 1e-05,
      "loss": 0.5358,
      "mean_token_accuracy": 0.8357667326927185,
      "num_tokens": 190990350.0,
      "step": 1198
    },
    {
      "epoch": 0.6098677517802645,
      "grad_norm": 1.2216440439224243,
      "learning_rate": 1e-05,
      "loss": 0.5632,
      "mean_token_accuracy": 0.827555239200592,
      "num_tokens": 191150121.0,
      "step": 1199
    },
    {
      "epoch": 0.6103763987792472,
      "grad_norm": 1.1574770212173462,
      "learning_rate": 1e-05,
      "loss": 0.5383,
      "mean_token_accuracy": 0.8342841267585754,
      "num_tokens": 191298356.0,
      "step": 1200
    },
    {
      "epoch": 0.6108850457782299,
      "grad_norm": 1.108181357383728,
      "learning_rate": 1e-05,
      "loss": 0.5168,
      "mean_token_accuracy": 0.8400946855545044,
      "num_tokens": 191454560.0,
      "step": 1201
    },
    {
      "epoch": 0.6113936927772126,
      "grad_norm": 1.1643569469451904,
      "learning_rate": 1e-05,
      "loss": 0.4993,
      "mean_token_accuracy": 0.8451148867607117,
      "num_tokens": 191607061.0,
      "step": 1202
    },
    {
      "epoch": 0.6119023397761953,
      "grad_norm": 1.0500165224075317,
      "learning_rate": 1e-05,
      "loss": 0.536,
      "mean_token_accuracy": 0.8340716361999512,
      "num_tokens": 191764683.0,
      "step": 1203
    },
    {
      "epoch": 0.612410986775178,
      "grad_norm": 1.1072869300842285,
      "learning_rate": 1e-05,
      "loss": 0.5273,
      "mean_token_accuracy": 0.8370859622955322,
      "num_tokens": 191920982.0,
      "step": 1204
    },
    {
      "epoch": 0.6129196337741607,
      "grad_norm": 1.034971833229065,
      "learning_rate": 1e-05,
      "loss": 0.5239,
      "mean_token_accuracy": 0.8371504545211792,
      "num_tokens": 192084464.0,
      "step": 1205
    },
    {
      "epoch": 0.6134282807731435,
      "grad_norm": 1.0584032535552979,
      "learning_rate": 1e-05,
      "loss": 0.5443,
      "mean_token_accuracy": 0.8337501883506775,
      "num_tokens": 192244544.0,
      "step": 1206
    },
    {
      "epoch": 0.6139369277721262,
      "grad_norm": 1.1733242273330688,
      "learning_rate": 1e-05,
      "loss": 0.5212,
      "mean_token_accuracy": 0.8383514881134033,
      "num_tokens": 192398895.0,
      "step": 1207
    },
    {
      "epoch": 0.6144455747711088,
      "grad_norm": 1.1394904851913452,
      "learning_rate": 1e-05,
      "loss": 0.5679,
      "mean_token_accuracy": 0.8283772468566895,
      "num_tokens": 192568952.0,
      "step": 1208
    },
    {
      "epoch": 0.6149542217700915,
      "grad_norm": 1.1953922510147095,
      "learning_rate": 1e-05,
      "loss": 0.5564,
      "mean_token_accuracy": 0.8289297819137573,
      "num_tokens": 192726852.0,
      "step": 1209
    },
    {
      "epoch": 0.6154628687690743,
      "grad_norm": 1.1235586404800415,
      "learning_rate": 1e-05,
      "loss": 0.4998,
      "mean_token_accuracy": 0.8443677425384521,
      "num_tokens": 192871340.0,
      "step": 1210
    },
    {
      "epoch": 0.615971515768057,
      "grad_norm": 1.1388280391693115,
      "learning_rate": 1e-05,
      "loss": 0.5144,
      "mean_token_accuracy": 0.8400600552558899,
      "num_tokens": 193042111.0,
      "step": 1211
    },
    {
      "epoch": 0.6164801627670397,
      "grad_norm": 1.1637824773788452,
      "learning_rate": 1e-05,
      "loss": 0.5454,
      "mean_token_accuracy": 0.8325772881507874,
      "num_tokens": 193194025.0,
      "step": 1212
    },
    {
      "epoch": 0.6169888097660223,
      "grad_norm": 1.2491023540496826,
      "learning_rate": 1e-05,
      "loss": 0.5025,
      "mean_token_accuracy": 0.8449327945709229,
      "num_tokens": 193349718.0,
      "step": 1213
    },
    {
      "epoch": 0.6174974567650051,
      "grad_norm": 1.1179122924804688,
      "learning_rate": 1e-05,
      "loss": 0.5388,
      "mean_token_accuracy": 0.8341668844223022,
      "num_tokens": 193505599.0,
      "step": 1214
    },
    {
      "epoch": 0.6180061037639878,
      "grad_norm": 1.0245755910873413,
      "learning_rate": 1e-05,
      "loss": 0.5785,
      "mean_token_accuracy": 0.8252415060997009,
      "num_tokens": 193681994.0,
      "step": 1215
    },
    {
      "epoch": 0.6185147507629705,
      "grad_norm": 1.0279556512832642,
      "learning_rate": 1e-05,
      "loss": 0.5564,
      "mean_token_accuracy": 0.8292478919029236,
      "num_tokens": 193846107.0,
      "step": 1216
    },
    {
      "epoch": 0.6190233977619533,
      "grad_norm": 1.0754578113555908,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.835884690284729,
      "num_tokens": 194001101.0,
      "step": 1217
    },
    {
      "epoch": 0.6195320447609359,
      "grad_norm": 0.9642245173454285,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8447218537330627,
      "num_tokens": 194162440.0,
      "step": 1218
    },
    {
      "epoch": 0.6200406917599186,
      "grad_norm": 1.0722028017044067,
      "learning_rate": 1e-05,
      "loss": 0.5374,
      "mean_token_accuracy": 0.8334408402442932,
      "num_tokens": 194316830.0,
      "step": 1219
    },
    {
      "epoch": 0.6205493387589013,
      "grad_norm": 0.9881809949874878,
      "learning_rate": 1e-05,
      "loss": 0.5179,
      "mean_token_accuracy": 0.8374440670013428,
      "num_tokens": 194471827.0,
      "step": 1220
    },
    {
      "epoch": 0.6210579857578841,
      "grad_norm": 1.063515305519104,
      "learning_rate": 1e-05,
      "loss": 0.5354,
      "mean_token_accuracy": 0.833842396736145,
      "num_tokens": 194631877.0,
      "step": 1221
    },
    {
      "epoch": 0.6215666327568667,
      "grad_norm": 0.9928879737854004,
      "learning_rate": 1e-05,
      "loss": 0.53,
      "mean_token_accuracy": 0.8351453542709351,
      "num_tokens": 194798176.0,
      "step": 1222
    },
    {
      "epoch": 0.6220752797558494,
      "grad_norm": 1.0375672578811646,
      "learning_rate": 1e-05,
      "loss": 0.5402,
      "mean_token_accuracy": 0.8339008688926697,
      "num_tokens": 194954556.0,
      "step": 1223
    },
    {
      "epoch": 0.6225839267548321,
      "grad_norm": 1.1741032600402832,
      "learning_rate": 1e-05,
      "loss": 0.5645,
      "mean_token_accuracy": 0.8256007432937622,
      "num_tokens": 195115010.0,
      "step": 1224
    },
    {
      "epoch": 0.6230925737538149,
      "grad_norm": 1.037577748298645,
      "learning_rate": 1e-05,
      "loss": 0.5128,
      "mean_token_accuracy": 0.8424239754676819,
      "num_tokens": 195277332.0,
      "step": 1225
    },
    {
      "epoch": 0.6236012207527976,
      "grad_norm": 1.1055760383605957,
      "learning_rate": 1e-05,
      "loss": 0.5151,
      "mean_token_accuracy": 0.8401551246643066,
      "num_tokens": 195442166.0,
      "step": 1226
    },
    {
      "epoch": 0.6241098677517802,
      "grad_norm": 0.9993090629577637,
      "learning_rate": 1e-05,
      "loss": 0.5111,
      "mean_token_accuracy": 0.8418375849723816,
      "num_tokens": 195598203.0,
      "step": 1227
    },
    {
      "epoch": 0.624618514750763,
      "grad_norm": 1.062853217124939,
      "learning_rate": 1e-05,
      "loss": 0.5156,
      "mean_token_accuracy": 0.8409103155136108,
      "num_tokens": 195755945.0,
      "step": 1228
    },
    {
      "epoch": 0.6251271617497457,
      "grad_norm": 0.9994860291481018,
      "learning_rate": 1e-05,
      "loss": 0.5165,
      "mean_token_accuracy": 0.8400459289550781,
      "num_tokens": 195923721.0,
      "step": 1229
    },
    {
      "epoch": 0.6256358087487284,
      "grad_norm": 1.10517418384552,
      "learning_rate": 1e-05,
      "loss": 0.5064,
      "mean_token_accuracy": 0.8433425426483154,
      "num_tokens": 196076642.0,
      "step": 1230
    },
    {
      "epoch": 0.626144455747711,
      "grad_norm": 1.0957393646240234,
      "learning_rate": 1e-05,
      "loss": 0.5347,
      "mean_token_accuracy": 0.8352852463722229,
      "num_tokens": 196232622.0,
      "step": 1231
    },
    {
      "epoch": 0.6266531027466938,
      "grad_norm": 1.0993536710739136,
      "learning_rate": 1e-05,
      "loss": 0.5545,
      "mean_token_accuracy": 0.8281955122947693,
      "num_tokens": 196384640.0,
      "step": 1232
    },
    {
      "epoch": 0.6271617497456765,
      "grad_norm": 1.0579915046691895,
      "learning_rate": 1e-05,
      "loss": 0.5706,
      "mean_token_accuracy": 0.8254827857017517,
      "num_tokens": 196546059.0,
      "step": 1233
    },
    {
      "epoch": 0.6276703967446592,
      "grad_norm": 1.0779695510864258,
      "learning_rate": 1e-05,
      "loss": 0.5603,
      "mean_token_accuracy": 0.8296513557434082,
      "num_tokens": 196709416.0,
      "step": 1234
    },
    {
      "epoch": 0.6281790437436419,
      "grad_norm": 0.9989771246910095,
      "learning_rate": 1e-05,
      "loss": 0.542,
      "mean_token_accuracy": 0.8338413834571838,
      "num_tokens": 196874595.0,
      "step": 1235
    },
    {
      "epoch": 0.6286876907426246,
      "grad_norm": 1.0544723272323608,
      "learning_rate": 1e-05,
      "loss": 0.5132,
      "mean_token_accuracy": 0.8396172523498535,
      "num_tokens": 197038688.0,
      "step": 1236
    },
    {
      "epoch": 0.6291963377416073,
      "grad_norm": 1.1366897821426392,
      "learning_rate": 1e-05,
      "loss": 0.5745,
      "mean_token_accuracy": 0.8230605125427246,
      "num_tokens": 197192465.0,
      "step": 1237
    },
    {
      "epoch": 0.62970498474059,
      "grad_norm": 1.061204433441162,
      "learning_rate": 1e-05,
      "loss": 0.5529,
      "mean_token_accuracy": 0.8329615592956543,
      "num_tokens": 197361082.0,
      "step": 1238
    },
    {
      "epoch": 0.6302136317395728,
      "grad_norm": 1.0565414428710938,
      "learning_rate": 1e-05,
      "loss": 0.4889,
      "mean_token_accuracy": 0.8487517833709717,
      "num_tokens": 197512368.0,
      "step": 1239
    },
    {
      "epoch": 0.6307222787385555,
      "grad_norm": 1.0725106000900269,
      "learning_rate": 1e-05,
      "loss": 0.5561,
      "mean_token_accuracy": 0.829352617263794,
      "num_tokens": 197654672.0,
      "step": 1240
    },
    {
      "epoch": 0.6312309257375381,
      "grad_norm": 1.1748433113098145,
      "learning_rate": 1e-05,
      "loss": 0.5574,
      "mean_token_accuracy": 0.8295996189117432,
      "num_tokens": 197816279.0,
      "step": 1241
    },
    {
      "epoch": 0.6317395727365208,
      "grad_norm": 1.0469516515731812,
      "learning_rate": 1e-05,
      "loss": 0.5124,
      "mean_token_accuracy": 0.8417311906814575,
      "num_tokens": 197972971.0,
      "step": 1242
    },
    {
      "epoch": 0.6322482197355036,
      "grad_norm": 1.1066620349884033,
      "learning_rate": 1e-05,
      "loss": 0.526,
      "mean_token_accuracy": 0.8364421129226685,
      "num_tokens": 198125745.0,
      "step": 1243
    },
    {
      "epoch": 0.6327568667344863,
      "grad_norm": 0.9559982419013977,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8472656607627869,
      "num_tokens": 198288270.0,
      "step": 1244
    },
    {
      "epoch": 0.633265513733469,
      "grad_norm": 1.037071943283081,
      "learning_rate": 1e-05,
      "loss": 0.5423,
      "mean_token_accuracy": 0.8322509527206421,
      "num_tokens": 198445798.0,
      "step": 1245
    },
    {
      "epoch": 0.6337741607324516,
      "grad_norm": 1.0956902503967285,
      "learning_rate": 1e-05,
      "loss": 0.5091,
      "mean_token_accuracy": 0.8409451246261597,
      "num_tokens": 198600564.0,
      "step": 1246
    },
    {
      "epoch": 0.6342828077314344,
      "grad_norm": 0.9643290638923645,
      "learning_rate": 1e-05,
      "loss": 0.507,
      "mean_token_accuracy": 0.843378484249115,
      "num_tokens": 198768145.0,
      "step": 1247
    },
    {
      "epoch": 0.6347914547304171,
      "grad_norm": 1.114691138267517,
      "learning_rate": 1e-05,
      "loss": 0.5142,
      "mean_token_accuracy": 0.8397768139839172,
      "num_tokens": 198919816.0,
      "step": 1248
    },
    {
      "epoch": 0.6353001017293998,
      "grad_norm": 1.0919520854949951,
      "learning_rate": 1e-05,
      "loss": 0.5617,
      "mean_token_accuracy": 0.8267562389373779,
      "num_tokens": 199084288.0,
      "step": 1249
    },
    {
      "epoch": 0.6358087487283826,
      "grad_norm": 1.0663131475448608,
      "learning_rate": 1e-05,
      "loss": 0.5508,
      "mean_token_accuracy": 0.830141544342041,
      "num_tokens": 199234621.0,
      "step": 1250
    },
    {
      "epoch": 0.6363173957273652,
      "grad_norm": 0.9842988848686218,
      "learning_rate": 1e-05,
      "loss": 0.5205,
      "mean_token_accuracy": 0.8383627533912659,
      "num_tokens": 199415992.0,
      "step": 1251
    },
    {
      "epoch": 0.6368260427263479,
      "grad_norm": 1.0851852893829346,
      "learning_rate": 1e-05,
      "loss": 0.5415,
      "mean_token_accuracy": 0.8332940340042114,
      "num_tokens": 199581880.0,
      "step": 1252
    },
    {
      "epoch": 0.6373346897253306,
      "grad_norm": 1.1240712404251099,
      "learning_rate": 1e-05,
      "loss": 0.5182,
      "mean_token_accuracy": 0.8378025889396667,
      "num_tokens": 199728425.0,
      "step": 1253
    },
    {
      "epoch": 0.6378433367243134,
      "grad_norm": 1.100454568862915,
      "learning_rate": 1e-05,
      "loss": 0.5766,
      "mean_token_accuracy": 0.8229030966758728,
      "num_tokens": 199887035.0,
      "step": 1254
    },
    {
      "epoch": 0.638351983723296,
      "grad_norm": 1.128618597984314,
      "learning_rate": 1e-05,
      "loss": 0.5494,
      "mean_token_accuracy": 0.8319924473762512,
      "num_tokens": 200042873.0,
      "step": 1255
    },
    {
      "epoch": 0.6388606307222787,
      "grad_norm": 1.069466471672058,
      "learning_rate": 1e-05,
      "loss": 0.5192,
      "mean_token_accuracy": 0.8399227857589722,
      "num_tokens": 200218153.0,
      "step": 1256
    },
    {
      "epoch": 0.6393692777212614,
      "grad_norm": 1.1159545183181763,
      "learning_rate": 1e-05,
      "loss": 0.5434,
      "mean_token_accuracy": 0.8340426683425903,
      "num_tokens": 200375972.0,
      "step": 1257
    },
    {
      "epoch": 0.6398779247202442,
      "grad_norm": 1.0914868116378784,
      "learning_rate": 1e-05,
      "loss": 0.5517,
      "mean_token_accuracy": 0.8304837942123413,
      "num_tokens": 200538925.0,
      "step": 1258
    },
    {
      "epoch": 0.6403865717192269,
      "grad_norm": 1.016735315322876,
      "learning_rate": 1e-05,
      "loss": 0.5222,
      "mean_token_accuracy": 0.8377134799957275,
      "num_tokens": 200702936.0,
      "step": 1259
    },
    {
      "epoch": 0.6408952187182095,
      "grad_norm": 1.1668028831481934,
      "learning_rate": 1e-05,
      "loss": 0.5199,
      "mean_token_accuracy": 0.8381738662719727,
      "num_tokens": 200852605.0,
      "step": 1260
    },
    {
      "epoch": 0.6414038657171923,
      "grad_norm": 1.106713891029358,
      "learning_rate": 1e-05,
      "loss": 0.5192,
      "mean_token_accuracy": 0.8384389877319336,
      "num_tokens": 201011038.0,
      "step": 1261
    },
    {
      "epoch": 0.641912512716175,
      "grad_norm": 1.081728458404541,
      "learning_rate": 1e-05,
      "loss": 0.5503,
      "mean_token_accuracy": 0.8293412327766418,
      "num_tokens": 201171353.0,
      "step": 1262
    },
    {
      "epoch": 0.6424211597151577,
      "grad_norm": 1.0619851350784302,
      "learning_rate": 1e-05,
      "loss": 0.5578,
      "mean_token_accuracy": 0.8283439874649048,
      "num_tokens": 201335064.0,
      "step": 1263
    },
    {
      "epoch": 0.6429298067141404,
      "grad_norm": 1.0915756225585938,
      "learning_rate": 1e-05,
      "loss": 0.5157,
      "mean_token_accuracy": 0.8394124507904053,
      "num_tokens": 201496911.0,
      "step": 1264
    },
    {
      "epoch": 0.6434384537131231,
      "grad_norm": 1.0689359903335571,
      "learning_rate": 1e-05,
      "loss": 0.5414,
      "mean_token_accuracy": 0.8329365253448486,
      "num_tokens": 201650881.0,
      "step": 1265
    },
    {
      "epoch": 0.6439471007121058,
      "grad_norm": 1.0703952312469482,
      "learning_rate": 1e-05,
      "loss": 0.5636,
      "mean_token_accuracy": 0.8266373872756958,
      "num_tokens": 201822744.0,
      "step": 1266
    },
    {
      "epoch": 0.6444557477110885,
      "grad_norm": 0.9987941980361938,
      "learning_rate": 1e-05,
      "loss": 0.52,
      "mean_token_accuracy": 0.8399134278297424,
      "num_tokens": 201988138.0,
      "step": 1267
    },
    {
      "epoch": 0.6449643947100712,
      "grad_norm": 0.9540848731994629,
      "learning_rate": 1e-05,
      "loss": 0.5188,
      "mean_token_accuracy": 0.8398693203926086,
      "num_tokens": 202163172.0,
      "step": 1268
    },
    {
      "epoch": 0.645473041709054,
      "grad_norm": 1.0938210487365723,
      "learning_rate": 1e-05,
      "loss": 0.5251,
      "mean_token_accuracy": 0.8377525806427002,
      "num_tokens": 202330354.0,
      "step": 1269
    },
    {
      "epoch": 0.6459816887080366,
      "grad_norm": 1.0494672060012817,
      "learning_rate": 1e-05,
      "loss": 0.5443,
      "mean_token_accuracy": 0.8317760229110718,
      "num_tokens": 202477704.0,
      "step": 1270
    },
    {
      "epoch": 0.6464903357070193,
      "grad_norm": 1.0470654964447021,
      "learning_rate": 1e-05,
      "loss": 0.5715,
      "mean_token_accuracy": 0.8244767785072327,
      "num_tokens": 202638535.0,
      "step": 1271
    },
    {
      "epoch": 0.646998982706002,
      "grad_norm": 1.0278775691986084,
      "learning_rate": 1e-05,
      "loss": 0.5526,
      "mean_token_accuracy": 0.8294790983200073,
      "num_tokens": 202799169.0,
      "step": 1272
    },
    {
      "epoch": 0.6475076297049848,
      "grad_norm": 1.0553829669952393,
      "learning_rate": 1e-05,
      "loss": 0.5316,
      "mean_token_accuracy": 0.8347764015197754,
      "num_tokens": 202948731.0,
      "step": 1273
    },
    {
      "epoch": 0.6480162767039674,
      "grad_norm": 1.0590755939483643,
      "learning_rate": 1e-05,
      "loss": 0.5301,
      "mean_token_accuracy": 0.8380409479141235,
      "num_tokens": 203103019.0,
      "step": 1274
    },
    {
      "epoch": 0.6485249237029501,
      "grad_norm": 1.1748765707015991,
      "learning_rate": 1e-05,
      "loss": 0.5641,
      "mean_token_accuracy": 0.8264802694320679,
      "num_tokens": 203258243.0,
      "step": 1275
    },
    {
      "epoch": 0.6490335707019329,
      "grad_norm": 0.9675571918487549,
      "learning_rate": 1e-05,
      "loss": 0.5274,
      "mean_token_accuracy": 0.8368325233459473,
      "num_tokens": 203417686.0,
      "step": 1276
    },
    {
      "epoch": 0.6495422177009156,
      "grad_norm": 1.0740697383880615,
      "learning_rate": 1e-05,
      "loss": 0.5338,
      "mean_token_accuracy": 0.8351479768753052,
      "num_tokens": 203581925.0,
      "step": 1277
    },
    {
      "epoch": 0.6500508646998983,
      "grad_norm": 1.079943299293518,
      "learning_rate": 1e-05,
      "loss": 0.5607,
      "mean_token_accuracy": 0.8276282548904419,
      "num_tokens": 203737547.0,
      "step": 1278
    },
    {
      "epoch": 0.6505595116988809,
      "grad_norm": 1.1141538619995117,
      "learning_rate": 1e-05,
      "loss": 0.5272,
      "mean_token_accuracy": 0.8363447189331055,
      "num_tokens": 203886501.0,
      "step": 1279
    },
    {
      "epoch": 0.6510681586978637,
      "grad_norm": 1.2464641332626343,
      "learning_rate": 1e-05,
      "loss": 0.5374,
      "mean_token_accuracy": 0.8330584764480591,
      "num_tokens": 204040525.0,
      "step": 1280
    },
    {
      "epoch": 0.6515768056968464,
      "grad_norm": 1.1051416397094727,
      "learning_rate": 1e-05,
      "loss": 0.5219,
      "mean_token_accuracy": 0.8374311923980713,
      "num_tokens": 204205921.0,
      "step": 1281
    },
    {
      "epoch": 0.6520854526958291,
      "grad_norm": 1.125686764717102,
      "learning_rate": 1e-05,
      "loss": 0.5151,
      "mean_token_accuracy": 0.8404279947280884,
      "num_tokens": 204371623.0,
      "step": 1282
    },
    {
      "epoch": 0.6525940996948117,
      "grad_norm": 1.036682367324829,
      "learning_rate": 1e-05,
      "loss": 0.5175,
      "mean_token_accuracy": 0.8396049737930298,
      "num_tokens": 204531555.0,
      "step": 1283
    },
    {
      "epoch": 0.6531027466937945,
      "grad_norm": 1.085217833518982,
      "learning_rate": 1e-05,
      "loss": 0.5249,
      "mean_token_accuracy": 0.8371258974075317,
      "num_tokens": 204695860.0,
      "step": 1284
    },
    {
      "epoch": 0.6536113936927772,
      "grad_norm": 1.0670760869979858,
      "learning_rate": 1e-05,
      "loss": 0.5212,
      "mean_token_accuracy": 0.8381375074386597,
      "num_tokens": 204849971.0,
      "step": 1285
    },
    {
      "epoch": 0.6541200406917599,
      "grad_norm": 0.9454727172851562,
      "learning_rate": 1e-05,
      "loss": 0.4977,
      "mean_token_accuracy": 0.8452643752098083,
      "num_tokens": 205022471.0,
      "step": 1286
    },
    {
      "epoch": 0.6546286876907427,
      "grad_norm": 1.0735646486282349,
      "learning_rate": 1e-05,
      "loss": 0.5106,
      "mean_token_accuracy": 0.8420166969299316,
      "num_tokens": 205191250.0,
      "step": 1287
    },
    {
      "epoch": 0.6551373346897253,
      "grad_norm": 1.1079990863800049,
      "learning_rate": 1e-05,
      "loss": 0.5513,
      "mean_token_accuracy": 0.828972578048706,
      "num_tokens": 205338248.0,
      "step": 1288
    },
    {
      "epoch": 0.655645981688708,
      "grad_norm": 1.0452680587768555,
      "learning_rate": 1e-05,
      "loss": 0.5566,
      "mean_token_accuracy": 0.8274991512298584,
      "num_tokens": 205510777.0,
      "step": 1289
    },
    {
      "epoch": 0.6561546286876907,
      "grad_norm": 1.0797021389007568,
      "learning_rate": 1e-05,
      "loss": 0.5465,
      "mean_token_accuracy": 0.8306173086166382,
      "num_tokens": 205654689.0,
      "step": 1290
    },
    {
      "epoch": 0.6566632756866735,
      "grad_norm": 0.9831796884536743,
      "learning_rate": 1e-05,
      "loss": 0.5565,
      "mean_token_accuracy": 0.8293613195419312,
      "num_tokens": 205818547.0,
      "step": 1291
    },
    {
      "epoch": 0.6571719226856562,
      "grad_norm": 1.095979928970337,
      "learning_rate": 1e-05,
      "loss": 0.5728,
      "mean_token_accuracy": 0.8247345685958862,
      "num_tokens": 205984399.0,
      "step": 1292
    },
    {
      "epoch": 0.6576805696846388,
      "grad_norm": 1.0510746240615845,
      "learning_rate": 1e-05,
      "loss": 0.5789,
      "mean_token_accuracy": 0.8231319189071655,
      "num_tokens": 206148673.0,
      "step": 1293
    },
    {
      "epoch": 0.6581892166836215,
      "grad_norm": 1.2273945808410645,
      "learning_rate": 1e-05,
      "loss": 0.5174,
      "mean_token_accuracy": 0.8394807577133179,
      "num_tokens": 206308245.0,
      "step": 1294
    },
    {
      "epoch": 0.6586978636826043,
      "grad_norm": 1.0817188024520874,
      "learning_rate": 1e-05,
      "loss": 0.4983,
      "mean_token_accuracy": 0.8441655039787292,
      "num_tokens": 206471247.0,
      "step": 1295
    },
    {
      "epoch": 0.659206510681587,
      "grad_norm": 1.0500448942184448,
      "learning_rate": 1e-05,
      "loss": 0.5385,
      "mean_token_accuracy": 0.8328456282615662,
      "num_tokens": 206613022.0,
      "step": 1296
    },
    {
      "epoch": 0.6597151576805697,
      "grad_norm": 1.0789843797683716,
      "learning_rate": 1e-05,
      "loss": 0.5426,
      "mean_token_accuracy": 0.8310835957527161,
      "num_tokens": 206781121.0,
      "step": 1297
    },
    {
      "epoch": 0.6602238046795524,
      "grad_norm": 1.0747429132461548,
      "learning_rate": 1e-05,
      "loss": 0.5464,
      "mean_token_accuracy": 0.8310299515724182,
      "num_tokens": 206935916.0,
      "step": 1298
    },
    {
      "epoch": 0.6607324516785351,
      "grad_norm": 1.1076595783233643,
      "learning_rate": 1e-05,
      "loss": 0.4949,
      "mean_token_accuracy": 0.8450763821601868,
      "num_tokens": 207089114.0,
      "step": 1299
    },
    {
      "epoch": 0.6612410986775178,
      "grad_norm": 1.0206778049468994,
      "learning_rate": 1e-05,
      "loss": 0.5213,
      "mean_token_accuracy": 0.8379412293434143,
      "num_tokens": 207249857.0,
      "step": 1300
    },
    {
      "epoch": 0.6617497456765005,
      "grad_norm": 1.0830273628234863,
      "learning_rate": 1e-05,
      "loss": 0.5693,
      "mean_token_accuracy": 0.8260539174079895,
      "num_tokens": 207415519.0,
      "step": 1301
    },
    {
      "epoch": 0.6622583926754833,
      "grad_norm": 1.032319188117981,
      "learning_rate": 1e-05,
      "loss": 0.5331,
      "mean_token_accuracy": 0.8347049951553345,
      "num_tokens": 207585946.0,
      "step": 1302
    },
    {
      "epoch": 0.6627670396744659,
      "grad_norm": 1.0059924125671387,
      "learning_rate": 1e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8431366682052612,
      "num_tokens": 207742283.0,
      "step": 1303
    },
    {
      "epoch": 0.6632756866734486,
      "grad_norm": 1.0041545629501343,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.8438891768455505,
      "num_tokens": 207905907.0,
      "step": 1304
    },
    {
      "epoch": 0.6637843336724313,
      "grad_norm": 1.0187472105026245,
      "learning_rate": 1e-05,
      "loss": 0.5053,
      "mean_token_accuracy": 0.8417147397994995,
      "num_tokens": 208066296.0,
      "step": 1305
    },
    {
      "epoch": 0.6642929806714141,
      "grad_norm": 1.0259668827056885,
      "learning_rate": 1e-05,
      "loss": 0.5098,
      "mean_token_accuracy": 0.8407272100448608,
      "num_tokens": 208221915.0,
      "step": 1306
    },
    {
      "epoch": 0.6648016276703967,
      "grad_norm": 1.1533657312393188,
      "learning_rate": 1e-05,
      "loss": 0.4846,
      "mean_token_accuracy": 0.8486118316650391,
      "num_tokens": 208374510.0,
      "step": 1307
    },
    {
      "epoch": 0.6653102746693794,
      "grad_norm": 1.092882513999939,
      "learning_rate": 1e-05,
      "loss": 0.5408,
      "mean_token_accuracy": 0.8338409662246704,
      "num_tokens": 208541953.0,
      "step": 1308
    },
    {
      "epoch": 0.6658189216683622,
      "grad_norm": 1.1926003694534302,
      "learning_rate": 1e-05,
      "loss": 0.5219,
      "mean_token_accuracy": 0.8373793959617615,
      "num_tokens": 208698721.0,
      "step": 1309
    },
    {
      "epoch": 0.6663275686673449,
      "grad_norm": 1.087581992149353,
      "learning_rate": 1e-05,
      "loss": 0.5256,
      "mean_token_accuracy": 0.8379611968994141,
      "num_tokens": 208855549.0,
      "step": 1310
    },
    {
      "epoch": 0.6668362156663276,
      "grad_norm": 1.1358734369277954,
      "learning_rate": 1e-05,
      "loss": 0.5506,
      "mean_token_accuracy": 0.8312427997589111,
      "num_tokens": 209015229.0,
      "step": 1311
    },
    {
      "epoch": 0.6673448626653102,
      "grad_norm": 1.130253791809082,
      "learning_rate": 1e-05,
      "loss": 0.54,
      "mean_token_accuracy": 0.8337419033050537,
      "num_tokens": 209167815.0,
      "step": 1312
    },
    {
      "epoch": 0.667853509664293,
      "grad_norm": 0.9735279083251953,
      "learning_rate": 1e-05,
      "loss": 0.5478,
      "mean_token_accuracy": 0.8318684101104736,
      "num_tokens": 209339913.0,
      "step": 1313
    },
    {
      "epoch": 0.6683621566632757,
      "grad_norm": 1.1487743854522705,
      "learning_rate": 1e-05,
      "loss": 0.5725,
      "mean_token_accuracy": 0.8254092931747437,
      "num_tokens": 209506826.0,
      "step": 1314
    },
    {
      "epoch": 0.6688708036622584,
      "grad_norm": 1.1382986307144165,
      "learning_rate": 1e-05,
      "loss": 0.5195,
      "mean_token_accuracy": 0.8394043445587158,
      "num_tokens": 209658534.0,
      "step": 1315
    },
    {
      "epoch": 0.669379450661241,
      "grad_norm": 1.1165403127670288,
      "learning_rate": 1e-05,
      "loss": 0.5122,
      "mean_token_accuracy": 0.8393932580947876,
      "num_tokens": 209810131.0,
      "step": 1316
    },
    {
      "epoch": 0.6698880976602238,
      "grad_norm": 1.0738933086395264,
      "learning_rate": 1e-05,
      "loss": 0.4967,
      "mean_token_accuracy": 0.8456311225891113,
      "num_tokens": 209972848.0,
      "step": 1317
    },
    {
      "epoch": 0.6703967446592065,
      "grad_norm": 1.0726091861724854,
      "learning_rate": 1e-05,
      "loss": 0.5678,
      "mean_token_accuracy": 0.8266811370849609,
      "num_tokens": 210123131.0,
      "step": 1318
    },
    {
      "epoch": 0.6709053916581892,
      "grad_norm": 1.0094369649887085,
      "learning_rate": 1e-05,
      "loss": 0.5373,
      "mean_token_accuracy": 0.835343599319458,
      "num_tokens": 210281302.0,
      "step": 1319
    },
    {
      "epoch": 0.671414038657172,
      "grad_norm": 1.08613920211792,
      "learning_rate": 1e-05,
      "loss": 0.5162,
      "mean_token_accuracy": 0.8415795564651489,
      "num_tokens": 210444300.0,
      "step": 1320
    },
    {
      "epoch": 0.6719226856561547,
      "grad_norm": 1.0181106328964233,
      "learning_rate": 1e-05,
      "loss": 0.5228,
      "mean_token_accuracy": 0.8374539613723755,
      "num_tokens": 210610063.0,
      "step": 1321
    },
    {
      "epoch": 0.6724313326551373,
      "grad_norm": 1.0776914358139038,
      "learning_rate": 1e-05,
      "loss": 0.5356,
      "mean_token_accuracy": 0.8327101469039917,
      "num_tokens": 210771386.0,
      "step": 1322
    },
    {
      "epoch": 0.67293997965412,
      "grad_norm": 1.0810972452163696,
      "learning_rate": 1e-05,
      "loss": 0.5113,
      "mean_token_accuracy": 0.8414871692657471,
      "num_tokens": 210919989.0,
      "step": 1323
    },
    {
      "epoch": 0.6734486266531028,
      "grad_norm": 1.2551976442337036,
      "learning_rate": 1e-05,
      "loss": 0.5499,
      "mean_token_accuracy": 0.831946611404419,
      "num_tokens": 211073799.0,
      "step": 1324
    },
    {
      "epoch": 0.6739572736520855,
      "grad_norm": 1.0054110288619995,
      "learning_rate": 1e-05,
      "loss": 0.5276,
      "mean_token_accuracy": 0.8376283645629883,
      "num_tokens": 211241360.0,
      "step": 1325
    },
    {
      "epoch": 0.6744659206510681,
      "grad_norm": 1.1998662948608398,
      "learning_rate": 1e-05,
      "loss": 0.5513,
      "mean_token_accuracy": 0.8302354216575623,
      "num_tokens": 211407370.0,
      "step": 1326
    },
    {
      "epoch": 0.6749745676500508,
      "grad_norm": 0.9885632991790771,
      "learning_rate": 1e-05,
      "loss": 0.5349,
      "mean_token_accuracy": 0.8335657119750977,
      "num_tokens": 211581491.0,
      "step": 1327
    },
    {
      "epoch": 0.6754832146490336,
      "grad_norm": 1.1440314054489136,
      "learning_rate": 1e-05,
      "loss": 0.5158,
      "mean_token_accuracy": 0.8405625820159912,
      "num_tokens": 211739189.0,
      "step": 1328
    },
    {
      "epoch": 0.6759918616480163,
      "grad_norm": 1.19303297996521,
      "learning_rate": 1e-05,
      "loss": 0.5386,
      "mean_token_accuracy": 0.8338267803192139,
      "num_tokens": 211893052.0,
      "step": 1329
    },
    {
      "epoch": 0.676500508646999,
      "grad_norm": 0.9618560075759888,
      "learning_rate": 1e-05,
      "loss": 0.5025,
      "mean_token_accuracy": 0.8445158004760742,
      "num_tokens": 212047699.0,
      "step": 1330
    },
    {
      "epoch": 0.6770091556459817,
      "grad_norm": 1.1449192762374878,
      "learning_rate": 1e-05,
      "loss": 0.5301,
      "mean_token_accuracy": 0.834658682346344,
      "num_tokens": 212189491.0,
      "step": 1331
    },
    {
      "epoch": 0.6775178026449644,
      "grad_norm": 1.1187528371810913,
      "learning_rate": 1e-05,
      "loss": 0.5411,
      "mean_token_accuracy": 0.8329986333847046,
      "num_tokens": 212350768.0,
      "step": 1332
    },
    {
      "epoch": 0.6780264496439471,
      "grad_norm": 1.0808460712432861,
      "learning_rate": 1e-05,
      "loss": 0.5457,
      "mean_token_accuracy": 0.8318155407905579,
      "num_tokens": 212505424.0,
      "step": 1333
    },
    {
      "epoch": 0.6785350966429298,
      "grad_norm": 1.0829100608825684,
      "learning_rate": 1e-05,
      "loss": 0.5315,
      "mean_token_accuracy": 0.8344807028770447,
      "num_tokens": 212656414.0,
      "step": 1334
    },
    {
      "epoch": 0.6790437436419126,
      "grad_norm": 1.0648113489151,
      "learning_rate": 1e-05,
      "loss": 0.5307,
      "mean_token_accuracy": 0.8352001905441284,
      "num_tokens": 212814380.0,
      "step": 1335
    },
    {
      "epoch": 0.6795523906408952,
      "grad_norm": 0.9756075143814087,
      "learning_rate": 1e-05,
      "loss": 0.5248,
      "mean_token_accuracy": 0.8369483947753906,
      "num_tokens": 212970630.0,
      "step": 1336
    },
    {
      "epoch": 0.6800610376398779,
      "grad_norm": 1.0037658214569092,
      "learning_rate": 1e-05,
      "loss": 0.5312,
      "mean_token_accuracy": 0.8363174200057983,
      "num_tokens": 213135584.0,
      "step": 1337
    },
    {
      "epoch": 0.6805696846388606,
      "grad_norm": 1.008080005645752,
      "learning_rate": 1e-05,
      "loss": 0.5304,
      "mean_token_accuracy": 0.8350541591644287,
      "num_tokens": 213296625.0,
      "step": 1338
    },
    {
      "epoch": 0.6810783316378434,
      "grad_norm": 1.0014381408691406,
      "learning_rate": 1e-05,
      "loss": 0.526,
      "mean_token_accuracy": 0.8380976915359497,
      "num_tokens": 213466408.0,
      "step": 1339
    },
    {
      "epoch": 0.681586978636826,
      "grad_norm": 1.0413174629211426,
      "learning_rate": 1e-05,
      "loss": 0.4952,
      "mean_token_accuracy": 0.846105694770813,
      "num_tokens": 213608058.0,
      "step": 1340
    },
    {
      "epoch": 0.6820956256358087,
      "grad_norm": 0.990386962890625,
      "learning_rate": 1e-05,
      "loss": 0.5326,
      "mean_token_accuracy": 0.835658848285675,
      "num_tokens": 213757840.0,
      "step": 1341
    },
    {
      "epoch": 0.6826042726347915,
      "grad_norm": 1.2512933015823364,
      "learning_rate": 1e-05,
      "loss": 0.5094,
      "mean_token_accuracy": 0.8409282565116882,
      "num_tokens": 213919287.0,
      "step": 1342
    },
    {
      "epoch": 0.6831129196337742,
      "grad_norm": 0.9421449303627014,
      "learning_rate": 1e-05,
      "loss": 0.5271,
      "mean_token_accuracy": 0.8369660973548889,
      "num_tokens": 214089237.0,
      "step": 1343
    },
    {
      "epoch": 0.6836215666327569,
      "grad_norm": 1.0769293308258057,
      "learning_rate": 1e-05,
      "loss": 0.5197,
      "mean_token_accuracy": 0.8391597270965576,
      "num_tokens": 214248836.0,
      "step": 1344
    },
    {
      "epoch": 0.6841302136317395,
      "grad_norm": 0.948756754398346,
      "learning_rate": 1e-05,
      "loss": 0.5196,
      "mean_token_accuracy": 0.8406849503517151,
      "num_tokens": 214419340.0,
      "step": 1345
    },
    {
      "epoch": 0.6846388606307223,
      "grad_norm": 0.9942221641540527,
      "learning_rate": 1e-05,
      "loss": 0.5296,
      "mean_token_accuracy": 0.8333039879798889,
      "num_tokens": 214578559.0,
      "step": 1346
    },
    {
      "epoch": 0.685147507629705,
      "grad_norm": 1.0239410400390625,
      "learning_rate": 1e-05,
      "loss": 0.5269,
      "mean_token_accuracy": 0.83641517162323,
      "num_tokens": 214735210.0,
      "step": 1347
    },
    {
      "epoch": 0.6856561546286877,
      "grad_norm": 1.0552555322647095,
      "learning_rate": 1e-05,
      "loss": 0.5374,
      "mean_token_accuracy": 0.8334919214248657,
      "num_tokens": 214902909.0,
      "step": 1348
    },
    {
      "epoch": 0.6861648016276704,
      "grad_norm": 1.0437448024749756,
      "learning_rate": 1e-05,
      "loss": 0.5295,
      "mean_token_accuracy": 0.8360548615455627,
      "num_tokens": 215064535.0,
      "step": 1349
    },
    {
      "epoch": 0.6866734486266531,
      "grad_norm": 0.9955970048904419,
      "learning_rate": 1e-05,
      "loss": 0.5288,
      "mean_token_accuracy": 0.835818886756897,
      "num_tokens": 215220787.0,
      "step": 1350
    },
    {
      "epoch": 0.6871820956256358,
      "grad_norm": 1.1820982694625854,
      "learning_rate": 1e-05,
      "loss": 0.5509,
      "mean_token_accuracy": 0.8299815654754639,
      "num_tokens": 215375205.0,
      "step": 1351
    },
    {
      "epoch": 0.6876907426246185,
      "grad_norm": 1.0263277292251587,
      "learning_rate": 1e-05,
      "loss": 0.5418,
      "mean_token_accuracy": 0.8360893726348877,
      "num_tokens": 215532596.0,
      "step": 1352
    },
    {
      "epoch": 0.6881993896236012,
      "grad_norm": 0.972446084022522,
      "learning_rate": 1e-05,
      "loss": 0.54,
      "mean_token_accuracy": 0.8345771431922913,
      "num_tokens": 215686435.0,
      "step": 1353
    },
    {
      "epoch": 0.688708036622584,
      "grad_norm": 1.0651381015777588,
      "learning_rate": 1e-05,
      "loss": 0.5407,
      "mean_token_accuracy": 0.8325681686401367,
      "num_tokens": 215854411.0,
      "step": 1354
    },
    {
      "epoch": 0.6892166836215666,
      "grad_norm": 1.0745397806167603,
      "learning_rate": 1e-05,
      "loss": 0.5056,
      "mean_token_accuracy": 0.8420690298080444,
      "num_tokens": 216016683.0,
      "step": 1355
    },
    {
      "epoch": 0.6897253306205493,
      "grad_norm": 1.0858376026153564,
      "learning_rate": 1e-05,
      "loss": 0.5441,
      "mean_token_accuracy": 0.8298423886299133,
      "num_tokens": 216176025.0,
      "step": 1356
    },
    {
      "epoch": 0.6902339776195321,
      "grad_norm": 1.060082197189331,
      "learning_rate": 1e-05,
      "loss": 0.5168,
      "mean_token_accuracy": 0.8398188352584839,
      "num_tokens": 216335904.0,
      "step": 1357
    },
    {
      "epoch": 0.6907426246185148,
      "grad_norm": 1.125820279121399,
      "learning_rate": 1e-05,
      "loss": 0.5228,
      "mean_token_accuracy": 0.8364633321762085,
      "num_tokens": 216490560.0,
      "step": 1358
    },
    {
      "epoch": 0.6912512716174974,
      "grad_norm": 1.0891369581222534,
      "learning_rate": 1e-05,
      "loss": 0.542,
      "mean_token_accuracy": 0.8322654962539673,
      "num_tokens": 216636941.0,
      "step": 1359
    },
    {
      "epoch": 0.6917599186164801,
      "grad_norm": 1.0387393236160278,
      "learning_rate": 1e-05,
      "loss": 0.5247,
      "mean_token_accuracy": 0.8380802869796753,
      "num_tokens": 216788892.0,
      "step": 1360
    },
    {
      "epoch": 0.6922685656154629,
      "grad_norm": 1.1280726194381714,
      "learning_rate": 1e-05,
      "loss": 0.5075,
      "mean_token_accuracy": 0.8425440192222595,
      "num_tokens": 216933108.0,
      "step": 1361
    },
    {
      "epoch": 0.6927772126144456,
      "grad_norm": 1.1861090660095215,
      "learning_rate": 1e-05,
      "loss": 0.5248,
      "mean_token_accuracy": 0.8362356424331665,
      "num_tokens": 217091789.0,
      "step": 1362
    },
    {
      "epoch": 0.6932858596134283,
      "grad_norm": 1.0378928184509277,
      "learning_rate": 1e-05,
      "loss": 0.5254,
      "mean_token_accuracy": 0.8354607820510864,
      "num_tokens": 217237394.0,
      "step": 1363
    },
    {
      "epoch": 0.6937945066124109,
      "grad_norm": 1.079342007637024,
      "learning_rate": 1e-05,
      "loss": 0.5319,
      "mean_token_accuracy": 0.8338744044303894,
      "num_tokens": 217391649.0,
      "step": 1364
    },
    {
      "epoch": 0.6943031536113937,
      "grad_norm": 1.0112749338150024,
      "learning_rate": 1e-05,
      "loss": 0.5069,
      "mean_token_accuracy": 0.8414908051490784,
      "num_tokens": 217554324.0,
      "step": 1365
    },
    {
      "epoch": 0.6948118006103764,
      "grad_norm": 1.1308326721191406,
      "learning_rate": 1e-05,
      "loss": 0.5229,
      "mean_token_accuracy": 0.837146520614624,
      "num_tokens": 217705676.0,
      "step": 1366
    },
    {
      "epoch": 0.6953204476093591,
      "grad_norm": 1.031843662261963,
      "learning_rate": 1e-05,
      "loss": 0.544,
      "mean_token_accuracy": 0.8329341411590576,
      "num_tokens": 217870837.0,
      "step": 1367
    },
    {
      "epoch": 0.6958290946083419,
      "grad_norm": 1.030326247215271,
      "learning_rate": 1e-05,
      "loss": 0.5513,
      "mean_token_accuracy": 0.830209493637085,
      "num_tokens": 218034192.0,
      "step": 1368
    },
    {
      "epoch": 0.6963377416073245,
      "grad_norm": 1.058383822441101,
      "learning_rate": 1e-05,
      "loss": 0.5228,
      "mean_token_accuracy": 0.8376389741897583,
      "num_tokens": 218201741.0,
      "step": 1369
    },
    {
      "epoch": 0.6968463886063072,
      "grad_norm": 1.0227019786834717,
      "learning_rate": 1e-05,
      "loss": 0.5384,
      "mean_token_accuracy": 0.8350127935409546,
      "num_tokens": 218361880.0,
      "step": 1370
    },
    {
      "epoch": 0.6973550356052899,
      "grad_norm": 1.0567728281021118,
      "learning_rate": 1e-05,
      "loss": 0.5002,
      "mean_token_accuracy": 0.8430039882659912,
      "num_tokens": 218507302.0,
      "step": 1371
    },
    {
      "epoch": 0.6978636826042727,
      "grad_norm": 0.9965653419494629,
      "learning_rate": 1e-05,
      "loss": 0.527,
      "mean_token_accuracy": 0.8355226516723633,
      "num_tokens": 218677170.0,
      "step": 1372
    },
    {
      "epoch": 0.6983723296032553,
      "grad_norm": 1.1475521326065063,
      "learning_rate": 1e-05,
      "loss": 0.5284,
      "mean_token_accuracy": 0.8367115259170532,
      "num_tokens": 218830583.0,
      "step": 1373
    },
    {
      "epoch": 0.698880976602238,
      "grad_norm": 1.0074726343154907,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8365511894226074,
      "num_tokens": 218990542.0,
      "step": 1374
    },
    {
      "epoch": 0.6993896236012207,
      "grad_norm": 1.1430093050003052,
      "learning_rate": 1e-05,
      "loss": 0.5373,
      "mean_token_accuracy": 0.834356427192688,
      "num_tokens": 219150118.0,
      "step": 1375
    },
    {
      "epoch": 0.6998982706002035,
      "grad_norm": 1.0113097429275513,
      "learning_rate": 1e-05,
      "loss": 0.512,
      "mean_token_accuracy": 0.8409308195114136,
      "num_tokens": 219320664.0,
      "step": 1376
    },
    {
      "epoch": 0.7004069175991862,
      "grad_norm": 1.0833232402801514,
      "learning_rate": 1e-05,
      "loss": 0.5412,
      "mean_token_accuracy": 0.8324387073516846,
      "num_tokens": 219483168.0,
      "step": 1377
    },
    {
      "epoch": 0.7009155645981688,
      "grad_norm": 1.008895993232727,
      "learning_rate": 1e-05,
      "loss": 0.5301,
      "mean_token_accuracy": 0.8351591229438782,
      "num_tokens": 219643460.0,
      "step": 1378
    },
    {
      "epoch": 0.7014242115971516,
      "grad_norm": 1.0583291053771973,
      "learning_rate": 1e-05,
      "loss": 0.5234,
      "mean_token_accuracy": 0.8379659056663513,
      "num_tokens": 219801607.0,
      "step": 1379
    },
    {
      "epoch": 0.7019328585961343,
      "grad_norm": 1.1074832677841187,
      "learning_rate": 1e-05,
      "loss": 0.5377,
      "mean_token_accuracy": 0.834449291229248,
      "num_tokens": 219953874.0,
      "step": 1380
    },
    {
      "epoch": 0.702441505595117,
      "grad_norm": 1.0328203439712524,
      "learning_rate": 1e-05,
      "loss": 0.4908,
      "mean_token_accuracy": 0.8464456796646118,
      "num_tokens": 220101993.0,
      "step": 1381
    },
    {
      "epoch": 0.7029501525940997,
      "grad_norm": 0.9966577887535095,
      "learning_rate": 1e-05,
      "loss": 0.5292,
      "mean_token_accuracy": 0.8362774848937988,
      "num_tokens": 220272230.0,
      "step": 1382
    },
    {
      "epoch": 0.7034587995930824,
      "grad_norm": 0.9913000464439392,
      "learning_rate": 1e-05,
      "loss": 0.5257,
      "mean_token_accuracy": 0.8375822305679321,
      "num_tokens": 220432579.0,
      "step": 1383
    },
    {
      "epoch": 0.7039674465920651,
      "grad_norm": 1.0375653505325317,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.8354050517082214,
      "num_tokens": 220588939.0,
      "step": 1384
    },
    {
      "epoch": 0.7044760935910478,
      "grad_norm": 0.9509556293487549,
      "learning_rate": 1e-05,
      "loss": 0.5041,
      "mean_token_accuracy": 0.8444002270698547,
      "num_tokens": 220759155.0,
      "step": 1385
    },
    {
      "epoch": 0.7049847405900305,
      "grad_norm": 0.9617030620574951,
      "learning_rate": 1e-05,
      "loss": 0.5248,
      "mean_token_accuracy": 0.8367959260940552,
      "num_tokens": 220916843.0,
      "step": 1386
    },
    {
      "epoch": 0.7054933875890133,
      "grad_norm": 1.019019603729248,
      "learning_rate": 1e-05,
      "loss": 0.5339,
      "mean_token_accuracy": 0.8349019885063171,
      "num_tokens": 221083858.0,
      "step": 1387
    },
    {
      "epoch": 0.7060020345879959,
      "grad_norm": 1.065538763999939,
      "learning_rate": 1e-05,
      "loss": 0.5757,
      "mean_token_accuracy": 0.8243120908737183,
      "num_tokens": 221258649.0,
      "step": 1388
    },
    {
      "epoch": 0.7065106815869786,
      "grad_norm": 1.0140337944030762,
      "learning_rate": 1e-05,
      "loss": 0.5384,
      "mean_token_accuracy": 0.832329273223877,
      "num_tokens": 221425144.0,
      "step": 1389
    },
    {
      "epoch": 0.7070193285859614,
      "grad_norm": 1.085599660873413,
      "learning_rate": 1e-05,
      "loss": 0.5375,
      "mean_token_accuracy": 0.834128201007843,
      "num_tokens": 221583215.0,
      "step": 1390
    },
    {
      "epoch": 0.7075279755849441,
      "grad_norm": 0.9903988242149353,
      "learning_rate": 1e-05,
      "loss": 0.542,
      "mean_token_accuracy": 0.8328906893730164,
      "num_tokens": 221755255.0,
      "step": 1391
    },
    {
      "epoch": 0.7080366225839267,
      "grad_norm": 1.087711215019226,
      "learning_rate": 1e-05,
      "loss": 0.5549,
      "mean_token_accuracy": 0.8285048007965088,
      "num_tokens": 221923372.0,
      "step": 1392
    },
    {
      "epoch": 0.7085452695829094,
      "grad_norm": 0.9438535571098328,
      "learning_rate": 1e-05,
      "loss": 0.5308,
      "mean_token_accuracy": 0.8352342247962952,
      "num_tokens": 222080701.0,
      "step": 1393
    },
    {
      "epoch": 0.7090539165818922,
      "grad_norm": 0.9913089871406555,
      "learning_rate": 1e-05,
      "loss": 0.4932,
      "mean_token_accuracy": 0.8428090214729309,
      "num_tokens": 222233287.0,
      "step": 1394
    },
    {
      "epoch": 0.7095625635808749,
      "grad_norm": 1.000691294670105,
      "learning_rate": 1e-05,
      "loss": 0.5478,
      "mean_token_accuracy": 0.8302925825119019,
      "num_tokens": 222399597.0,
      "step": 1395
    },
    {
      "epoch": 0.7100712105798576,
      "grad_norm": 1.1176949739456177,
      "learning_rate": 1e-05,
      "loss": 0.5346,
      "mean_token_accuracy": 0.8353996276855469,
      "num_tokens": 222549595.0,
      "step": 1396
    },
    {
      "epoch": 0.7105798575788402,
      "grad_norm": 0.968652069568634,
      "learning_rate": 1e-05,
      "loss": 0.5561,
      "mean_token_accuracy": 0.8303560018539429,
      "num_tokens": 222711422.0,
      "step": 1397
    },
    {
      "epoch": 0.711088504577823,
      "grad_norm": 1.0685300827026367,
      "learning_rate": 1e-05,
      "loss": 0.5077,
      "mean_token_accuracy": 0.8400074243545532,
      "num_tokens": 222873104.0,
      "step": 1398
    },
    {
      "epoch": 0.7115971515768057,
      "grad_norm": 1.0692155361175537,
      "learning_rate": 1e-05,
      "loss": 0.5379,
      "mean_token_accuracy": 0.8347507119178772,
      "num_tokens": 223020625.0,
      "step": 1399
    },
    {
      "epoch": 0.7121057985757884,
      "grad_norm": 1.057916283607483,
      "learning_rate": 1e-05,
      "loss": 0.5137,
      "mean_token_accuracy": 0.8409354090690613,
      "num_tokens": 223175189.0,
      "step": 1400
    },
    {
      "epoch": 0.7126144455747712,
      "grad_norm": 1.0396839380264282,
      "learning_rate": 1e-05,
      "loss": 0.5184,
      "mean_token_accuracy": 0.836387038230896,
      "num_tokens": 223335073.0,
      "step": 1401
    },
    {
      "epoch": 0.7131230925737538,
      "grad_norm": 0.9966652393341064,
      "learning_rate": 1e-05,
      "loss": 0.5288,
      "mean_token_accuracy": 0.8362771272659302,
      "num_tokens": 223499553.0,
      "step": 1402
    },
    {
      "epoch": 0.7136317395727365,
      "grad_norm": 1.064444899559021,
      "learning_rate": 1e-05,
      "loss": 0.5123,
      "mean_token_accuracy": 0.8392417430877686,
      "num_tokens": 223657911.0,
      "step": 1403
    },
    {
      "epoch": 0.7141403865717192,
      "grad_norm": 1.1067264080047607,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.8337592482566833,
      "num_tokens": 223808765.0,
      "step": 1404
    },
    {
      "epoch": 0.714649033570702,
      "grad_norm": 1.0045108795166016,
      "learning_rate": 1e-05,
      "loss": 0.5086,
      "mean_token_accuracy": 0.8400107622146606,
      "num_tokens": 223974533.0,
      "step": 1405
    },
    {
      "epoch": 0.7151576805696847,
      "grad_norm": 1.034972906112671,
      "learning_rate": 1e-05,
      "loss": 0.4909,
      "mean_token_accuracy": 0.8464032411575317,
      "num_tokens": 224133481.0,
      "step": 1406
    },
    {
      "epoch": 0.7156663275686673,
      "grad_norm": 1.1038155555725098,
      "learning_rate": 1e-05,
      "loss": 0.5127,
      "mean_token_accuracy": 0.8401011824607849,
      "num_tokens": 224282640.0,
      "step": 1407
    },
    {
      "epoch": 0.71617497456765,
      "grad_norm": 1.0537794828414917,
      "learning_rate": 1e-05,
      "loss": 0.5516,
      "mean_token_accuracy": 0.8289327621459961,
      "num_tokens": 224447298.0,
      "step": 1408
    },
    {
      "epoch": 0.7166836215666328,
      "grad_norm": 1.0741024017333984,
      "learning_rate": 1e-05,
      "loss": 0.5102,
      "mean_token_accuracy": 0.8410897254943848,
      "num_tokens": 224596045.0,
      "step": 1409
    },
    {
      "epoch": 0.7171922685656155,
      "grad_norm": 1.037316083908081,
      "learning_rate": 1e-05,
      "loss": 0.5357,
      "mean_token_accuracy": 0.8368381261825562,
      "num_tokens": 224757199.0,
      "step": 1410
    },
    {
      "epoch": 0.7177009155645981,
      "grad_norm": 1.006191372871399,
      "learning_rate": 1e-05,
      "loss": 0.5133,
      "mean_token_accuracy": 0.8390223383903503,
      "num_tokens": 224913566.0,
      "step": 1411
    },
    {
      "epoch": 0.7182095625635809,
      "grad_norm": 1.0501819849014282,
      "learning_rate": 1e-05,
      "loss": 0.5285,
      "mean_token_accuracy": 0.8382315039634705,
      "num_tokens": 225078638.0,
      "step": 1412
    },
    {
      "epoch": 0.7187182095625636,
      "grad_norm": 0.9292073845863342,
      "learning_rate": 1e-05,
      "loss": 0.5319,
      "mean_token_accuracy": 0.8357728123664856,
      "num_tokens": 225242585.0,
      "step": 1413
    },
    {
      "epoch": 0.7192268565615463,
      "grad_norm": 1.1309953927993774,
      "learning_rate": 1e-05,
      "loss": 0.503,
      "mean_token_accuracy": 0.8418674468994141,
      "num_tokens": 225402143.0,
      "step": 1414
    },
    {
      "epoch": 0.719735503560529,
      "grad_norm": 1.0340876579284668,
      "learning_rate": 1e-05,
      "loss": 0.5559,
      "mean_token_accuracy": 0.83051598072052,
      "num_tokens": 225561481.0,
      "step": 1415
    },
    {
      "epoch": 0.7202441505595117,
      "grad_norm": 1.0265415906906128,
      "learning_rate": 1e-05,
      "loss": 0.5083,
      "mean_token_accuracy": 0.8427028059959412,
      "num_tokens": 225720513.0,
      "step": 1416
    },
    {
      "epoch": 0.7207527975584944,
      "grad_norm": 1.1584059000015259,
      "learning_rate": 1e-05,
      "loss": 0.5592,
      "mean_token_accuracy": 0.8283109664916992,
      "num_tokens": 225873349.0,
      "step": 1417
    },
    {
      "epoch": 0.7212614445574771,
      "grad_norm": 1.0035206079483032,
      "learning_rate": 1e-05,
      "loss": 0.5073,
      "mean_token_accuracy": 0.8412970900535583,
      "num_tokens": 226039357.0,
      "step": 1418
    },
    {
      "epoch": 0.7217700915564598,
      "grad_norm": 1.0063248872756958,
      "learning_rate": 1e-05,
      "loss": 0.5033,
      "mean_token_accuracy": 0.8425672054290771,
      "num_tokens": 226189348.0,
      "step": 1419
    },
    {
      "epoch": 0.7222787385554426,
      "grad_norm": 1.0380918979644775,
      "learning_rate": 1e-05,
      "loss": 0.5352,
      "mean_token_accuracy": 0.8346319198608398,
      "num_tokens": 226335920.0,
      "step": 1420
    },
    {
      "epoch": 0.7227873855544252,
      "grad_norm": 1.1460973024368286,
      "learning_rate": 1e-05,
      "loss": 0.5388,
      "mean_token_accuracy": 0.8333281874656677,
      "num_tokens": 226495005.0,
      "step": 1421
    },
    {
      "epoch": 0.7232960325534079,
      "grad_norm": 1.0217316150665283,
      "learning_rate": 1e-05,
      "loss": 0.4966,
      "mean_token_accuracy": 0.8454707860946655,
      "num_tokens": 226657995.0,
      "step": 1422
    },
    {
      "epoch": 0.7238046795523907,
      "grad_norm": 1.2179800271987915,
      "learning_rate": 1e-05,
      "loss": 0.5315,
      "mean_token_accuracy": 0.836176335811615,
      "num_tokens": 226828202.0,
      "step": 1423
    },
    {
      "epoch": 0.7243133265513734,
      "grad_norm": 1.0598959922790527,
      "learning_rate": 1e-05,
      "loss": 0.542,
      "mean_token_accuracy": 0.8340860605239868,
      "num_tokens": 226982248.0,
      "step": 1424
    },
    {
      "epoch": 0.724821973550356,
      "grad_norm": 1.1656885147094727,
      "learning_rate": 1e-05,
      "loss": 0.5318,
      "mean_token_accuracy": 0.8348649144172668,
      "num_tokens": 227129474.0,
      "step": 1425
    },
    {
      "epoch": 0.7253306205493387,
      "grad_norm": 1.257928490638733,
      "learning_rate": 1e-05,
      "loss": 0.5091,
      "mean_token_accuracy": 0.8404781818389893,
      "num_tokens": 227274613.0,
      "step": 1426
    },
    {
      "epoch": 0.7258392675483215,
      "grad_norm": 1.0831841230392456,
      "learning_rate": 1e-05,
      "loss": 0.5461,
      "mean_token_accuracy": 0.8328593373298645,
      "num_tokens": 227440413.0,
      "step": 1427
    },
    {
      "epoch": 0.7263479145473042,
      "grad_norm": 1.2265989780426025,
      "learning_rate": 1e-05,
      "loss": 0.5218,
      "mean_token_accuracy": 0.8368296027183533,
      "num_tokens": 227605134.0,
      "step": 1428
    },
    {
      "epoch": 0.7268565615462869,
      "grad_norm": 0.9749830365180969,
      "learning_rate": 1e-05,
      "loss": 0.5134,
      "mean_token_accuracy": 0.8399536609649658,
      "num_tokens": 227761956.0,
      "step": 1429
    },
    {
      "epoch": 0.7273652085452695,
      "grad_norm": 1.2150827646255493,
      "learning_rate": 1e-05,
      "loss": 0.5796,
      "mean_token_accuracy": 0.8240715265274048,
      "num_tokens": 227922627.0,
      "step": 1430
    },
    {
      "epoch": 0.7278738555442523,
      "grad_norm": 1.0878562927246094,
      "learning_rate": 1e-05,
      "loss": 0.5264,
      "mean_token_accuracy": 0.835365891456604,
      "num_tokens": 228080800.0,
      "step": 1431
    },
    {
      "epoch": 0.728382502543235,
      "grad_norm": 1.1435154676437378,
      "learning_rate": 1e-05,
      "loss": 0.5195,
      "mean_token_accuracy": 0.8392224311828613,
      "num_tokens": 228242141.0,
      "step": 1432
    },
    {
      "epoch": 0.7288911495422177,
      "grad_norm": 1.3589378595352173,
      "learning_rate": 1e-05,
      "loss": 0.5227,
      "mean_token_accuracy": 0.8367589712142944,
      "num_tokens": 228387653.0,
      "step": 1433
    },
    {
      "epoch": 0.7293997965412004,
      "grad_norm": 1.0286848545074463,
      "learning_rate": 1e-05,
      "loss": 0.5023,
      "mean_token_accuracy": 0.8427996635437012,
      "num_tokens": 228545473.0,
      "step": 1434
    },
    {
      "epoch": 0.7299084435401831,
      "grad_norm": 1.1867001056671143,
      "learning_rate": 1e-05,
      "loss": 0.5364,
      "mean_token_accuracy": 0.8338367938995361,
      "num_tokens": 228713308.0,
      "step": 1435
    },
    {
      "epoch": 0.7304170905391658,
      "grad_norm": 1.0601273775100708,
      "learning_rate": 1e-05,
      "loss": 0.5541,
      "mean_token_accuracy": 0.8306837677955627,
      "num_tokens": 228884216.0,
      "step": 1436
    },
    {
      "epoch": 0.7309257375381485,
      "grad_norm": 1.101950764656067,
      "learning_rate": 1e-05,
      "loss": 0.5667,
      "mean_token_accuracy": 0.8280445337295532,
      "num_tokens": 229029827.0,
      "step": 1437
    },
    {
      "epoch": 0.7314343845371313,
      "grad_norm": 1.1425142288208008,
      "learning_rate": 1e-05,
      "loss": 0.5437,
      "mean_token_accuracy": 0.8314026594161987,
      "num_tokens": 229181920.0,
      "step": 1438
    },
    {
      "epoch": 0.731943031536114,
      "grad_norm": 0.9540426731109619,
      "learning_rate": 1e-05,
      "loss": 0.5452,
      "mean_token_accuracy": 0.8314418792724609,
      "num_tokens": 229349488.0,
      "step": 1439
    },
    {
      "epoch": 0.7324516785350966,
      "grad_norm": 1.516485333442688,
      "learning_rate": 1e-05,
      "loss": 0.5179,
      "mean_token_accuracy": 0.8390694260597229,
      "num_tokens": 229505884.0,
      "step": 1440
    },
    {
      "epoch": 0.7329603255340793,
      "grad_norm": 1.3585577011108398,
      "learning_rate": 1e-05,
      "loss": 0.5283,
      "mean_token_accuracy": 0.8348572254180908,
      "num_tokens": 229673440.0,
      "step": 1441
    },
    {
      "epoch": 0.7334689725330621,
      "grad_norm": 1.0918457508087158,
      "learning_rate": 1e-05,
      "loss": 0.516,
      "mean_token_accuracy": 0.8398845195770264,
      "num_tokens": 229844345.0,
      "step": 1442
    },
    {
      "epoch": 0.7339776195320448,
      "grad_norm": 1.2803473472595215,
      "learning_rate": 1e-05,
      "loss": 0.5754,
      "mean_token_accuracy": 0.8247080445289612,
      "num_tokens": 230001344.0,
      "step": 1443
    },
    {
      "epoch": 0.7344862665310274,
      "grad_norm": 1.09883451461792,
      "learning_rate": 1e-05,
      "loss": 0.5085,
      "mean_token_accuracy": 0.8424191474914551,
      "num_tokens": 230160760.0,
      "step": 1444
    },
    {
      "epoch": 0.7349949135300101,
      "grad_norm": 1.2438838481903076,
      "learning_rate": 1e-05,
      "loss": 0.5198,
      "mean_token_accuracy": 0.8385148644447327,
      "num_tokens": 230318252.0,
      "step": 1445
    },
    {
      "epoch": 0.7355035605289929,
      "grad_norm": 1.1230305433273315,
      "learning_rate": 1e-05,
      "loss": 0.534,
      "mean_token_accuracy": 0.8361711502075195,
      "num_tokens": 230480542.0,
      "step": 1446
    },
    {
      "epoch": 0.7360122075279756,
      "grad_norm": 1.1084116697311401,
      "learning_rate": 1e-05,
      "loss": 0.5423,
      "mean_token_accuracy": 0.8342896103858948,
      "num_tokens": 230631226.0,
      "step": 1447
    },
    {
      "epoch": 0.7365208545269583,
      "grad_norm": 1.1472543478012085,
      "learning_rate": 1e-05,
      "loss": 0.5009,
      "mean_token_accuracy": 0.8427757024765015,
      "num_tokens": 230793950.0,
      "step": 1448
    },
    {
      "epoch": 0.737029501525941,
      "grad_norm": 1.0803556442260742,
      "learning_rate": 1e-05,
      "loss": 0.5749,
      "mean_token_accuracy": 0.8247697949409485,
      "num_tokens": 230949258.0,
      "step": 1449
    },
    {
      "epoch": 0.7375381485249237,
      "grad_norm": 1.0975403785705566,
      "learning_rate": 1e-05,
      "loss": 0.5154,
      "mean_token_accuracy": 0.8405905961990356,
      "num_tokens": 231116995.0,
      "step": 1450
    },
    {
      "epoch": 0.7380467955239064,
      "grad_norm": 0.9167547821998596,
      "learning_rate": 1e-05,
      "loss": 0.4993,
      "mean_token_accuracy": 0.8458133339881897,
      "num_tokens": 231286840.0,
      "step": 1451
    },
    {
      "epoch": 0.7385554425228891,
      "grad_norm": 1.0844895839691162,
      "learning_rate": 1e-05,
      "loss": 0.5208,
      "mean_token_accuracy": 0.8365471959114075,
      "num_tokens": 231446504.0,
      "step": 1452
    },
    {
      "epoch": 0.7390640895218719,
      "grad_norm": 0.9481300711631775,
      "learning_rate": 1e-05,
      "loss": 0.5508,
      "mean_token_accuracy": 0.8306936025619507,
      "num_tokens": 231614914.0,
      "step": 1453
    },
    {
      "epoch": 0.7395727365208545,
      "grad_norm": 1.0026079416275024,
      "learning_rate": 1e-05,
      "loss": 0.5209,
      "mean_token_accuracy": 0.8390685319900513,
      "num_tokens": 231772874.0,
      "step": 1454
    },
    {
      "epoch": 0.7400813835198372,
      "grad_norm": 0.9564797282218933,
      "learning_rate": 1e-05,
      "loss": 0.5323,
      "mean_token_accuracy": 0.834564208984375,
      "num_tokens": 231937051.0,
      "step": 1455
    },
    {
      "epoch": 0.7405900305188199,
      "grad_norm": 1.0185620784759521,
      "learning_rate": 1e-05,
      "loss": 0.517,
      "mean_token_accuracy": 0.8383815884590149,
      "num_tokens": 232087611.0,
      "step": 1456
    },
    {
      "epoch": 0.7410986775178027,
      "grad_norm": 1.0209577083587646,
      "learning_rate": 1e-05,
      "loss": 0.5688,
      "mean_token_accuracy": 0.8257862329483032,
      "num_tokens": 232246891.0,
      "step": 1457
    },
    {
      "epoch": 0.7416073245167853,
      "grad_norm": 0.9779701232910156,
      "learning_rate": 1e-05,
      "loss": 0.5092,
      "mean_token_accuracy": 0.8411809206008911,
      "num_tokens": 232405181.0,
      "step": 1458
    },
    {
      "epoch": 0.742115971515768,
      "grad_norm": 0.9819640517234802,
      "learning_rate": 1e-05,
      "loss": 0.5713,
      "mean_token_accuracy": 0.8215596675872803,
      "num_tokens": 232569215.0,
      "step": 1459
    },
    {
      "epoch": 0.7426246185147508,
      "grad_norm": 0.9781466126441956,
      "learning_rate": 1e-05,
      "loss": 0.5015,
      "mean_token_accuracy": 0.8428027629852295,
      "num_tokens": 232730814.0,
      "step": 1460
    },
    {
      "epoch": 0.7431332655137335,
      "grad_norm": 1.0607805252075195,
      "learning_rate": 1e-05,
      "loss": 0.5266,
      "mean_token_accuracy": 0.8380945324897766,
      "num_tokens": 232892427.0,
      "step": 1461
    },
    {
      "epoch": 0.7436419125127162,
      "grad_norm": 1.0117738246917725,
      "learning_rate": 1e-05,
      "loss": 0.5172,
      "mean_token_accuracy": 0.8377254009246826,
      "num_tokens": 233062413.0,
      "step": 1462
    },
    {
      "epoch": 0.7441505595116988,
      "grad_norm": 1.094952940940857,
      "learning_rate": 1e-05,
      "loss": 0.5061,
      "mean_token_accuracy": 0.8411229848861694,
      "num_tokens": 233223449.0,
      "step": 1463
    },
    {
      "epoch": 0.7446592065106816,
      "grad_norm": 1.0199710130691528,
      "learning_rate": 1e-05,
      "loss": 0.5188,
      "mean_token_accuracy": 0.8373159170150757,
      "num_tokens": 233382314.0,
      "step": 1464
    },
    {
      "epoch": 0.7451678535096643,
      "grad_norm": 0.9318074584007263,
      "learning_rate": 1e-05,
      "loss": 0.5111,
      "mean_token_accuracy": 0.8436965346336365,
      "num_tokens": 233549471.0,
      "step": 1465
    },
    {
      "epoch": 0.745676500508647,
      "grad_norm": 0.9865319728851318,
      "learning_rate": 1e-05,
      "loss": 0.5337,
      "mean_token_accuracy": 0.8344173431396484,
      "num_tokens": 233714244.0,
      "step": 1466
    },
    {
      "epoch": 0.7461851475076297,
      "grad_norm": 1.060027837753296,
      "learning_rate": 1e-05,
      "loss": 0.5224,
      "mean_token_accuracy": 0.8394288420677185,
      "num_tokens": 233873031.0,
      "step": 1467
    },
    {
      "epoch": 0.7466937945066124,
      "grad_norm": 0.9694737195968628,
      "learning_rate": 1e-05,
      "loss": 0.5047,
      "mean_token_accuracy": 0.8412796854972839,
      "num_tokens": 234028924.0,
      "step": 1468
    },
    {
      "epoch": 0.7472024415055951,
      "grad_norm": 1.07378089427948,
      "learning_rate": 1e-05,
      "loss": 0.5137,
      "mean_token_accuracy": 0.8402303457260132,
      "num_tokens": 234187316.0,
      "step": 1469
    },
    {
      "epoch": 0.7477110885045778,
      "grad_norm": 1.0076998472213745,
      "learning_rate": 1e-05,
      "loss": 0.5046,
      "mean_token_accuracy": 0.8435081243515015,
      "num_tokens": 234352052.0,
      "step": 1470
    },
    {
      "epoch": 0.7482197355035606,
      "grad_norm": 1.1248302459716797,
      "learning_rate": 1e-05,
      "loss": 0.5054,
      "mean_token_accuracy": 0.8418548703193665,
      "num_tokens": 234522725.0,
      "step": 1471
    },
    {
      "epoch": 0.7487283825025433,
      "grad_norm": 1.0523717403411865,
      "learning_rate": 1e-05,
      "loss": 0.5446,
      "mean_token_accuracy": 0.8295474052429199,
      "num_tokens": 234685164.0,
      "step": 1472
    },
    {
      "epoch": 0.7492370295015259,
      "grad_norm": 1.009028673171997,
      "learning_rate": 1e-05,
      "loss": 0.5101,
      "mean_token_accuracy": 0.8410037755966187,
      "num_tokens": 234844848.0,
      "step": 1473
    },
    {
      "epoch": 0.7497456765005086,
      "grad_norm": 1.1979455947875977,
      "learning_rate": 1e-05,
      "loss": 0.497,
      "mean_token_accuracy": 0.8434077501296997,
      "num_tokens": 234999779.0,
      "step": 1474
    },
    {
      "epoch": 0.7502543234994914,
      "grad_norm": 1.081596851348877,
      "learning_rate": 1e-05,
      "loss": 0.5912,
      "mean_token_accuracy": 0.8188192844390869,
      "num_tokens": 235162263.0,
      "step": 1475
    },
    {
      "epoch": 0.7507629704984741,
      "grad_norm": 1.0012668371200562,
      "learning_rate": 1e-05,
      "loss": 0.5698,
      "mean_token_accuracy": 0.8237411975860596,
      "num_tokens": 235319603.0,
      "step": 1476
    },
    {
      "epoch": 0.7512716174974567,
      "grad_norm": 1.0081019401550293,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.8393720984458923,
      "num_tokens": 235479108.0,
      "step": 1477
    },
    {
      "epoch": 0.7517802644964394,
      "grad_norm": 1.0606874227523804,
      "learning_rate": 1e-05,
      "loss": 0.5275,
      "mean_token_accuracy": 0.8372191786766052,
      "num_tokens": 235635662.0,
      "step": 1478
    },
    {
      "epoch": 0.7522889114954222,
      "grad_norm": 1.1336381435394287,
      "learning_rate": 1e-05,
      "loss": 0.5402,
      "mean_token_accuracy": 0.8327375054359436,
      "num_tokens": 235778848.0,
      "step": 1479
    },
    {
      "epoch": 0.7527975584944049,
      "grad_norm": 0.9920240044593811,
      "learning_rate": 1e-05,
      "loss": 0.5048,
      "mean_token_accuracy": 0.8434538841247559,
      "num_tokens": 235935563.0,
      "step": 1480
    },
    {
      "epoch": 0.7533062054933876,
      "grad_norm": 1.0288413763046265,
      "learning_rate": 1e-05,
      "loss": 0.5292,
      "mean_token_accuracy": 0.8377621173858643,
      "num_tokens": 236102712.0,
      "step": 1481
    },
    {
      "epoch": 0.7538148524923703,
      "grad_norm": 1.0564460754394531,
      "learning_rate": 1e-05,
      "loss": 0.5207,
      "mean_token_accuracy": 0.8376520872116089,
      "num_tokens": 236247865.0,
      "step": 1482
    },
    {
      "epoch": 0.754323499491353,
      "grad_norm": 1.0536723136901855,
      "learning_rate": 1e-05,
      "loss": 0.527,
      "mean_token_accuracy": 0.8366076350212097,
      "num_tokens": 236406459.0,
      "step": 1483
    },
    {
      "epoch": 0.7548321464903357,
      "grad_norm": 0.9918603301048279,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.8441921472549438,
      "num_tokens": 236548474.0,
      "step": 1484
    },
    {
      "epoch": 0.7553407934893184,
      "grad_norm": 1.125241994857788,
      "learning_rate": 1e-05,
      "loss": 0.5428,
      "mean_token_accuracy": 0.8297998905181885,
      "num_tokens": 236699987.0,
      "step": 1485
    },
    {
      "epoch": 0.7558494404883012,
      "grad_norm": 1.0547477006912231,
      "learning_rate": 1e-05,
      "loss": 0.4975,
      "mean_token_accuracy": 0.844157338142395,
      "num_tokens": 236859308.0,
      "step": 1486
    },
    {
      "epoch": 0.7563580874872838,
      "grad_norm": 1.0355929136276245,
      "learning_rate": 1e-05,
      "loss": 0.4996,
      "mean_token_accuracy": 0.843138575553894,
      "num_tokens": 237011851.0,
      "step": 1487
    },
    {
      "epoch": 0.7568667344862665,
      "grad_norm": 1.0818662643432617,
      "learning_rate": 1e-05,
      "loss": 0.5279,
      "mean_token_accuracy": 0.8360995650291443,
      "num_tokens": 237170473.0,
      "step": 1488
    },
    {
      "epoch": 0.7573753814852492,
      "grad_norm": 1.2234809398651123,
      "learning_rate": 1e-05,
      "loss": 0.5049,
      "mean_token_accuracy": 0.8414808511734009,
      "num_tokens": 237317583.0,
      "step": 1489
    },
    {
      "epoch": 0.757884028484232,
      "grad_norm": 1.0319559574127197,
      "learning_rate": 1e-05,
      "loss": 0.5018,
      "mean_token_accuracy": 0.8432955145835876,
      "num_tokens": 237483382.0,
      "step": 1490
    },
    {
      "epoch": 0.7583926754832147,
      "grad_norm": 1.0208332538604736,
      "learning_rate": 1e-05,
      "loss": 0.5151,
      "mean_token_accuracy": 0.8389483094215393,
      "num_tokens": 237642857.0,
      "step": 1491
    },
    {
      "epoch": 0.7589013224821973,
      "grad_norm": 1.0073045492172241,
      "learning_rate": 1e-05,
      "loss": 0.5174,
      "mean_token_accuracy": 0.8391669988632202,
      "num_tokens": 237809071.0,
      "step": 1492
    },
    {
      "epoch": 0.7594099694811801,
      "grad_norm": 1.1355136632919312,
      "learning_rate": 1e-05,
      "loss": 0.5396,
      "mean_token_accuracy": 0.8332517147064209,
      "num_tokens": 237977918.0,
      "step": 1493
    },
    {
      "epoch": 0.7599186164801628,
      "grad_norm": 1.0361236333847046,
      "learning_rate": 1e-05,
      "loss": 0.537,
      "mean_token_accuracy": 0.8351563215255737,
      "num_tokens": 238146507.0,
      "step": 1494
    },
    {
      "epoch": 0.7604272634791455,
      "grad_norm": 0.9729189872741699,
      "learning_rate": 1e-05,
      "loss": 0.5075,
      "mean_token_accuracy": 0.8422070741653442,
      "num_tokens": 238317537.0,
      "step": 1495
    },
    {
      "epoch": 0.7609359104781281,
      "grad_norm": 1.0099557638168335,
      "learning_rate": 1e-05,
      "loss": 0.5084,
      "mean_token_accuracy": 0.8419430255889893,
      "num_tokens": 238483591.0,
      "step": 1496
    },
    {
      "epoch": 0.7614445574771109,
      "grad_norm": 0.9618970155715942,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8457642793655396,
      "num_tokens": 238641826.0,
      "step": 1497
    },
    {
      "epoch": 0.7619532044760936,
      "grad_norm": 1.0791717767715454,
      "learning_rate": 1e-05,
      "loss": 0.5176,
      "mean_token_accuracy": 0.8391422033309937,
      "num_tokens": 238793523.0,
      "step": 1498
    },
    {
      "epoch": 0.7624618514750763,
      "grad_norm": 1.0534913539886475,
      "learning_rate": 1e-05,
      "loss": 0.5243,
      "mean_token_accuracy": 0.8369989395141602,
      "num_tokens": 238956488.0,
      "step": 1499
    },
    {
      "epoch": 0.762970498474059,
      "grad_norm": 0.9715752601623535,
      "learning_rate": 1e-05,
      "loss": 0.4962,
      "mean_token_accuracy": 0.8451597094535828,
      "num_tokens": 239112968.0,
      "step": 1500
    },
    {
      "epoch": 0.7634791454730417,
      "grad_norm": 1.042768955230713,
      "learning_rate": 1e-05,
      "loss": 0.5518,
      "mean_token_accuracy": 0.83104407787323,
      "num_tokens": 239278918.0,
      "step": 1501
    },
    {
      "epoch": 0.7639877924720244,
      "grad_norm": 1.1038720607757568,
      "learning_rate": 1e-05,
      "loss": 0.5063,
      "mean_token_accuracy": 0.8409518003463745,
      "num_tokens": 239429305.0,
      "step": 1502
    },
    {
      "epoch": 0.7644964394710071,
      "grad_norm": 1.003100037574768,
      "learning_rate": 1e-05,
      "loss": 0.4916,
      "mean_token_accuracy": 0.8468797206878662,
      "num_tokens": 239592603.0,
      "step": 1503
    },
    {
      "epoch": 0.7650050864699899,
      "grad_norm": 1.1595673561096191,
      "learning_rate": 1e-05,
      "loss": 0.5495,
      "mean_token_accuracy": 0.8297867178916931,
      "num_tokens": 239745054.0,
      "step": 1504
    },
    {
      "epoch": 0.7655137334689726,
      "grad_norm": 1.01282799243927,
      "learning_rate": 1e-05,
      "loss": 0.5121,
      "mean_token_accuracy": 0.8400813341140747,
      "num_tokens": 239909253.0,
      "step": 1505
    },
    {
      "epoch": 0.7660223804679552,
      "grad_norm": 1.0849213600158691,
      "learning_rate": 1e-05,
      "loss": 0.513,
      "mean_token_accuracy": 0.8397663831710815,
      "num_tokens": 240051112.0,
      "step": 1506
    },
    {
      "epoch": 0.7665310274669379,
      "grad_norm": 1.0645639896392822,
      "learning_rate": 1e-05,
      "loss": 0.484,
      "mean_token_accuracy": 0.847365140914917,
      "num_tokens": 240211700.0,
      "step": 1507
    },
    {
      "epoch": 0.7670396744659207,
      "grad_norm": 1.019892930984497,
      "learning_rate": 1e-05,
      "loss": 0.5217,
      "mean_token_accuracy": 0.8375338912010193,
      "num_tokens": 240373170.0,
      "step": 1508
    },
    {
      "epoch": 0.7675483214649034,
      "grad_norm": 1.1592234373092651,
      "learning_rate": 1e-05,
      "loss": 0.5409,
      "mean_token_accuracy": 0.8331730365753174,
      "num_tokens": 240518936.0,
      "step": 1509
    },
    {
      "epoch": 0.768056968463886,
      "grad_norm": 1.040312647819519,
      "learning_rate": 1e-05,
      "loss": 0.5215,
      "mean_token_accuracy": 0.8367902636528015,
      "num_tokens": 240668912.0,
      "step": 1510
    },
    {
      "epoch": 0.7685656154628687,
      "grad_norm": 1.066318154335022,
      "learning_rate": 1e-05,
      "loss": 0.5073,
      "mean_token_accuracy": 0.8416951894760132,
      "num_tokens": 240828027.0,
      "step": 1511
    },
    {
      "epoch": 0.7690742624618515,
      "grad_norm": 0.9991273283958435,
      "learning_rate": 1e-05,
      "loss": 0.5108,
      "mean_token_accuracy": 0.840943455696106,
      "num_tokens": 240996115.0,
      "step": 1512
    },
    {
      "epoch": 0.7695829094608342,
      "grad_norm": 1.1857500076293945,
      "learning_rate": 1e-05,
      "loss": 0.5093,
      "mean_token_accuracy": 0.8417667746543884,
      "num_tokens": 241154020.0,
      "step": 1513
    },
    {
      "epoch": 0.7700915564598169,
      "grad_norm": 1.0184717178344727,
      "learning_rate": 1e-05,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8427598476409912,
      "num_tokens": 241313175.0,
      "step": 1514
    },
    {
      "epoch": 0.7706002034587996,
      "grad_norm": 1.2196848392486572,
      "learning_rate": 1e-05,
      "loss": 0.5209,
      "mean_token_accuracy": 0.8368412852287292,
      "num_tokens": 241461371.0,
      "step": 1515
    },
    {
      "epoch": 0.7711088504577823,
      "grad_norm": 1.0729477405548096,
      "learning_rate": 1e-05,
      "loss": 0.5422,
      "mean_token_accuracy": 0.8319259881973267,
      "num_tokens": 241631176.0,
      "step": 1516
    },
    {
      "epoch": 0.771617497456765,
      "grad_norm": 1.0394853353500366,
      "learning_rate": 1e-05,
      "loss": 0.5506,
      "mean_token_accuracy": 0.829657793045044,
      "num_tokens": 241791904.0,
      "step": 1517
    },
    {
      "epoch": 0.7721261444557477,
      "grad_norm": 1.1062549352645874,
      "learning_rate": 1e-05,
      "loss": 0.5106,
      "mean_token_accuracy": 0.8426122069358826,
      "num_tokens": 241954029.0,
      "step": 1518
    },
    {
      "epoch": 0.7726347914547305,
      "grad_norm": 1.0023213624954224,
      "learning_rate": 1e-05,
      "loss": 0.573,
      "mean_token_accuracy": 0.8251248598098755,
      "num_tokens": 242106558.0,
      "step": 1519
    },
    {
      "epoch": 0.7731434384537131,
      "grad_norm": 1.079978585243225,
      "learning_rate": 1e-05,
      "loss": 0.536,
      "mean_token_accuracy": 0.8343393802642822,
      "num_tokens": 242276624.0,
      "step": 1520
    },
    {
      "epoch": 0.7736520854526958,
      "grad_norm": 1.0364209413528442,
      "learning_rate": 1e-05,
      "loss": 0.529,
      "mean_token_accuracy": 0.8353438973426819,
      "num_tokens": 242434754.0,
      "step": 1521
    },
    {
      "epoch": 0.7741607324516785,
      "grad_norm": 1.0392584800720215,
      "learning_rate": 1e-05,
      "loss": 0.4997,
      "mean_token_accuracy": 0.8424927592277527,
      "num_tokens": 242576931.0,
      "step": 1522
    },
    {
      "epoch": 0.7746693794506613,
      "grad_norm": 1.045095443725586,
      "learning_rate": 1e-05,
      "loss": 0.5205,
      "mean_token_accuracy": 0.8376889228820801,
      "num_tokens": 242733793.0,
      "step": 1523
    },
    {
      "epoch": 0.775178026449644,
      "grad_norm": 1.132960557937622,
      "learning_rate": 1e-05,
      "loss": 0.5181,
      "mean_token_accuracy": 0.838951587677002,
      "num_tokens": 242882978.0,
      "step": 1524
    },
    {
      "epoch": 0.7756866734486266,
      "grad_norm": 1.0570266246795654,
      "learning_rate": 1e-05,
      "loss": 0.5178,
      "mean_token_accuracy": 0.8384246826171875,
      "num_tokens": 243038462.0,
      "step": 1525
    },
    {
      "epoch": 0.7761953204476093,
      "grad_norm": 1.1150120496749878,
      "learning_rate": 1e-05,
      "loss": 0.5308,
      "mean_token_accuracy": 0.8351293802261353,
      "num_tokens": 243206785.0,
      "step": 1526
    },
    {
      "epoch": 0.7767039674465921,
      "grad_norm": 0.9758465886116028,
      "learning_rate": 1e-05,
      "loss": 0.5166,
      "mean_token_accuracy": 0.839024543762207,
      "num_tokens": 243357645.0,
      "step": 1527
    },
    {
      "epoch": 0.7772126144455748,
      "grad_norm": 1.1001802682876587,
      "learning_rate": 1e-05,
      "loss": 0.5088,
      "mean_token_accuracy": 0.8403192758560181,
      "num_tokens": 243512602.0,
      "step": 1528
    },
    {
      "epoch": 0.7777212614445574,
      "grad_norm": 1.087286353111267,
      "learning_rate": 1e-05,
      "loss": 0.5116,
      "mean_token_accuracy": 0.839819073677063,
      "num_tokens": 243687260.0,
      "step": 1529
    },
    {
      "epoch": 0.7782299084435402,
      "grad_norm": 1.0088173151016235,
      "learning_rate": 1e-05,
      "loss": 0.5192,
      "mean_token_accuracy": 0.8393971920013428,
      "num_tokens": 243845125.0,
      "step": 1530
    },
    {
      "epoch": 0.7787385554425229,
      "grad_norm": 1.0728702545166016,
      "learning_rate": 1e-05,
      "loss": 0.4801,
      "mean_token_accuracy": 0.8486957550048828,
      "num_tokens": 243985343.0,
      "step": 1531
    },
    {
      "epoch": 0.7792472024415056,
      "grad_norm": 1.0450102090835571,
      "learning_rate": 1e-05,
      "loss": 0.5358,
      "mean_token_accuracy": 0.8348969221115112,
      "num_tokens": 244139431.0,
      "step": 1532
    },
    {
      "epoch": 0.7797558494404883,
      "grad_norm": 1.116742730140686,
      "learning_rate": 1e-05,
      "loss": 0.5322,
      "mean_token_accuracy": 0.8358186483383179,
      "num_tokens": 244296684.0,
      "step": 1533
    },
    {
      "epoch": 0.780264496439471,
      "grad_norm": 1.0600138902664185,
      "learning_rate": 1e-05,
      "loss": 0.527,
      "mean_token_accuracy": 0.8371978402137756,
      "num_tokens": 244456341.0,
      "step": 1534
    },
    {
      "epoch": 0.7807731434384537,
      "grad_norm": 1.1324113607406616,
      "learning_rate": 1e-05,
      "loss": 0.4989,
      "mean_token_accuracy": 0.8429820537567139,
      "num_tokens": 244612234.0,
      "step": 1535
    },
    {
      "epoch": 0.7812817904374364,
      "grad_norm": 1.0579875707626343,
      "learning_rate": 1e-05,
      "loss": 0.5438,
      "mean_token_accuracy": 0.8308662176132202,
      "num_tokens": 244763392.0,
      "step": 1536
    },
    {
      "epoch": 0.7817904374364191,
      "grad_norm": 1.105968713760376,
      "learning_rate": 1e-05,
      "loss": 0.5399,
      "mean_token_accuracy": 0.8314764499664307,
      "num_tokens": 244921544.0,
      "step": 1537
    },
    {
      "epoch": 0.7822990844354019,
      "grad_norm": 1.1011391878128052,
      "learning_rate": 1e-05,
      "loss": 0.5108,
      "mean_token_accuracy": 0.8399180173873901,
      "num_tokens": 245082597.0,
      "step": 1538
    },
    {
      "epoch": 0.7828077314343845,
      "grad_norm": 1.0996878147125244,
      "learning_rate": 1e-05,
      "loss": 0.5323,
      "mean_token_accuracy": 0.8340091705322266,
      "num_tokens": 245229522.0,
      "step": 1539
    },
    {
      "epoch": 0.7833163784333672,
      "grad_norm": 1.0516421794891357,
      "learning_rate": 1e-05,
      "loss": 0.5278,
      "mean_token_accuracy": 0.835968554019928,
      "num_tokens": 245379953.0,
      "step": 1540
    },
    {
      "epoch": 0.78382502543235,
      "grad_norm": 1.0144932270050049,
      "learning_rate": 1e-05,
      "loss": 0.5131,
      "mean_token_accuracy": 0.839628279209137,
      "num_tokens": 245531845.0,
      "step": 1541
    },
    {
      "epoch": 0.7843336724313327,
      "grad_norm": 1.0560461282730103,
      "learning_rate": 1e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8432492613792419,
      "num_tokens": 245696677.0,
      "step": 1542
    },
    {
      "epoch": 0.7848423194303153,
      "grad_norm": 1.0136598348617554,
      "learning_rate": 1e-05,
      "loss": 0.5166,
      "mean_token_accuracy": 0.8379706144332886,
      "num_tokens": 245866942.0,
      "step": 1543
    },
    {
      "epoch": 0.785350966429298,
      "grad_norm": 1.0351725816726685,
      "learning_rate": 1e-05,
      "loss": 0.5417,
      "mean_token_accuracy": 0.8329663276672363,
      "num_tokens": 246030786.0,
      "step": 1544
    },
    {
      "epoch": 0.7858596134282808,
      "grad_norm": 1.1024549007415771,
      "learning_rate": 1e-05,
      "loss": 0.5229,
      "mean_token_accuracy": 0.836799681186676,
      "num_tokens": 246181272.0,
      "step": 1545
    },
    {
      "epoch": 0.7863682604272635,
      "grad_norm": 1.1076233386993408,
      "learning_rate": 1e-05,
      "loss": 0.4818,
      "mean_token_accuracy": 0.8464345335960388,
      "num_tokens": 246332311.0,
      "step": 1546
    },
    {
      "epoch": 0.7868769074262462,
      "grad_norm": 1.084722876548767,
      "learning_rate": 1e-05,
      "loss": 0.5509,
      "mean_token_accuracy": 0.830653190612793,
      "num_tokens": 246491365.0,
      "step": 1547
    },
    {
      "epoch": 0.7873855544252288,
      "grad_norm": 1.0792385339736938,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8466219902038574,
      "num_tokens": 246646398.0,
      "step": 1548
    },
    {
      "epoch": 0.7878942014242116,
      "grad_norm": 1.161864995956421,
      "learning_rate": 1e-05,
      "loss": 0.5148,
      "mean_token_accuracy": 0.83869469165802,
      "num_tokens": 246806249.0,
      "step": 1549
    },
    {
      "epoch": 0.7884028484231943,
      "grad_norm": 1.0989301204681396,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.8468911051750183,
      "num_tokens": 246957754.0,
      "step": 1550
    },
    {
      "epoch": 0.788911495422177,
      "grad_norm": 0.945274829864502,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8512423038482666,
      "num_tokens": 247122388.0,
      "step": 1551
    },
    {
      "epoch": 0.7894201424211598,
      "grad_norm": 1.0815494060516357,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8495436310768127,
      "num_tokens": 247284678.0,
      "step": 1552
    },
    {
      "epoch": 0.7899287894201424,
      "grad_norm": 1.0339261293411255,
      "learning_rate": 1e-05,
      "loss": 0.5423,
      "mean_token_accuracy": 0.8329569101333618,
      "num_tokens": 247442396.0,
      "step": 1553
    },
    {
      "epoch": 0.7904374364191251,
      "grad_norm": 3.8189172744750977,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.8405753970146179,
      "num_tokens": 247601438.0,
      "step": 1554
    },
    {
      "epoch": 0.7909460834181078,
      "grad_norm": 1.2736577987670898,
      "learning_rate": 1e-05,
      "loss": 0.5051,
      "mean_token_accuracy": 0.8410643339157104,
      "num_tokens": 247766929.0,
      "step": 1555
    },
    {
      "epoch": 0.7914547304170906,
      "grad_norm": 1.0385481119155884,
      "learning_rate": 1e-05,
      "loss": 0.533,
      "mean_token_accuracy": 0.835472583770752,
      "num_tokens": 247935788.0,
      "step": 1556
    },
    {
      "epoch": 0.7919633774160733,
      "grad_norm": 1.1767652034759521,
      "learning_rate": 1e-05,
      "loss": 0.499,
      "mean_token_accuracy": 0.845313310623169,
      "num_tokens": 248102197.0,
      "step": 1557
    },
    {
      "epoch": 0.7924720244150559,
      "grad_norm": 1.1459380388259888,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.8455806374549866,
      "num_tokens": 248252748.0,
      "step": 1558
    },
    {
      "epoch": 0.7929806714140386,
      "grad_norm": 1.030868649482727,
      "learning_rate": 1e-05,
      "loss": 0.5086,
      "mean_token_accuracy": 0.8415130376815796,
      "num_tokens": 248410690.0,
      "step": 1559
    },
    {
      "epoch": 0.7934893184130214,
      "grad_norm": 1.288917899131775,
      "learning_rate": 1e-05,
      "loss": 0.5049,
      "mean_token_accuracy": 0.8433955907821655,
      "num_tokens": 248562629.0,
      "step": 1560
    },
    {
      "epoch": 0.7939979654120041,
      "grad_norm": 1.0981899499893188,
      "learning_rate": 1e-05,
      "loss": 0.5357,
      "mean_token_accuracy": 0.8328940868377686,
      "num_tokens": 248724426.0,
      "step": 1561
    },
    {
      "epoch": 0.7945066124109867,
      "grad_norm": 1.112394094467163,
      "learning_rate": 1e-05,
      "loss": 0.533,
      "mean_token_accuracy": 0.8356070518493652,
      "num_tokens": 248884711.0,
      "step": 1562
    },
    {
      "epoch": 0.7950152594099695,
      "grad_norm": 1.0571964979171753,
      "learning_rate": 1e-05,
      "loss": 0.5299,
      "mean_token_accuracy": 0.8359705209732056,
      "num_tokens": 249039319.0,
      "step": 1563
    },
    {
      "epoch": 0.7955239064089522,
      "grad_norm": 1.1130516529083252,
      "learning_rate": 1e-05,
      "loss": 0.5171,
      "mean_token_accuracy": 0.8393397927284241,
      "num_tokens": 249198668.0,
      "step": 1564
    },
    {
      "epoch": 0.7960325534079349,
      "grad_norm": 1.1128276586532593,
      "learning_rate": 1e-05,
      "loss": 0.5168,
      "mean_token_accuracy": 0.837859570980072,
      "num_tokens": 249355386.0,
      "step": 1565
    },
    {
      "epoch": 0.7965412004069176,
      "grad_norm": 1.033722162246704,
      "learning_rate": 1e-05,
      "loss": 0.5328,
      "mean_token_accuracy": 0.8348727226257324,
      "num_tokens": 249506726.0,
      "step": 1566
    },
    {
      "epoch": 0.7970498474059003,
      "grad_norm": 1.0181803703308105,
      "learning_rate": 1e-05,
      "loss": 0.5501,
      "mean_token_accuracy": 0.830032467842102,
      "num_tokens": 249668433.0,
      "step": 1567
    },
    {
      "epoch": 0.797558494404883,
      "grad_norm": 1.0626704692840576,
      "learning_rate": 1e-05,
      "loss": 0.4986,
      "mean_token_accuracy": 0.8434553146362305,
      "num_tokens": 249815595.0,
      "step": 1568
    },
    {
      "epoch": 0.7980671414038657,
      "grad_norm": 1.022221565246582,
      "learning_rate": 1e-05,
      "loss": 0.5164,
      "mean_token_accuracy": 0.837891697883606,
      "num_tokens": 249977103.0,
      "step": 1569
    },
    {
      "epoch": 0.7985757884028484,
      "grad_norm": 0.9750449657440186,
      "learning_rate": 1e-05,
      "loss": 0.5242,
      "mean_token_accuracy": 0.8367565870285034,
      "num_tokens": 250145992.0,
      "step": 1570
    },
    {
      "epoch": 0.7990844354018312,
      "grad_norm": 1.1764782667160034,
      "learning_rate": 1e-05,
      "loss": 0.5263,
      "mean_token_accuracy": 0.8365598320960999,
      "num_tokens": 250280246.0,
      "step": 1571
    },
    {
      "epoch": 0.7995930824008138,
      "grad_norm": 0.9848682880401611,
      "learning_rate": 1e-05,
      "loss": 0.5299,
      "mean_token_accuracy": 0.8356618881225586,
      "num_tokens": 250443792.0,
      "step": 1572
    },
    {
      "epoch": 0.8001017293997965,
      "grad_norm": 0.9653362035751343,
      "learning_rate": 1e-05,
      "loss": 0.5148,
      "mean_token_accuracy": 0.8378146290779114,
      "num_tokens": 250599631.0,
      "step": 1573
    },
    {
      "epoch": 0.8006103763987793,
      "grad_norm": 1.061545968055725,
      "learning_rate": 1e-05,
      "loss": 0.5423,
      "mean_token_accuracy": 0.8325547575950623,
      "num_tokens": 250747877.0,
      "step": 1574
    },
    {
      "epoch": 0.801119023397762,
      "grad_norm": 1.0658345222473145,
      "learning_rate": 1e-05,
      "loss": 0.5348,
      "mean_token_accuracy": 0.833031952381134,
      "num_tokens": 250900217.0,
      "step": 1575
    },
    {
      "epoch": 0.8016276703967447,
      "grad_norm": 0.9847412109375,
      "learning_rate": 1e-05,
      "loss": 0.4963,
      "mean_token_accuracy": 0.8432402610778809,
      "num_tokens": 251052502.0,
      "step": 1576
    },
    {
      "epoch": 0.8021363173957273,
      "grad_norm": 0.9033166170120239,
      "learning_rate": 1e-05,
      "loss": 0.4988,
      "mean_token_accuracy": 0.8441382050514221,
      "num_tokens": 251208564.0,
      "step": 1577
    },
    {
      "epoch": 0.8026449643947101,
      "grad_norm": 1.023217797279358,
      "learning_rate": 1e-05,
      "loss": 0.5155,
      "mean_token_accuracy": 0.8398065567016602,
      "num_tokens": 251352775.0,
      "step": 1578
    },
    {
      "epoch": 0.8031536113936928,
      "grad_norm": 0.9612480401992798,
      "learning_rate": 1e-05,
      "loss": 0.5119,
      "mean_token_accuracy": 0.8414446711540222,
      "num_tokens": 251518171.0,
      "step": 1579
    },
    {
      "epoch": 0.8036622583926755,
      "grad_norm": 0.9663166999816895,
      "learning_rate": 1e-05,
      "loss": 0.5203,
      "mean_token_accuracy": 0.8399222493171692,
      "num_tokens": 251677652.0,
      "step": 1580
    },
    {
      "epoch": 0.8041709053916581,
      "grad_norm": 1.036909580230713,
      "learning_rate": 1e-05,
      "loss": 0.5303,
      "mean_token_accuracy": 0.8345316648483276,
      "num_tokens": 251821333.0,
      "step": 1581
    },
    {
      "epoch": 0.8046795523906409,
      "grad_norm": 0.984411358833313,
      "learning_rate": 1e-05,
      "loss": 0.575,
      "mean_token_accuracy": 0.8244682550430298,
      "num_tokens": 251995232.0,
      "step": 1582
    },
    {
      "epoch": 0.8051881993896236,
      "grad_norm": 0.9926067590713501,
      "learning_rate": 1e-05,
      "loss": 0.5079,
      "mean_token_accuracy": 0.8403122425079346,
      "num_tokens": 252151382.0,
      "step": 1583
    },
    {
      "epoch": 0.8056968463886063,
      "grad_norm": 0.9969213008880615,
      "learning_rate": 1e-05,
      "loss": 0.5407,
      "mean_token_accuracy": 0.8337632417678833,
      "num_tokens": 252318139.0,
      "step": 1584
    },
    {
      "epoch": 0.8062054933875891,
      "grad_norm": 1.0000914335250854,
      "learning_rate": 1e-05,
      "loss": 0.5066,
      "mean_token_accuracy": 0.8416249752044678,
      "num_tokens": 252479916.0,
      "step": 1585
    },
    {
      "epoch": 0.8067141403865717,
      "grad_norm": 0.9684925079345703,
      "learning_rate": 1e-05,
      "loss": 0.5192,
      "mean_token_accuracy": 0.8398261070251465,
      "num_tokens": 252655015.0,
      "step": 1586
    },
    {
      "epoch": 0.8072227873855544,
      "grad_norm": 1.0065932273864746,
      "learning_rate": 1e-05,
      "loss": 0.5135,
      "mean_token_accuracy": 0.8411272168159485,
      "num_tokens": 252810669.0,
      "step": 1587
    },
    {
      "epoch": 0.8077314343845371,
      "grad_norm": 1.0063505172729492,
      "learning_rate": 1e-05,
      "loss": 0.5104,
      "mean_token_accuracy": 0.8413547873497009,
      "num_tokens": 252975872.0,
      "step": 1588
    },
    {
      "epoch": 0.8082400813835199,
      "grad_norm": 1.00690495967865,
      "learning_rate": 1e-05,
      "loss": 0.5231,
      "mean_token_accuracy": 0.8373821377754211,
      "num_tokens": 253140927.0,
      "step": 1589
    },
    {
      "epoch": 0.8087487283825026,
      "grad_norm": 1.0648139715194702,
      "learning_rate": 1e-05,
      "loss": 0.5302,
      "mean_token_accuracy": 0.8346395492553711,
      "num_tokens": 253298562.0,
      "step": 1590
    },
    {
      "epoch": 0.8092573753814852,
      "grad_norm": 0.9983528256416321,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.836251974105835,
      "num_tokens": 253469245.0,
      "step": 1591
    },
    {
      "epoch": 0.8097660223804679,
      "grad_norm": 0.987586259841919,
      "learning_rate": 1e-05,
      "loss": 0.4635,
      "mean_token_accuracy": 0.8542689085006714,
      "num_tokens": 253623711.0,
      "step": 1592
    },
    {
      "epoch": 0.8102746693794507,
      "grad_norm": 1.0418356657028198,
      "learning_rate": 1e-05,
      "loss": 0.5004,
      "mean_token_accuracy": 0.8439185619354248,
      "num_tokens": 253793013.0,
      "step": 1593
    },
    {
      "epoch": 0.8107833163784334,
      "grad_norm": 1.1564536094665527,
      "learning_rate": 1e-05,
      "loss": 0.556,
      "mean_token_accuracy": 0.827933669090271,
      "num_tokens": 253951728.0,
      "step": 1594
    },
    {
      "epoch": 0.811291963377416,
      "grad_norm": 1.0996955633163452,
      "learning_rate": 1e-05,
      "loss": 0.5163,
      "mean_token_accuracy": 0.8392601013183594,
      "num_tokens": 254107635.0,
      "step": 1595
    },
    {
      "epoch": 0.8118006103763988,
      "grad_norm": 1.0114479064941406,
      "learning_rate": 1e-05,
      "loss": 0.5291,
      "mean_token_accuracy": 0.8350151777267456,
      "num_tokens": 254256769.0,
      "step": 1596
    },
    {
      "epoch": 0.8123092573753815,
      "grad_norm": 1.1125953197479248,
      "learning_rate": 1e-05,
      "loss": 0.5573,
      "mean_token_accuracy": 0.8287286162376404,
      "num_tokens": 254426121.0,
      "step": 1597
    },
    {
      "epoch": 0.8128179043743642,
      "grad_norm": 0.9474158883094788,
      "learning_rate": 1e-05,
      "loss": 0.5213,
      "mean_token_accuracy": 0.8386734127998352,
      "num_tokens": 254592698.0,
      "step": 1598
    },
    {
      "epoch": 0.8133265513733469,
      "grad_norm": 0.9946539998054504,
      "learning_rate": 1e-05,
      "loss": 0.5251,
      "mean_token_accuracy": 0.8364081382751465,
      "num_tokens": 254757687.0,
      "step": 1599
    },
    {
      "epoch": 0.8138351983723296,
      "grad_norm": 1.0554031133651733,
      "learning_rate": 1e-05,
      "loss": 0.5254,
      "mean_token_accuracy": 0.8361053466796875,
      "num_tokens": 254907582.0,
      "step": 1600
    },
    {
      "epoch": 0.8143438453713123,
      "grad_norm": 1.0031455755233765,
      "learning_rate": 1e-05,
      "loss": 0.5018,
      "mean_token_accuracy": 0.8430084586143494,
      "num_tokens": 255060864.0,
      "step": 1601
    },
    {
      "epoch": 0.814852492370295,
      "grad_norm": 1.090097427368164,
      "learning_rate": 1e-05,
      "loss": 0.4965,
      "mean_token_accuracy": 0.8441165685653687,
      "num_tokens": 255220612.0,
      "step": 1602
    },
    {
      "epoch": 0.8153611393692777,
      "grad_norm": 1.1100863218307495,
      "learning_rate": 1e-05,
      "loss": 0.5731,
      "mean_token_accuracy": 0.8228561878204346,
      "num_tokens": 255381676.0,
      "step": 1603
    },
    {
      "epoch": 0.8158697863682605,
      "grad_norm": 1.1037805080413818,
      "learning_rate": 1e-05,
      "loss": 0.4941,
      "mean_token_accuracy": 0.8476219177246094,
      "num_tokens": 255530491.0,
      "step": 1604
    },
    {
      "epoch": 0.8163784333672431,
      "grad_norm": 1.0004202127456665,
      "learning_rate": 1e-05,
      "loss": 0.5148,
      "mean_token_accuracy": 0.840594470500946,
      "num_tokens": 255693669.0,
      "step": 1605
    },
    {
      "epoch": 0.8168870803662258,
      "grad_norm": 1.0348905324935913,
      "learning_rate": 1e-05,
      "loss": 0.5237,
      "mean_token_accuracy": 0.8360775709152222,
      "num_tokens": 255861693.0,
      "step": 1606
    },
    {
      "epoch": 0.8173957273652085,
      "grad_norm": 1.048621416091919,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8353121876716614,
      "num_tokens": 256028864.0,
      "step": 1607
    },
    {
      "epoch": 0.8179043743641913,
      "grad_norm": 1.1196389198303223,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8500303626060486,
      "num_tokens": 256184434.0,
      "step": 1608
    },
    {
      "epoch": 0.818413021363174,
      "grad_norm": 1.0022071599960327,
      "learning_rate": 1e-05,
      "loss": 0.484,
      "mean_token_accuracy": 0.847318708896637,
      "num_tokens": 256342000.0,
      "step": 1609
    },
    {
      "epoch": 0.8189216683621566,
      "grad_norm": 1.1261740922927856,
      "learning_rate": 1e-05,
      "loss": 0.5286,
      "mean_token_accuracy": 0.8349587917327881,
      "num_tokens": 256517073.0,
      "step": 1610
    },
    {
      "epoch": 0.8194303153611394,
      "grad_norm": 1.0318447351455688,
      "learning_rate": 1e-05,
      "loss": 0.4905,
      "mean_token_accuracy": 0.8450832366943359,
      "num_tokens": 256674055.0,
      "step": 1611
    },
    {
      "epoch": 0.8199389623601221,
      "grad_norm": 1.1506813764572144,
      "learning_rate": 1e-05,
      "loss": 0.5446,
      "mean_token_accuracy": 0.8306101560592651,
      "num_tokens": 256848111.0,
      "step": 1612
    },
    {
      "epoch": 0.8204476093591048,
      "grad_norm": 1.055451512336731,
      "learning_rate": 1e-05,
      "loss": 0.5006,
      "mean_token_accuracy": 0.8442692756652832,
      "num_tokens": 257016873.0,
      "step": 1613
    },
    {
      "epoch": 0.8209562563580874,
      "grad_norm": 1.036156415939331,
      "learning_rate": 1e-05,
      "loss": 0.5416,
      "mean_token_accuracy": 0.8340879678726196,
      "num_tokens": 257182035.0,
      "step": 1614
    },
    {
      "epoch": 0.8214649033570702,
      "grad_norm": 1.0551000833511353,
      "learning_rate": 1e-05,
      "loss": 0.5548,
      "mean_token_accuracy": 0.827684223651886,
      "num_tokens": 257351888.0,
      "step": 1615
    },
    {
      "epoch": 0.8219735503560529,
      "grad_norm": 0.9917570352554321,
      "learning_rate": 1e-05,
      "loss": 0.5212,
      "mean_token_accuracy": 0.8370511531829834,
      "num_tokens": 257508833.0,
      "step": 1616
    },
    {
      "epoch": 0.8224821973550356,
      "grad_norm": 1.1158199310302734,
      "learning_rate": 1e-05,
      "loss": 0.5041,
      "mean_token_accuracy": 0.8426960706710815,
      "num_tokens": 257666306.0,
      "step": 1617
    },
    {
      "epoch": 0.8229908443540183,
      "grad_norm": 0.9963831901550293,
      "learning_rate": 1e-05,
      "loss": 0.522,
      "mean_token_accuracy": 0.8374859690666199,
      "num_tokens": 257832236.0,
      "step": 1618
    },
    {
      "epoch": 0.823499491353001,
      "grad_norm": 1.141318917274475,
      "learning_rate": 1e-05,
      "loss": 0.4946,
      "mean_token_accuracy": 0.8444522619247437,
      "num_tokens": 257996091.0,
      "step": 1619
    },
    {
      "epoch": 0.8240081383519837,
      "grad_norm": 0.9614141583442688,
      "learning_rate": 1e-05,
      "loss": 0.5153,
      "mean_token_accuracy": 0.8377129435539246,
      "num_tokens": 258156268.0,
      "step": 1620
    },
    {
      "epoch": 0.8245167853509664,
      "grad_norm": 1.0855903625488281,
      "learning_rate": 1e-05,
      "loss": 0.5324,
      "mean_token_accuracy": 0.8347781300544739,
      "num_tokens": 258323374.0,
      "step": 1621
    },
    {
      "epoch": 0.8250254323499492,
      "grad_norm": 0.993126630783081,
      "learning_rate": 1e-05,
      "loss": 0.5211,
      "mean_token_accuracy": 0.8383423089981079,
      "num_tokens": 258495850.0,
      "step": 1622
    },
    {
      "epoch": 0.8255340793489319,
      "grad_norm": 1.0016838312149048,
      "learning_rate": 1e-05,
      "loss": 0.519,
      "mean_token_accuracy": 0.8375632166862488,
      "num_tokens": 258653458.0,
      "step": 1623
    },
    {
      "epoch": 0.8260427263479145,
      "grad_norm": 1.0147916078567505,
      "learning_rate": 1e-05,
      "loss": 0.5316,
      "mean_token_accuracy": 0.8373314738273621,
      "num_tokens": 258823978.0,
      "step": 1624
    },
    {
      "epoch": 0.8265513733468972,
      "grad_norm": 1.0405744314193726,
      "learning_rate": 1e-05,
      "loss": 0.5138,
      "mean_token_accuracy": 0.8393408060073853,
      "num_tokens": 258981414.0,
      "step": 1625
    },
    {
      "epoch": 0.82706002034588,
      "grad_norm": 0.9844928979873657,
      "learning_rate": 1e-05,
      "loss": 0.5174,
      "mean_token_accuracy": 0.837151288986206,
      "num_tokens": 259143758.0,
      "step": 1626
    },
    {
      "epoch": 0.8275686673448627,
      "grad_norm": 1.1062935590744019,
      "learning_rate": 1e-05,
      "loss": 0.5362,
      "mean_token_accuracy": 0.834156334400177,
      "num_tokens": 259297261.0,
      "step": 1627
    },
    {
      "epoch": 0.8280773143438453,
      "grad_norm": 1.0649558305740356,
      "learning_rate": 1e-05,
      "loss": 0.5231,
      "mean_token_accuracy": 0.8374667167663574,
      "num_tokens": 259459034.0,
      "step": 1628
    },
    {
      "epoch": 0.828585961342828,
      "grad_norm": 0.9770439267158508,
      "learning_rate": 1e-05,
      "loss": 0.5036,
      "mean_token_accuracy": 0.8402736783027649,
      "num_tokens": 259614641.0,
      "step": 1629
    },
    {
      "epoch": 0.8290946083418108,
      "grad_norm": 1.07271146774292,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.8429797291755676,
      "num_tokens": 259769628.0,
      "step": 1630
    },
    {
      "epoch": 0.8296032553407935,
      "grad_norm": 1.238389253616333,
      "learning_rate": 1e-05,
      "loss": 0.5086,
      "mean_token_accuracy": 0.8421880006790161,
      "num_tokens": 259925515.0,
      "step": 1631
    },
    {
      "epoch": 0.8301119023397762,
      "grad_norm": 1.0082037448883057,
      "learning_rate": 1e-05,
      "loss": 0.5637,
      "mean_token_accuracy": 0.8264005184173584,
      "num_tokens": 260091045.0,
      "step": 1632
    },
    {
      "epoch": 0.830620549338759,
      "grad_norm": 1.0819079875946045,
      "learning_rate": 1e-05,
      "loss": 0.5048,
      "mean_token_accuracy": 0.8421937227249146,
      "num_tokens": 260243318.0,
      "step": 1633
    },
    {
      "epoch": 0.8311291963377416,
      "grad_norm": 1.0576622486114502,
      "learning_rate": 1e-05,
      "loss": 0.5265,
      "mean_token_accuracy": 0.8375747203826904,
      "num_tokens": 260414955.0,
      "step": 1634
    },
    {
      "epoch": 0.8316378433367243,
      "grad_norm": 1.120271921157837,
      "learning_rate": 1e-05,
      "loss": 0.515,
      "mean_token_accuracy": 0.8407070636749268,
      "num_tokens": 260573372.0,
      "step": 1635
    },
    {
      "epoch": 0.832146490335707,
      "grad_norm": 0.9665607213973999,
      "learning_rate": 1e-05,
      "loss": 0.4649,
      "mean_token_accuracy": 0.8517748117446899,
      "num_tokens": 260738547.0,
      "step": 1636
    },
    {
      "epoch": 0.8326551373346898,
      "grad_norm": 1.0370997190475464,
      "learning_rate": 1e-05,
      "loss": 0.5054,
      "mean_token_accuracy": 0.8430832624435425,
      "num_tokens": 260892133.0,
      "step": 1637
    },
    {
      "epoch": 0.8331637843336724,
      "grad_norm": 1.1478748321533203,
      "learning_rate": 1e-05,
      "loss": 0.5364,
      "mean_token_accuracy": 0.8316822052001953,
      "num_tokens": 261051308.0,
      "step": 1638
    },
    {
      "epoch": 0.8336724313326551,
      "grad_norm": 1.0233328342437744,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8349559307098389,
      "num_tokens": 261205776.0,
      "step": 1639
    },
    {
      "epoch": 0.8341810783316378,
      "grad_norm": 1.065025806427002,
      "learning_rate": 1e-05,
      "loss": 0.4871,
      "mean_token_accuracy": 0.8467060327529907,
      "num_tokens": 261358402.0,
      "step": 1640
    },
    {
      "epoch": 0.8346897253306206,
      "grad_norm": 1.1216325759887695,
      "learning_rate": 1e-05,
      "loss": 0.5393,
      "mean_token_accuracy": 0.8320754766464233,
      "num_tokens": 261524731.0,
      "step": 1641
    },
    {
      "epoch": 0.8351983723296033,
      "grad_norm": 0.9606276750564575,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.8425238132476807,
      "num_tokens": 261687233.0,
      "step": 1642
    },
    {
      "epoch": 0.8357070193285859,
      "grad_norm": 1.090203046798706,
      "learning_rate": 1e-05,
      "loss": 0.5293,
      "mean_token_accuracy": 0.8350657224655151,
      "num_tokens": 261854243.0,
      "step": 1643
    },
    {
      "epoch": 0.8362156663275687,
      "grad_norm": 0.9650431275367737,
      "learning_rate": 1e-05,
      "loss": 0.5342,
      "mean_token_accuracy": 0.8349401950836182,
      "num_tokens": 262026963.0,
      "step": 1644
    },
    {
      "epoch": 0.8367243133265514,
      "grad_norm": 0.9442290663719177,
      "learning_rate": 1e-05,
      "loss": 0.4959,
      "mean_token_accuracy": 0.8443392515182495,
      "num_tokens": 262196156.0,
      "step": 1645
    },
    {
      "epoch": 0.8372329603255341,
      "grad_norm": 1.0255963802337646,
      "learning_rate": 1e-05,
      "loss": 0.5325,
      "mean_token_accuracy": 0.8355637788772583,
      "num_tokens": 262365347.0,
      "step": 1646
    },
    {
      "epoch": 0.8377416073245167,
      "grad_norm": 1.0480962991714478,
      "learning_rate": 1e-05,
      "loss": 0.5263,
      "mean_token_accuracy": 0.837949812412262,
      "num_tokens": 262519307.0,
      "step": 1647
    },
    {
      "epoch": 0.8382502543234995,
      "grad_norm": 1.1212228536605835,
      "learning_rate": 1e-05,
      "loss": 0.5002,
      "mean_token_accuracy": 0.84267258644104,
      "num_tokens": 262673650.0,
      "step": 1648
    },
    {
      "epoch": 0.8387589013224822,
      "grad_norm": 1.1384960412979126,
      "learning_rate": 1e-05,
      "loss": 0.518,
      "mean_token_accuracy": 0.8377788066864014,
      "num_tokens": 262812546.0,
      "step": 1649
    },
    {
      "epoch": 0.8392675483214649,
      "grad_norm": 1.0220293998718262,
      "learning_rate": 1e-05,
      "loss": 0.5455,
      "mean_token_accuracy": 0.8309825658798218,
      "num_tokens": 262970304.0,
      "step": 1650
    },
    {
      "epoch": 0.8397761953204476,
      "grad_norm": 1.1437078714370728,
      "learning_rate": 1e-05,
      "loss": 0.5425,
      "mean_token_accuracy": 0.8323178887367249,
      "num_tokens": 263132275.0,
      "step": 1651
    },
    {
      "epoch": 0.8402848423194303,
      "grad_norm": 1.028165578842163,
      "learning_rate": 1e-05,
      "loss": 0.518,
      "mean_token_accuracy": 0.8380722403526306,
      "num_tokens": 263293537.0,
      "step": 1652
    },
    {
      "epoch": 0.840793489318413,
      "grad_norm": 0.9466443657875061,
      "learning_rate": 1e-05,
      "loss": 0.5183,
      "mean_token_accuracy": 0.8380733728408813,
      "num_tokens": 263455476.0,
      "step": 1653
    },
    {
      "epoch": 0.8413021363173957,
      "grad_norm": 1.03770911693573,
      "learning_rate": 1e-05,
      "loss": 0.5061,
      "mean_token_accuracy": 0.8408149480819702,
      "num_tokens": 263615516.0,
      "step": 1654
    },
    {
      "epoch": 0.8418107833163785,
      "grad_norm": 1.0772814750671387,
      "learning_rate": 1e-05,
      "loss": 0.5046,
      "mean_token_accuracy": 0.8437204360961914,
      "num_tokens": 263759112.0,
      "step": 1655
    },
    {
      "epoch": 0.8423194303153612,
      "grad_norm": 0.9608617424964905,
      "learning_rate": 1e-05,
      "loss": 0.5035,
      "mean_token_accuracy": 0.8425559401512146,
      "num_tokens": 263924035.0,
      "step": 1656
    },
    {
      "epoch": 0.8428280773143438,
      "grad_norm": 0.9779507517814636,
      "learning_rate": 1e-05,
      "loss": 0.5685,
      "mean_token_accuracy": 0.8232150077819824,
      "num_tokens": 264096716.0,
      "step": 1657
    },
    {
      "epoch": 0.8433367243133265,
      "grad_norm": 0.9801053404808044,
      "learning_rate": 1e-05,
      "loss": 0.5298,
      "mean_token_accuracy": 0.833958625793457,
      "num_tokens": 264253124.0,
      "step": 1658
    },
    {
      "epoch": 0.8438453713123093,
      "grad_norm": 0.9987591505050659,
      "learning_rate": 1e-05,
      "loss": 0.4906,
      "mean_token_accuracy": 0.8448872566223145,
      "num_tokens": 264412558.0,
      "step": 1659
    },
    {
      "epoch": 0.844354018311292,
      "grad_norm": 0.9199841022491455,
      "learning_rate": 1e-05,
      "loss": 0.5281,
      "mean_token_accuracy": 0.8362597227096558,
      "num_tokens": 264578027.0,
      "step": 1660
    },
    {
      "epoch": 0.8448626653102747,
      "grad_norm": 1.0133132934570312,
      "learning_rate": 1e-05,
      "loss": 0.5204,
      "mean_token_accuracy": 0.8379366397857666,
      "num_tokens": 264737779.0,
      "step": 1661
    },
    {
      "epoch": 0.8453713123092573,
      "grad_norm": 1.0267291069030762,
      "learning_rate": 1e-05,
      "loss": 0.5308,
      "mean_token_accuracy": 0.8346647024154663,
      "num_tokens": 264888054.0,
      "step": 1662
    },
    {
      "epoch": 0.8458799593082401,
      "grad_norm": 0.9384878873825073,
      "learning_rate": 1e-05,
      "loss": 0.5277,
      "mean_token_accuracy": 0.8353392481803894,
      "num_tokens": 265048672.0,
      "step": 1663
    },
    {
      "epoch": 0.8463886063072228,
      "grad_norm": 3.248255729675293,
      "learning_rate": 1e-05,
      "loss": 0.515,
      "mean_token_accuracy": 0.840934157371521,
      "num_tokens": 265191225.0,
      "step": 1664
    },
    {
      "epoch": 0.8468972533062055,
      "grad_norm": 1.18904447555542,
      "learning_rate": 1e-05,
      "loss": 0.5388,
      "mean_token_accuracy": 0.8337969183921814,
      "num_tokens": 265357998.0,
      "step": 1665
    },
    {
      "epoch": 0.8474059003051883,
      "grad_norm": 0.9972271919250488,
      "learning_rate": 1e-05,
      "loss": 0.5118,
      "mean_token_accuracy": 0.8408945202827454,
      "num_tokens": 265529823.0,
      "step": 1666
    },
    {
      "epoch": 0.8479145473041709,
      "grad_norm": 1.0963144302368164,
      "learning_rate": 1e-05,
      "loss": 0.5272,
      "mean_token_accuracy": 0.8357887864112854,
      "num_tokens": 265683573.0,
      "step": 1667
    },
    {
      "epoch": 0.8484231943031536,
      "grad_norm": 1.0239129066467285,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8485143184661865,
      "num_tokens": 265845965.0,
      "step": 1668
    },
    {
      "epoch": 0.8489318413021363,
      "grad_norm": 1.1555248498916626,
      "learning_rate": 1e-05,
      "loss": 0.5508,
      "mean_token_accuracy": 0.8298073410987854,
      "num_tokens": 265991385.0,
      "step": 1669
    },
    {
      "epoch": 0.8494404883011191,
      "grad_norm": 1.1584995985031128,
      "learning_rate": 1e-05,
      "loss": 0.5253,
      "mean_token_accuracy": 0.8365756273269653,
      "num_tokens": 266161615.0,
      "step": 1670
    },
    {
      "epoch": 0.8499491353001017,
      "grad_norm": 1.168226957321167,
      "learning_rate": 1e-05,
      "loss": 0.5393,
      "mean_token_accuracy": 0.8331434726715088,
      "num_tokens": 266310118.0,
      "step": 1671
    },
    {
      "epoch": 0.8504577822990844,
      "grad_norm": 1.0551339387893677,
      "learning_rate": 1e-05,
      "loss": 0.5057,
      "mean_token_accuracy": 0.8417863845825195,
      "num_tokens": 266462436.0,
      "step": 1672
    },
    {
      "epoch": 0.8509664292980671,
      "grad_norm": 1.0659514665603638,
      "learning_rate": 1e-05,
      "loss": 0.5139,
      "mean_token_accuracy": 0.8392844200134277,
      "num_tokens": 266613253.0,
      "step": 1673
    },
    {
      "epoch": 0.8514750762970499,
      "grad_norm": 1.0820661783218384,
      "learning_rate": 1e-05,
      "loss": 0.52,
      "mean_token_accuracy": 0.8378826379776001,
      "num_tokens": 266773293.0,
      "step": 1674
    },
    {
      "epoch": 0.8519837232960326,
      "grad_norm": 1.0080983638763428,
      "learning_rate": 1e-05,
      "loss": 0.5274,
      "mean_token_accuracy": 0.835020899772644,
      "num_tokens": 266935261.0,
      "step": 1675
    },
    {
      "epoch": 0.8524923702950152,
      "grad_norm": 1.1683578491210938,
      "learning_rate": 1e-05,
      "loss": 0.5339,
      "mean_token_accuracy": 0.8347437977790833,
      "num_tokens": 267100828.0,
      "step": 1676
    },
    {
      "epoch": 0.853001017293998,
      "grad_norm": 1.112990140914917,
      "learning_rate": 1e-05,
      "loss": 0.5207,
      "mean_token_accuracy": 0.8374387621879578,
      "num_tokens": 267269826.0,
      "step": 1677
    },
    {
      "epoch": 0.8535096642929807,
      "grad_norm": 0.9589158296585083,
      "learning_rate": 1e-05,
      "loss": 0.5052,
      "mean_token_accuracy": 0.8411589860916138,
      "num_tokens": 267429972.0,
      "step": 1678
    },
    {
      "epoch": 0.8540183112919634,
      "grad_norm": 1.1416692733764648,
      "learning_rate": 1e-05,
      "loss": 0.5185,
      "mean_token_accuracy": 0.8388045430183411,
      "num_tokens": 267590240.0,
      "step": 1679
    },
    {
      "epoch": 0.854526958290946,
      "grad_norm": 0.9843428730964661,
      "learning_rate": 1e-05,
      "loss": 0.4795,
      "mean_token_accuracy": 0.8487104773521423,
      "num_tokens": 267748631.0,
      "step": 1680
    },
    {
      "epoch": 0.8550356052899288,
      "grad_norm": 0.9813622236251831,
      "learning_rate": 1e-05,
      "loss": 0.5237,
      "mean_token_accuracy": 0.8359060287475586,
      "num_tokens": 267906365.0,
      "step": 1681
    },
    {
      "epoch": 0.8555442522889115,
      "grad_norm": 1.0184063911437988,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.843595027923584,
      "num_tokens": 268058483.0,
      "step": 1682
    },
    {
      "epoch": 0.8560528992878942,
      "grad_norm": 1.0338000059127808,
      "learning_rate": 1e-05,
      "loss": 0.4977,
      "mean_token_accuracy": 0.8444024324417114,
      "num_tokens": 268217173.0,
      "step": 1683
    },
    {
      "epoch": 0.8565615462868769,
      "grad_norm": 1.0170055627822876,
      "learning_rate": 1e-05,
      "loss": 0.5311,
      "mean_token_accuracy": 0.8358046412467957,
      "num_tokens": 268382188.0,
      "step": 1684
    },
    {
      "epoch": 0.8570701932858596,
      "grad_norm": 1.0744831562042236,
      "learning_rate": 1e-05,
      "loss": 0.5146,
      "mean_token_accuracy": 0.8386661410331726,
      "num_tokens": 268543730.0,
      "step": 1685
    },
    {
      "epoch": 0.8575788402848423,
      "grad_norm": 1.038621187210083,
      "learning_rate": 1e-05,
      "loss": 0.5027,
      "mean_token_accuracy": 0.8423762321472168,
      "num_tokens": 268701052.0,
      "step": 1686
    },
    {
      "epoch": 0.858087487283825,
      "grad_norm": 1.0548534393310547,
      "learning_rate": 1e-05,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8422161936759949,
      "num_tokens": 268852461.0,
      "step": 1687
    },
    {
      "epoch": 0.8585961342828077,
      "grad_norm": 1.250922441482544,
      "learning_rate": 1e-05,
      "loss": 0.4818,
      "mean_token_accuracy": 0.8479307889938354,
      "num_tokens": 268996708.0,
      "step": 1688
    },
    {
      "epoch": 0.8591047812817905,
      "grad_norm": 1.1026182174682617,
      "learning_rate": 1e-05,
      "loss": 0.5421,
      "mean_token_accuracy": 0.8324263095855713,
      "num_tokens": 269146990.0,
      "step": 1689
    },
    {
      "epoch": 0.8596134282807731,
      "grad_norm": 1.1113507747650146,
      "learning_rate": 1e-05,
      "loss": 0.4792,
      "mean_token_accuracy": 0.8493058681488037,
      "num_tokens": 269311951.0,
      "step": 1690
    },
    {
      "epoch": 0.8601220752797558,
      "grad_norm": 1.0716196298599243,
      "learning_rate": 1e-05,
      "loss": 0.522,
      "mean_token_accuracy": 0.8371376991271973,
      "num_tokens": 269450985.0,
      "step": 1691
    },
    {
      "epoch": 0.8606307222787386,
      "grad_norm": 1.0759919881820679,
      "learning_rate": 1e-05,
      "loss": 0.4785,
      "mean_token_accuracy": 0.8490256071090698,
      "num_tokens": 269598930.0,
      "step": 1692
    },
    {
      "epoch": 0.8611393692777213,
      "grad_norm": 1.248870849609375,
      "learning_rate": 1e-05,
      "loss": 0.5399,
      "mean_token_accuracy": 0.8343132734298706,
      "num_tokens": 269760034.0,
      "step": 1693
    },
    {
      "epoch": 0.861648016276704,
      "grad_norm": 1.086024522781372,
      "learning_rate": 1e-05,
      "loss": 0.5293,
      "mean_token_accuracy": 0.8356090188026428,
      "num_tokens": 269921692.0,
      "step": 1694
    },
    {
      "epoch": 0.8621566632756866,
      "grad_norm": 1.1465421915054321,
      "learning_rate": 1e-05,
      "loss": 0.4961,
      "mean_token_accuracy": 0.8453572392463684,
      "num_tokens": 270073024.0,
      "step": 1695
    },
    {
      "epoch": 0.8626653102746694,
      "grad_norm": 1.1328400373458862,
      "learning_rate": 1e-05,
      "loss": 0.5391,
      "mean_token_accuracy": 0.8332090377807617,
      "num_tokens": 270232620.0,
      "step": 1696
    },
    {
      "epoch": 0.8631739572736521,
      "grad_norm": 1.1193041801452637,
      "learning_rate": 1e-05,
      "loss": 0.5131,
      "mean_token_accuracy": 0.8391777276992798,
      "num_tokens": 270385773.0,
      "step": 1697
    },
    {
      "epoch": 0.8636826042726348,
      "grad_norm": 1.0104612112045288,
      "learning_rate": 1e-05,
      "loss": 0.5078,
      "mean_token_accuracy": 0.8431419134140015,
      "num_tokens": 270560619.0,
      "step": 1698
    },
    {
      "epoch": 0.8641912512716174,
      "grad_norm": 1.0650326013565063,
      "learning_rate": 1e-05,
      "loss": 0.5316,
      "mean_token_accuracy": 0.8359506130218506,
      "num_tokens": 270719944.0,
      "step": 1699
    },
    {
      "epoch": 0.8646998982706002,
      "grad_norm": 1.106897234916687,
      "learning_rate": 1e-05,
      "loss": 0.5024,
      "mean_token_accuracy": 0.8426070213317871,
      "num_tokens": 270874728.0,
      "step": 1700
    },
    {
      "epoch": 0.8652085452695829,
      "grad_norm": 0.972864031791687,
      "learning_rate": 1e-05,
      "loss": 0.514,
      "mean_token_accuracy": 0.8413708209991455,
      "num_tokens": 271036953.0,
      "step": 1701
    },
    {
      "epoch": 0.8657171922685656,
      "grad_norm": 1.0611705780029297,
      "learning_rate": 1e-05,
      "loss": 0.5096,
      "mean_token_accuracy": 0.8410873413085938,
      "num_tokens": 271202782.0,
      "step": 1702
    },
    {
      "epoch": 0.8662258392675484,
      "grad_norm": 1.021085262298584,
      "learning_rate": 1e-05,
      "loss": 0.5245,
      "mean_token_accuracy": 0.838525652885437,
      "num_tokens": 271367311.0,
      "step": 1703
    },
    {
      "epoch": 0.866734486266531,
      "grad_norm": 1.002583622932434,
      "learning_rate": 1e-05,
      "loss": 0.5201,
      "mean_token_accuracy": 0.838585615158081,
      "num_tokens": 271529123.0,
      "step": 1704
    },
    {
      "epoch": 0.8672431332655137,
      "grad_norm": 1.1310274600982666,
      "learning_rate": 1e-05,
      "loss": 0.5038,
      "mean_token_accuracy": 0.8432391881942749,
      "num_tokens": 271673733.0,
      "step": 1705
    },
    {
      "epoch": 0.8677517802644964,
      "grad_norm": 1.1639235019683838,
      "learning_rate": 1e-05,
      "loss": 0.4958,
      "mean_token_accuracy": 0.8447306156158447,
      "num_tokens": 271828281.0,
      "step": 1706
    },
    {
      "epoch": 0.8682604272634792,
      "grad_norm": 1.1261109113693237,
      "learning_rate": 1e-05,
      "loss": 0.5364,
      "mean_token_accuracy": 0.8337569236755371,
      "num_tokens": 271997170.0,
      "step": 1707
    },
    {
      "epoch": 0.8687690742624619,
      "grad_norm": 0.9615015387535095,
      "learning_rate": 1e-05,
      "loss": 0.4922,
      "mean_token_accuracy": 0.8460158109664917,
      "num_tokens": 272152291.0,
      "step": 1708
    },
    {
      "epoch": 0.8692777212614445,
      "grad_norm": 1.131570816040039,
      "learning_rate": 1e-05,
      "loss": 0.5436,
      "mean_token_accuracy": 0.8301962018013,
      "num_tokens": 272307347.0,
      "step": 1709
    },
    {
      "epoch": 0.8697863682604272,
      "grad_norm": 1.0198192596435547,
      "learning_rate": 1e-05,
      "loss": 0.5348,
      "mean_token_accuracy": 0.8341028690338135,
      "num_tokens": 272469049.0,
      "step": 1710
    },
    {
      "epoch": 0.87029501525941,
      "grad_norm": 1.0531460046768188,
      "learning_rate": 1e-05,
      "loss": 0.533,
      "mean_token_accuracy": 0.8349130153656006,
      "num_tokens": 272612517.0,
      "step": 1711
    },
    {
      "epoch": 0.8708036622583927,
      "grad_norm": 1.121743083000183,
      "learning_rate": 1e-05,
      "loss": 0.5492,
      "mean_token_accuracy": 0.8294485211372375,
      "num_tokens": 272768408.0,
      "step": 1712
    },
    {
      "epoch": 0.8713123092573754,
      "grad_norm": 0.9979460835456848,
      "learning_rate": 1e-05,
      "loss": 0.5217,
      "mean_token_accuracy": 0.8372259140014648,
      "num_tokens": 272919347.0,
      "step": 1713
    },
    {
      "epoch": 0.8718209562563581,
      "grad_norm": 1.1718113422393799,
      "learning_rate": 1e-05,
      "loss": 0.5065,
      "mean_token_accuracy": 0.841956377029419,
      "num_tokens": 273089732.0,
      "step": 1714
    },
    {
      "epoch": 0.8723296032553408,
      "grad_norm": 1.0107232332229614,
      "learning_rate": 1e-05,
      "loss": 0.4993,
      "mean_token_accuracy": 0.8463701009750366,
      "num_tokens": 273242780.0,
      "step": 1715
    },
    {
      "epoch": 0.8728382502543235,
      "grad_norm": 0.9980534315109253,
      "learning_rate": 1e-05,
      "loss": 0.5054,
      "mean_token_accuracy": 0.8422236442565918,
      "num_tokens": 273407433.0,
      "step": 1716
    },
    {
      "epoch": 0.8733468972533062,
      "grad_norm": 2.6608331203460693,
      "learning_rate": 1e-05,
      "loss": 0.5206,
      "mean_token_accuracy": 0.8389767408370972,
      "num_tokens": 273570166.0,
      "step": 1717
    },
    {
      "epoch": 0.873855544252289,
      "grad_norm": 1.1380822658538818,
      "learning_rate": 1e-05,
      "loss": 0.5361,
      "mean_token_accuracy": 0.8322861194610596,
      "num_tokens": 273726324.0,
      "step": 1718
    },
    {
      "epoch": 0.8743641912512716,
      "grad_norm": 1.0491878986358643,
      "learning_rate": 1e-05,
      "loss": 0.5264,
      "mean_token_accuracy": 0.8350399732589722,
      "num_tokens": 273890563.0,
      "step": 1719
    },
    {
      "epoch": 0.8748728382502543,
      "grad_norm": 1.0042964220046997,
      "learning_rate": 1e-05,
      "loss": 0.484,
      "mean_token_accuracy": 0.8472247123718262,
      "num_tokens": 274048187.0,
      "step": 1720
    },
    {
      "epoch": 0.875381485249237,
      "grad_norm": 1.0009158849716187,
      "learning_rate": 1e-05,
      "loss": 0.4828,
      "mean_token_accuracy": 0.8478761911392212,
      "num_tokens": 274221855.0,
      "step": 1721
    },
    {
      "epoch": 0.8758901322482198,
      "grad_norm": 0.9791696667671204,
      "learning_rate": 1e-05,
      "loss": 0.5279,
      "mean_token_accuracy": 0.8349093198776245,
      "num_tokens": 274377833.0,
      "step": 1722
    },
    {
      "epoch": 0.8763987792472024,
      "grad_norm": 0.9939912557601929,
      "learning_rate": 1e-05,
      "loss": 0.5423,
      "mean_token_accuracy": 0.8309345245361328,
      "num_tokens": 274541936.0,
      "step": 1723
    },
    {
      "epoch": 0.8769074262461851,
      "grad_norm": 1.0004394054412842,
      "learning_rate": 1e-05,
      "loss": 0.5375,
      "mean_token_accuracy": 0.8317330479621887,
      "num_tokens": 274705006.0,
      "step": 1724
    },
    {
      "epoch": 0.8774160732451679,
      "grad_norm": 0.9913457632064819,
      "learning_rate": 1e-05,
      "loss": 0.5086,
      "mean_token_accuracy": 0.8392433524131775,
      "num_tokens": 274854136.0,
      "step": 1725
    },
    {
      "epoch": 0.8779247202441506,
      "grad_norm": 1.064914345741272,
      "learning_rate": 1e-05,
      "loss": 0.4971,
      "mean_token_accuracy": 0.8464230298995972,
      "num_tokens": 275014059.0,
      "step": 1726
    },
    {
      "epoch": 0.8784333672431333,
      "grad_norm": 1.2501319646835327,
      "learning_rate": 1e-05,
      "loss": 0.5594,
      "mean_token_accuracy": 0.8279653787612915,
      "num_tokens": 275187361.0,
      "step": 1727
    },
    {
      "epoch": 0.8789420142421159,
      "grad_norm": 1.013777732849121,
      "learning_rate": 1e-05,
      "loss": 0.5058,
      "mean_token_accuracy": 0.8425623178482056,
      "num_tokens": 275338869.0,
      "step": 1728
    },
    {
      "epoch": 0.8794506612410987,
      "grad_norm": 0.9364235997200012,
      "learning_rate": 1e-05,
      "loss": 0.511,
      "mean_token_accuracy": 0.8407465219497681,
      "num_tokens": 275496774.0,
      "step": 1729
    },
    {
      "epoch": 0.8799593082400814,
      "grad_norm": 1.1220306158065796,
      "learning_rate": 1e-05,
      "loss": 0.5401,
      "mean_token_accuracy": 0.8303672075271606,
      "num_tokens": 275653401.0,
      "step": 1730
    },
    {
      "epoch": 0.8804679552390641,
      "grad_norm": 1.0587559938430786,
      "learning_rate": 1e-05,
      "loss": 0.4986,
      "mean_token_accuracy": 0.8438327312469482,
      "num_tokens": 275806070.0,
      "step": 1731
    },
    {
      "epoch": 0.8809766022380467,
      "grad_norm": 1.1446789503097534,
      "learning_rate": 1e-05,
      "loss": 0.5118,
      "mean_token_accuracy": 0.8396477103233337,
      "num_tokens": 275962373.0,
      "step": 1732
    },
    {
      "epoch": 0.8814852492370295,
      "grad_norm": 1.0831222534179688,
      "learning_rate": 1e-05,
      "loss": 0.5185,
      "mean_token_accuracy": 0.8397834300994873,
      "num_tokens": 276129118.0,
      "step": 1733
    },
    {
      "epoch": 0.8819938962360122,
      "grad_norm": 1.0126874446868896,
      "learning_rate": 1e-05,
      "loss": 0.5259,
      "mean_token_accuracy": 0.8371516466140747,
      "num_tokens": 276287064.0,
      "step": 1734
    },
    {
      "epoch": 0.8825025432349949,
      "grad_norm": 1.0329630374908447,
      "learning_rate": 1e-05,
      "loss": 0.5033,
      "mean_token_accuracy": 0.8428022861480713,
      "num_tokens": 276457678.0,
      "step": 1735
    },
    {
      "epoch": 0.8830111902339777,
      "grad_norm": 1.0754644870758057,
      "learning_rate": 1e-05,
      "loss": 0.5355,
      "mean_token_accuracy": 0.834283709526062,
      "num_tokens": 276610916.0,
      "step": 1736
    },
    {
      "epoch": 0.8835198372329603,
      "grad_norm": 0.9800335764884949,
      "learning_rate": 1e-05,
      "loss": 0.4946,
      "mean_token_accuracy": 0.8463819026947021,
      "num_tokens": 276764925.0,
      "step": 1737
    },
    {
      "epoch": 0.884028484231943,
      "grad_norm": 1.0709609985351562,
      "learning_rate": 1e-05,
      "loss": 0.574,
      "mean_token_accuracy": 0.8228265047073364,
      "num_tokens": 276928368.0,
      "step": 1738
    },
    {
      "epoch": 0.8845371312309257,
      "grad_norm": 1.0306966304779053,
      "learning_rate": 1e-05,
      "loss": 0.4964,
      "mean_token_accuracy": 0.8455370664596558,
      "num_tokens": 277094718.0,
      "step": 1739
    },
    {
      "epoch": 0.8850457782299085,
      "grad_norm": 1.026525855064392,
      "learning_rate": 1e-05,
      "loss": 0.5012,
      "mean_token_accuracy": 0.8419168591499329,
      "num_tokens": 277246859.0,
      "step": 1740
    },
    {
      "epoch": 0.8855544252288912,
      "grad_norm": 0.9500491619110107,
      "learning_rate": 1e-05,
      "loss": 0.528,
      "mean_token_accuracy": 0.83522629737854,
      "num_tokens": 277414068.0,
      "step": 1741
    },
    {
      "epoch": 0.8860630722278738,
      "grad_norm": 1.0432976484298706,
      "learning_rate": 1e-05,
      "loss": 0.5106,
      "mean_token_accuracy": 0.840704083442688,
      "num_tokens": 277573295.0,
      "step": 1742
    },
    {
      "epoch": 0.8865717192268565,
      "grad_norm": 1.0290638208389282,
      "learning_rate": 1e-05,
      "loss": 0.5246,
      "mean_token_accuracy": 0.8381516933441162,
      "num_tokens": 277726913.0,
      "step": 1743
    },
    {
      "epoch": 0.8870803662258393,
      "grad_norm": 0.9862432479858398,
      "learning_rate": 1e-05,
      "loss": 0.4924,
      "mean_token_accuracy": 0.8451281785964966,
      "num_tokens": 277877252.0,
      "step": 1744
    },
    {
      "epoch": 0.887589013224822,
      "grad_norm": 1.0773284435272217,
      "learning_rate": 1e-05,
      "loss": 0.5378,
      "mean_token_accuracy": 0.8321132659912109,
      "num_tokens": 278034976.0,
      "step": 1745
    },
    {
      "epoch": 0.8880976602238047,
      "grad_norm": 0.9903329014778137,
      "learning_rate": 1e-05,
      "loss": 0.4778,
      "mean_token_accuracy": 0.8489717841148376,
      "num_tokens": 278194339.0,
      "step": 1746
    },
    {
      "epoch": 0.8886063072227874,
      "grad_norm": 0.9978926777839661,
      "learning_rate": 1e-05,
      "loss": 0.5122,
      "mean_token_accuracy": 0.83891361951828,
      "num_tokens": 278352150.0,
      "step": 1747
    },
    {
      "epoch": 0.8891149542217701,
      "grad_norm": 1.0282410383224487,
      "learning_rate": 1e-05,
      "loss": 0.5348,
      "mean_token_accuracy": 0.832777738571167,
      "num_tokens": 278516575.0,
      "step": 1748
    },
    {
      "epoch": 0.8896236012207528,
      "grad_norm": 0.9718902111053467,
      "learning_rate": 1e-05,
      "loss": 0.5342,
      "mean_token_accuracy": 0.8338038921356201,
      "num_tokens": 278687157.0,
      "step": 1749
    },
    {
      "epoch": 0.8901322482197355,
      "grad_norm": 1.0574067831039429,
      "learning_rate": 1e-05,
      "loss": 0.5331,
      "mean_token_accuracy": 0.8335607051849365,
      "num_tokens": 278844290.0,
      "step": 1750
    },
    {
      "epoch": 0.8906408952187183,
      "grad_norm": 1.118514895439148,
      "learning_rate": 1e-05,
      "loss": 0.5178,
      "mean_token_accuracy": 0.8380013704299927,
      "num_tokens": 279002276.0,
      "step": 1751
    },
    {
      "epoch": 0.8911495422177009,
      "grad_norm": 1.017069935798645,
      "learning_rate": 1e-05,
      "loss": 0.5315,
      "mean_token_accuracy": 0.835065484046936,
      "num_tokens": 279161918.0,
      "step": 1752
    },
    {
      "epoch": 0.8916581892166836,
      "grad_norm": 1.1437948942184448,
      "learning_rate": 1e-05,
      "loss": 0.5303,
      "mean_token_accuracy": 0.8335355520248413,
      "num_tokens": 279321930.0,
      "step": 1753
    },
    {
      "epoch": 0.8921668362156663,
      "grad_norm": 1.0639396905899048,
      "learning_rate": 1e-05,
      "loss": 0.5353,
      "mean_token_accuracy": 0.8358509540557861,
      "num_tokens": 279481295.0,
      "step": 1754
    },
    {
      "epoch": 0.8926754832146491,
      "grad_norm": 0.9928602576255798,
      "learning_rate": 1e-05,
      "loss": 0.5236,
      "mean_token_accuracy": 0.8368161916732788,
      "num_tokens": 279633071.0,
      "step": 1755
    },
    {
      "epoch": 0.8931841302136317,
      "grad_norm": 1.0300835371017456,
      "learning_rate": 1e-05,
      "loss": 0.4902,
      "mean_token_accuracy": 0.8459980487823486,
      "num_tokens": 279800329.0,
      "step": 1756
    },
    {
      "epoch": 0.8936927772126144,
      "grad_norm": 1.0273916721343994,
      "learning_rate": 1e-05,
      "loss": 0.483,
      "mean_token_accuracy": 0.8479832410812378,
      "num_tokens": 279944707.0,
      "step": 1757
    },
    {
      "epoch": 0.8942014242115972,
      "grad_norm": 0.9553089737892151,
      "learning_rate": 1e-05,
      "loss": 0.5054,
      "mean_token_accuracy": 0.8420307636260986,
      "num_tokens": 280111037.0,
      "step": 1758
    },
    {
      "epoch": 0.8947100712105799,
      "grad_norm": 1.0497819185256958,
      "learning_rate": 1e-05,
      "loss": 0.5061,
      "mean_token_accuracy": 0.8428786993026733,
      "num_tokens": 280258749.0,
      "step": 1759
    },
    {
      "epoch": 0.8952187182095626,
      "grad_norm": 1.0025304555892944,
      "learning_rate": 1e-05,
      "loss": 0.5179,
      "mean_token_accuracy": 0.8372910618782043,
      "num_tokens": 280405852.0,
      "step": 1760
    },
    {
      "epoch": 0.8957273652085452,
      "grad_norm": 0.9646736979484558,
      "learning_rate": 1e-05,
      "loss": 0.5166,
      "mean_token_accuracy": 0.8398212194442749,
      "num_tokens": 280580217.0,
      "step": 1761
    },
    {
      "epoch": 0.896236012207528,
      "grad_norm": 1.0457273721694946,
      "learning_rate": 1e-05,
      "loss": 0.534,
      "mean_token_accuracy": 0.8361070156097412,
      "num_tokens": 280727104.0,
      "step": 1762
    },
    {
      "epoch": 0.8967446592065107,
      "grad_norm": 1.0271176099777222,
      "learning_rate": 1e-05,
      "loss": 0.4906,
      "mean_token_accuracy": 0.8460567593574524,
      "num_tokens": 280889043.0,
      "step": 1763
    },
    {
      "epoch": 0.8972533062054934,
      "grad_norm": 1.0649844408035278,
      "learning_rate": 1e-05,
      "loss": 0.5078,
      "mean_token_accuracy": 0.8414275646209717,
      "num_tokens": 281043942.0,
      "step": 1764
    },
    {
      "epoch": 0.897761953204476,
      "grad_norm": 1.0987054109573364,
      "learning_rate": 1e-05,
      "loss": 0.5214,
      "mean_token_accuracy": 0.8381832838058472,
      "num_tokens": 281216118.0,
      "step": 1765
    },
    {
      "epoch": 0.8982706002034588,
      "grad_norm": 1.0467629432678223,
      "learning_rate": 1e-05,
      "loss": 0.5184,
      "mean_token_accuracy": 0.839253842830658,
      "num_tokens": 281374849.0,
      "step": 1766
    },
    {
      "epoch": 0.8987792472024415,
      "grad_norm": 1.060920238494873,
      "learning_rate": 1e-05,
      "loss": 0.5083,
      "mean_token_accuracy": 0.8411917090415955,
      "num_tokens": 281528478.0,
      "step": 1767
    },
    {
      "epoch": 0.8992878942014242,
      "grad_norm": 1.1491779088974,
      "learning_rate": 1e-05,
      "loss": 0.5106,
      "mean_token_accuracy": 0.8416895866394043,
      "num_tokens": 281681982.0,
      "step": 1768
    },
    {
      "epoch": 0.8997965412004069,
      "grad_norm": 1.028265357017517,
      "learning_rate": 1e-05,
      "loss": 0.4975,
      "mean_token_accuracy": 0.8440154790878296,
      "num_tokens": 281846049.0,
      "step": 1769
    },
    {
      "epoch": 0.9003051881993896,
      "grad_norm": 1.1693898439407349,
      "learning_rate": 1e-05,
      "loss": 0.5579,
      "mean_token_accuracy": 0.8275507688522339,
      "num_tokens": 282019096.0,
      "step": 1770
    },
    {
      "epoch": 0.9008138351983723,
      "grad_norm": 1.0625677108764648,
      "learning_rate": 1e-05,
      "loss": 0.4945,
      "mean_token_accuracy": 0.8449739217758179,
      "num_tokens": 282196142.0,
      "step": 1771
    },
    {
      "epoch": 0.901322482197355,
      "grad_norm": 0.9961878061294556,
      "learning_rate": 1e-05,
      "loss": 0.5269,
      "mean_token_accuracy": 0.8377020359039307,
      "num_tokens": 282360553.0,
      "step": 1772
    },
    {
      "epoch": 0.9018311291963378,
      "grad_norm": 1.1472853422164917,
      "learning_rate": 1e-05,
      "loss": 0.4972,
      "mean_token_accuracy": 0.8453328609466553,
      "num_tokens": 282527151.0,
      "step": 1773
    },
    {
      "epoch": 0.9023397761953205,
      "grad_norm": 1.0794323682785034,
      "learning_rate": 1e-05,
      "loss": 0.5417,
      "mean_token_accuracy": 0.8325939178466797,
      "num_tokens": 282674995.0,
      "step": 1774
    },
    {
      "epoch": 0.9028484231943031,
      "grad_norm": 1.19264554977417,
      "learning_rate": 1e-05,
      "loss": 0.4699,
      "mean_token_accuracy": 0.8525264859199524,
      "num_tokens": 282832990.0,
      "step": 1775
    },
    {
      "epoch": 0.9033570701932858,
      "grad_norm": 1.1174076795578003,
      "learning_rate": 1e-05,
      "loss": 0.5375,
      "mean_token_accuracy": 0.8341360688209534,
      "num_tokens": 282997592.0,
      "step": 1776
    },
    {
      "epoch": 0.9038657171922686,
      "grad_norm": 1.1486574411392212,
      "learning_rate": 1e-05,
      "loss": 0.5164,
      "mean_token_accuracy": 0.8390986919403076,
      "num_tokens": 283155838.0,
      "step": 1777
    },
    {
      "epoch": 0.9043743641912513,
      "grad_norm": 0.977383017539978,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8494383096694946,
      "num_tokens": 283307527.0,
      "step": 1778
    },
    {
      "epoch": 0.904883011190234,
      "grad_norm": 1.0261812210083008,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.8334873914718628,
      "num_tokens": 283465966.0,
      "step": 1779
    },
    {
      "epoch": 0.9053916581892166,
      "grad_norm": 1.0674028396606445,
      "learning_rate": 1e-05,
      "loss": 0.5333,
      "mean_token_accuracy": 0.8340882062911987,
      "num_tokens": 283620377.0,
      "step": 1780
    },
    {
      "epoch": 0.9059003051881994,
      "grad_norm": 0.9716681241989136,
      "learning_rate": 1e-05,
      "loss": 0.4934,
      "mean_token_accuracy": 0.8443837761878967,
      "num_tokens": 283777695.0,
      "step": 1781
    },
    {
      "epoch": 0.9064089521871821,
      "grad_norm": 1.0851030349731445,
      "learning_rate": 1e-05,
      "loss": 0.5108,
      "mean_token_accuracy": 0.8417867422103882,
      "num_tokens": 283931044.0,
      "step": 1782
    },
    {
      "epoch": 0.9069175991861648,
      "grad_norm": 1.069290041923523,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.8441795110702515,
      "num_tokens": 284089124.0,
      "step": 1783
    },
    {
      "epoch": 0.9074262461851476,
      "grad_norm": 1.009878158569336,
      "learning_rate": 1e-05,
      "loss": 0.5346,
      "mean_token_accuracy": 0.8334824442863464,
      "num_tokens": 284248921.0,
      "step": 1784
    },
    {
      "epoch": 0.9079348931841302,
      "grad_norm": 1.088663935661316,
      "learning_rate": 1e-05,
      "loss": 0.501,
      "mean_token_accuracy": 0.8420149683952332,
      "num_tokens": 284406295.0,
      "step": 1785
    },
    {
      "epoch": 0.9084435401831129,
      "grad_norm": 1.072481393814087,
      "learning_rate": 1e-05,
      "loss": 0.4704,
      "mean_token_accuracy": 0.8499975800514221,
      "num_tokens": 284570062.0,
      "step": 1786
    },
    {
      "epoch": 0.9089521871820956,
      "grad_norm": 1.067689299583435,
      "learning_rate": 1e-05,
      "loss": 0.5181,
      "mean_token_accuracy": 0.8386754989624023,
      "num_tokens": 284731919.0,
      "step": 1787
    },
    {
      "epoch": 0.9094608341810784,
      "grad_norm": 1.002181053161621,
      "learning_rate": 1e-05,
      "loss": 0.5293,
      "mean_token_accuracy": 0.8357716798782349,
      "num_tokens": 284892788.0,
      "step": 1788
    },
    {
      "epoch": 0.909969481180061,
      "grad_norm": 1.1112078428268433,
      "learning_rate": 1e-05,
      "loss": 0.5359,
      "mean_token_accuracy": 0.8343470096588135,
      "num_tokens": 285053351.0,
      "step": 1789
    },
    {
      "epoch": 0.9104781281790437,
      "grad_norm": 1.0172306299209595,
      "learning_rate": 1e-05,
      "loss": 0.5216,
      "mean_token_accuracy": 0.8377166986465454,
      "num_tokens": 285203700.0,
      "step": 1790
    },
    {
      "epoch": 0.9109867751780264,
      "grad_norm": 1.0339511632919312,
      "learning_rate": 1e-05,
      "loss": 0.5193,
      "mean_token_accuracy": 0.8383980989456177,
      "num_tokens": 285375982.0,
      "step": 1791
    },
    {
      "epoch": 0.9114954221770092,
      "grad_norm": 0.9519206285476685,
      "learning_rate": 1e-05,
      "loss": 0.5129,
      "mean_token_accuracy": 0.8404895663261414,
      "num_tokens": 285550546.0,
      "step": 1792
    },
    {
      "epoch": 0.9120040691759919,
      "grad_norm": 1.0351858139038086,
      "learning_rate": 1e-05,
      "loss": 0.523,
      "mean_token_accuracy": 0.8380796909332275,
      "num_tokens": 285699042.0,
      "step": 1793
    },
    {
      "epoch": 0.9125127161749745,
      "grad_norm": 1.031059980392456,
      "learning_rate": 1e-05,
      "loss": 0.5172,
      "mean_token_accuracy": 0.8366599678993225,
      "num_tokens": 285845604.0,
      "step": 1794
    },
    {
      "epoch": 0.9130213631739573,
      "grad_norm": 1.0507643222808838,
      "learning_rate": 1e-05,
      "loss": 0.5195,
      "mean_token_accuracy": 0.838219165802002,
      "num_tokens": 285989582.0,
      "step": 1795
    },
    {
      "epoch": 0.91353001017294,
      "grad_norm": 1.0662038326263428,
      "learning_rate": 1e-05,
      "loss": 0.5288,
      "mean_token_accuracy": 0.8362869024276733,
      "num_tokens": 286144853.0,
      "step": 1796
    },
    {
      "epoch": 0.9140386571719227,
      "grad_norm": 1.100429892539978,
      "learning_rate": 1e-05,
      "loss": 0.5145,
      "mean_token_accuracy": 0.838462233543396,
      "num_tokens": 286306136.0,
      "step": 1797
    },
    {
      "epoch": 0.9145473041709054,
      "grad_norm": 1.0728235244750977,
      "learning_rate": 1e-05,
      "loss": 0.5172,
      "mean_token_accuracy": 0.8359116911888123,
      "num_tokens": 286458777.0,
      "step": 1798
    },
    {
      "epoch": 0.9150559511698881,
      "grad_norm": 0.9926060438156128,
      "learning_rate": 1e-05,
      "loss": 0.5225,
      "mean_token_accuracy": 0.8370160460472107,
      "num_tokens": 286618373.0,
      "step": 1799
    },
    {
      "epoch": 0.9155645981688708,
      "grad_norm": 1.0963010787963867,
      "learning_rate": 1e-05,
      "loss": 0.5096,
      "mean_token_accuracy": 0.8409956097602844,
      "num_tokens": 286774374.0,
      "step": 1800
    },
    {
      "epoch": 0.9160732451678535,
      "grad_norm": 1.0256036520004272,
      "learning_rate": 1e-05,
      "loss": 0.4933,
      "mean_token_accuracy": 0.8443099856376648,
      "num_tokens": 286926506.0,
      "step": 1801
    },
    {
      "epoch": 0.9165818921668362,
      "grad_norm": 1.0650215148925781,
      "learning_rate": 1e-05,
      "loss": 0.5427,
      "mean_token_accuracy": 0.8311566114425659,
      "num_tokens": 287085779.0,
      "step": 1802
    },
    {
      "epoch": 0.917090539165819,
      "grad_norm": 1.0663613080978394,
      "learning_rate": 1e-05,
      "loss": 0.5112,
      "mean_token_accuracy": 0.8394999504089355,
      "num_tokens": 287253165.0,
      "step": 1803
    },
    {
      "epoch": 0.9175991861648016,
      "grad_norm": 1.076520562171936,
      "learning_rate": 1e-05,
      "loss": 0.5377,
      "mean_token_accuracy": 0.832918643951416,
      "num_tokens": 287405700.0,
      "step": 1804
    },
    {
      "epoch": 0.9181078331637843,
      "grad_norm": 1.0021073818206787,
      "learning_rate": 1e-05,
      "loss": 0.477,
      "mean_token_accuracy": 0.849860668182373,
      "num_tokens": 287551897.0,
      "step": 1805
    },
    {
      "epoch": 0.9186164801627671,
      "grad_norm": 1.0759339332580566,
      "learning_rate": 1e-05,
      "loss": 0.5147,
      "mean_token_accuracy": 0.8378969430923462,
      "num_tokens": 287716716.0,
      "step": 1806
    },
    {
      "epoch": 0.9191251271617498,
      "grad_norm": 1.0549094676971436,
      "learning_rate": 1e-05,
      "loss": 0.5497,
      "mean_token_accuracy": 0.8297152519226074,
      "num_tokens": 287885440.0,
      "step": 1807
    },
    {
      "epoch": 0.9196337741607324,
      "grad_norm": 1.0533418655395508,
      "learning_rate": 1e-05,
      "loss": 0.4901,
      "mean_token_accuracy": 0.8451825380325317,
      "num_tokens": 288040043.0,
      "step": 1808
    },
    {
      "epoch": 0.9201424211597151,
      "grad_norm": 1.0616881847381592,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.8455978631973267,
      "num_tokens": 288190884.0,
      "step": 1809
    },
    {
      "epoch": 0.9206510681586979,
      "grad_norm": 1.0778988599777222,
      "learning_rate": 1e-05,
      "loss": 0.5411,
      "mean_token_accuracy": 0.8317572474479675,
      "num_tokens": 288352383.0,
      "step": 1810
    },
    {
      "epoch": 0.9211597151576806,
      "grad_norm": 0.968621015548706,
      "learning_rate": 1e-05,
      "loss": 0.475,
      "mean_token_accuracy": 0.8496113419532776,
      "num_tokens": 288513226.0,
      "step": 1811
    },
    {
      "epoch": 0.9216683621566633,
      "grad_norm": 1.3760226964950562,
      "learning_rate": 1e-05,
      "loss": 0.5276,
      "mean_token_accuracy": 0.8349040746688843,
      "num_tokens": 288681122.0,
      "step": 1812
    },
    {
      "epoch": 0.9221770091556459,
      "grad_norm": 1.0464659929275513,
      "learning_rate": 1e-05,
      "loss": 0.5393,
      "mean_token_accuracy": 0.8343937397003174,
      "num_tokens": 288840955.0,
      "step": 1813
    },
    {
      "epoch": 0.9226856561546287,
      "grad_norm": 0.9643142223358154,
      "learning_rate": 1e-05,
      "loss": 0.5309,
      "mean_token_accuracy": 0.8338766098022461,
      "num_tokens": 289007588.0,
      "step": 1814
    },
    {
      "epoch": 0.9231943031536114,
      "grad_norm": 1.0363613367080688,
      "learning_rate": 1e-05,
      "loss": 0.5117,
      "mean_token_accuracy": 0.840315043926239,
      "num_tokens": 289171963.0,
      "step": 1815
    },
    {
      "epoch": 0.9237029501525941,
      "grad_norm": 0.975771963596344,
      "learning_rate": 1e-05,
      "loss": 0.5043,
      "mean_token_accuracy": 0.8419756889343262,
      "num_tokens": 289338174.0,
      "step": 1816
    },
    {
      "epoch": 0.9242115971515769,
      "grad_norm": 0.9576640725135803,
      "learning_rate": 1e-05,
      "loss": 0.504,
      "mean_token_accuracy": 0.8424282670021057,
      "num_tokens": 289498369.0,
      "step": 1817
    },
    {
      "epoch": 0.9247202441505595,
      "grad_norm": 1.0540850162506104,
      "learning_rate": 1e-05,
      "loss": 0.5032,
      "mean_token_accuracy": 0.8419172763824463,
      "num_tokens": 289657315.0,
      "step": 1818
    },
    {
      "epoch": 0.9252288911495422,
      "grad_norm": 1.0213912725448608,
      "learning_rate": 1e-05,
      "loss": 0.4994,
      "mean_token_accuracy": 0.8438811302185059,
      "num_tokens": 289824552.0,
      "step": 1819
    },
    {
      "epoch": 0.9257375381485249,
      "grad_norm": 0.9739988446235657,
      "learning_rate": 1e-05,
      "loss": 0.5164,
      "mean_token_accuracy": 0.8370422124862671,
      "num_tokens": 289982984.0,
      "step": 1820
    },
    {
      "epoch": 0.9262461851475077,
      "grad_norm": 0.9719370007514954,
      "learning_rate": 1e-05,
      "loss": 0.5018,
      "mean_token_accuracy": 0.8428477644920349,
      "num_tokens": 290127147.0,
      "step": 1821
    },
    {
      "epoch": 0.9267548321464903,
      "grad_norm": 0.9787293672561646,
      "learning_rate": 1e-05,
      "loss": 0.5066,
      "mean_token_accuracy": 0.8425807952880859,
      "num_tokens": 290298649.0,
      "step": 1822
    },
    {
      "epoch": 0.927263479145473,
      "grad_norm": 1.0080795288085938,
      "learning_rate": 1e-05,
      "loss": 0.5337,
      "mean_token_accuracy": 0.8355758190155029,
      "num_tokens": 290460355.0,
      "step": 1823
    },
    {
      "epoch": 0.9277721261444557,
      "grad_norm": 0.9061076045036316,
      "learning_rate": 1e-05,
      "loss": 0.4846,
      "mean_token_accuracy": 0.8472140431404114,
      "num_tokens": 290617255.0,
      "step": 1824
    },
    {
      "epoch": 0.9282807731434385,
      "grad_norm": 1.1104837656021118,
      "learning_rate": 1e-05,
      "loss": 0.5201,
      "mean_token_accuracy": 0.8363896608352661,
      "num_tokens": 290773839.0,
      "step": 1825
    },
    {
      "epoch": 0.9287894201424212,
      "grad_norm": 0.9883270263671875,
      "learning_rate": 1e-05,
      "loss": 0.49,
      "mean_token_accuracy": 0.846264123916626,
      "num_tokens": 290937706.0,
      "step": 1826
    },
    {
      "epoch": 0.9292980671414038,
      "grad_norm": 0.9738947153091431,
      "learning_rate": 1e-05,
      "loss": 0.508,
      "mean_token_accuracy": 0.8415963649749756,
      "num_tokens": 291095257.0,
      "step": 1827
    },
    {
      "epoch": 0.9298067141403866,
      "grad_norm": 1.1056708097457886,
      "learning_rate": 1e-05,
      "loss": 0.5293,
      "mean_token_accuracy": 0.8362318277359009,
      "num_tokens": 291248241.0,
      "step": 1828
    },
    {
      "epoch": 0.9303153611393693,
      "grad_norm": 1.0047028064727783,
      "learning_rate": 1e-05,
      "loss": 0.5443,
      "mean_token_accuracy": 0.8309669494628906,
      "num_tokens": 291408732.0,
      "step": 1829
    },
    {
      "epoch": 0.930824008138352,
      "grad_norm": 0.943976104259491,
      "learning_rate": 1e-05,
      "loss": 0.5021,
      "mean_token_accuracy": 0.8414051532745361,
      "num_tokens": 291586275.0,
      "step": 1830
    },
    {
      "epoch": 0.9313326551373347,
      "grad_norm": 1.0755298137664795,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8507115840911865,
      "num_tokens": 291748574.0,
      "step": 1831
    },
    {
      "epoch": 0.9318413021363174,
      "grad_norm": 1.173866868019104,
      "learning_rate": 1e-05,
      "loss": 0.5208,
      "mean_token_accuracy": 0.8381999731063843,
      "num_tokens": 291904561.0,
      "step": 1832
    },
    {
      "epoch": 0.9323499491353001,
      "grad_norm": 3.691843271255493,
      "learning_rate": 1e-05,
      "loss": 0.5084,
      "mean_token_accuracy": 0.8408045768737793,
      "num_tokens": 292063897.0,
      "step": 1833
    },
    {
      "epoch": 0.9328585961342828,
      "grad_norm": 1.1650604009628296,
      "learning_rate": 1e-05,
      "loss": 0.5194,
      "mean_token_accuracy": 0.8382998704910278,
      "num_tokens": 292214056.0,
      "step": 1834
    },
    {
      "epoch": 0.9333672431332655,
      "grad_norm": 1.0645188093185425,
      "learning_rate": 1e-05,
      "loss": 0.5278,
      "mean_token_accuracy": 0.8340986371040344,
      "num_tokens": 292349124.0,
      "step": 1835
    },
    {
      "epoch": 0.9338758901322483,
      "grad_norm": 1.0586034059524536,
      "learning_rate": 1e-05,
      "loss": 0.5013,
      "mean_token_accuracy": 0.8424099087715149,
      "num_tokens": 292505385.0,
      "step": 1836
    },
    {
      "epoch": 0.9343845371312309,
      "grad_norm": 1.0245041847229004,
      "learning_rate": 1e-05,
      "loss": 0.5198,
      "mean_token_accuracy": 0.8401522636413574,
      "num_tokens": 292665526.0,
      "step": 1837
    },
    {
      "epoch": 0.9348931841302136,
      "grad_norm": 0.9853573441505432,
      "learning_rate": 1e-05,
      "loss": 0.5365,
      "mean_token_accuracy": 0.8344807624816895,
      "num_tokens": 292829795.0,
      "step": 1838
    },
    {
      "epoch": 0.9354018311291964,
      "grad_norm": 1.0970759391784668,
      "learning_rate": 1e-05,
      "loss": 0.5013,
      "mean_token_accuracy": 0.8453795313835144,
      "num_tokens": 292987496.0,
      "step": 1839
    },
    {
      "epoch": 0.9359104781281791,
      "grad_norm": 0.9906995892524719,
      "learning_rate": 1e-05,
      "loss": 0.4999,
      "mean_token_accuracy": 0.842957079410553,
      "num_tokens": 293144553.0,
      "step": 1840
    },
    {
      "epoch": 0.9364191251271617,
      "grad_norm": 1.069620966911316,
      "learning_rate": 1e-05,
      "loss": 0.5448,
      "mean_token_accuracy": 0.8327444791793823,
      "num_tokens": 293302965.0,
      "step": 1841
    },
    {
      "epoch": 0.9369277721261444,
      "grad_norm": 0.9693588614463806,
      "learning_rate": 1e-05,
      "loss": 0.4966,
      "mean_token_accuracy": 0.8441400527954102,
      "num_tokens": 293463953.0,
      "step": 1842
    },
    {
      "epoch": 0.9374364191251272,
      "grad_norm": 1.0056179761886597,
      "learning_rate": 1e-05,
      "loss": 0.5025,
      "mean_token_accuracy": 0.842492401599884,
      "num_tokens": 293629039.0,
      "step": 1843
    },
    {
      "epoch": 0.9379450661241099,
      "grad_norm": 0.977562665939331,
      "learning_rate": 1e-05,
      "loss": 0.4593,
      "mean_token_accuracy": 0.8544231057167053,
      "num_tokens": 293773864.0,
      "step": 1844
    },
    {
      "epoch": 0.9384537131230926,
      "grad_norm": 1.0503692626953125,
      "learning_rate": 1e-05,
      "loss": 0.5065,
      "mean_token_accuracy": 0.8430094718933105,
      "num_tokens": 293937000.0,
      "step": 1845
    },
    {
      "epoch": 0.9389623601220752,
      "grad_norm": 0.9278028011322021,
      "learning_rate": 1e-05,
      "loss": 0.4832,
      "mean_token_accuracy": 0.8477224111557007,
      "num_tokens": 294103688.0,
      "step": 1846
    },
    {
      "epoch": 0.939471007121058,
      "grad_norm": 0.9603849053382874,
      "learning_rate": 1e-05,
      "loss": 0.4943,
      "mean_token_accuracy": 0.8472484350204468,
      "num_tokens": 294266886.0,
      "step": 1847
    },
    {
      "epoch": 0.9399796541200407,
      "grad_norm": 0.9893496036529541,
      "learning_rate": 1e-05,
      "loss": 0.4778,
      "mean_token_accuracy": 0.8487132787704468,
      "num_tokens": 294425536.0,
      "step": 1848
    },
    {
      "epoch": 0.9404883011190234,
      "grad_norm": 0.9785003662109375,
      "learning_rate": 1e-05,
      "loss": 0.5227,
      "mean_token_accuracy": 0.8382445573806763,
      "num_tokens": 294590856.0,
      "step": 1849
    },
    {
      "epoch": 0.940996948118006,
      "grad_norm": 1.0731697082519531,
      "learning_rate": 1e-05,
      "loss": 0.5099,
      "mean_token_accuracy": 0.841046929359436,
      "num_tokens": 294743889.0,
      "step": 1850
    },
    {
      "epoch": 0.9415055951169888,
      "grad_norm": 1.0036062002182007,
      "learning_rate": 1e-05,
      "loss": 0.5033,
      "mean_token_accuracy": 0.8415602445602417,
      "num_tokens": 294899666.0,
      "step": 1851
    },
    {
      "epoch": 0.9420142421159715,
      "grad_norm": 0.9595556855201721,
      "learning_rate": 1e-05,
      "loss": 0.5585,
      "mean_token_accuracy": 0.8298701047897339,
      "num_tokens": 295078964.0,
      "step": 1852
    },
    {
      "epoch": 0.9425228891149542,
      "grad_norm": 1.0190085172653198,
      "learning_rate": 1e-05,
      "loss": 0.4816,
      "mean_token_accuracy": 0.8475434184074402,
      "num_tokens": 295221135.0,
      "step": 1853
    },
    {
      "epoch": 0.943031536113937,
      "grad_norm": 1.0703707933425903,
      "learning_rate": 1e-05,
      "loss": 0.4964,
      "mean_token_accuracy": 0.8452032804489136,
      "num_tokens": 295382594.0,
      "step": 1854
    },
    {
      "epoch": 0.9435401831129197,
      "grad_norm": 1.0177465677261353,
      "learning_rate": 1e-05,
      "loss": 0.5324,
      "mean_token_accuracy": 0.8352153301239014,
      "num_tokens": 295541772.0,
      "step": 1855
    },
    {
      "epoch": 0.9440488301119023,
      "grad_norm": 1.011955738067627,
      "learning_rate": 1e-05,
      "loss": 0.4917,
      "mean_token_accuracy": 0.8448024988174438,
      "num_tokens": 295694723.0,
      "step": 1856
    },
    {
      "epoch": 0.944557477110885,
      "grad_norm": 0.9742183685302734,
      "learning_rate": 1e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8474919199943542,
      "num_tokens": 295848132.0,
      "step": 1857
    },
    {
      "epoch": 0.9450661241098678,
      "grad_norm": 1.0941163301467896,
      "learning_rate": 1e-05,
      "loss": 0.5045,
      "mean_token_accuracy": 0.8426496386528015,
      "num_tokens": 295998034.0,
      "step": 1858
    },
    {
      "epoch": 0.9455747711088505,
      "grad_norm": 0.9956144690513611,
      "learning_rate": 1e-05,
      "loss": 0.5207,
      "mean_token_accuracy": 0.8375461101531982,
      "num_tokens": 296142665.0,
      "step": 1859
    },
    {
      "epoch": 0.9460834181078331,
      "grad_norm": 0.9501311182975769,
      "learning_rate": 1e-05,
      "loss": 0.4997,
      "mean_token_accuracy": 0.8430957794189453,
      "num_tokens": 296290081.0,
      "step": 1860
    },
    {
      "epoch": 0.9465920651068158,
      "grad_norm": 1.0607683658599854,
      "learning_rate": 1e-05,
      "loss": 0.5181,
      "mean_token_accuracy": 0.8405077457427979,
      "num_tokens": 296439741.0,
      "step": 1861
    },
    {
      "epoch": 0.9471007121057986,
      "grad_norm": 0.9842624664306641,
      "learning_rate": 1e-05,
      "loss": 0.49,
      "mean_token_accuracy": 0.8464711904525757,
      "num_tokens": 296598043.0,
      "step": 1862
    },
    {
      "epoch": 0.9476093591047813,
      "grad_norm": 1.018387794494629,
      "learning_rate": 1e-05,
      "loss": 0.4958,
      "mean_token_accuracy": 0.8460543751716614,
      "num_tokens": 296748422.0,
      "step": 1863
    },
    {
      "epoch": 0.948118006103764,
      "grad_norm": 1.0162421464920044,
      "learning_rate": 1e-05,
      "loss": 0.5107,
      "mean_token_accuracy": 0.8393254280090332,
      "num_tokens": 296907598.0,
      "step": 1864
    },
    {
      "epoch": 0.9486266531027467,
      "grad_norm": 1.0726654529571533,
      "learning_rate": 1e-05,
      "loss": 0.5236,
      "mean_token_accuracy": 0.8362458348274231,
      "num_tokens": 297068522.0,
      "step": 1865
    },
    {
      "epoch": 0.9491353001017294,
      "grad_norm": 0.9776252508163452,
      "learning_rate": 1e-05,
      "loss": 0.4941,
      "mean_token_accuracy": 0.8452169895172119,
      "num_tokens": 297229538.0,
      "step": 1866
    },
    {
      "epoch": 0.9496439471007121,
      "grad_norm": 1.0472756624221802,
      "learning_rate": 1e-05,
      "loss": 0.5119,
      "mean_token_accuracy": 0.8398129940032959,
      "num_tokens": 297390837.0,
      "step": 1867
    },
    {
      "epoch": 0.9501525940996948,
      "grad_norm": 1.1011399030685425,
      "learning_rate": 1e-05,
      "loss": 0.5551,
      "mean_token_accuracy": 0.8309721946716309,
      "num_tokens": 297545849.0,
      "step": 1868
    },
    {
      "epoch": 0.9506612410986776,
      "grad_norm": 1.0296998023986816,
      "learning_rate": 1e-05,
      "loss": 0.4996,
      "mean_token_accuracy": 0.8422638177871704,
      "num_tokens": 297702474.0,
      "step": 1869
    },
    {
      "epoch": 0.9511698880976602,
      "grad_norm": 1.101548194885254,
      "learning_rate": 1e-05,
      "loss": 0.5465,
      "mean_token_accuracy": 0.831012487411499,
      "num_tokens": 297857385.0,
      "step": 1870
    },
    {
      "epoch": 0.9516785350966429,
      "grad_norm": 1.0518685579299927,
      "learning_rate": 1e-05,
      "loss": 0.4988,
      "mean_token_accuracy": 0.8414741158485413,
      "num_tokens": 298007755.0,
      "step": 1871
    },
    {
      "epoch": 0.9521871820956256,
      "grad_norm": 1.0469207763671875,
      "learning_rate": 1e-05,
      "loss": 0.5145,
      "mean_token_accuracy": 0.8405040502548218,
      "num_tokens": 298178437.0,
      "step": 1872
    },
    {
      "epoch": 0.9526958290946084,
      "grad_norm": 1.1140328645706177,
      "learning_rate": 1e-05,
      "loss": 0.5155,
      "mean_token_accuracy": 0.8406937122344971,
      "num_tokens": 298335117.0,
      "step": 1873
    },
    {
      "epoch": 0.953204476093591,
      "grad_norm": 1.0772130489349365,
      "learning_rate": 1e-05,
      "loss": 0.4882,
      "mean_token_accuracy": 0.8448014855384827,
      "num_tokens": 298502544.0,
      "step": 1874
    },
    {
      "epoch": 0.9537131230925737,
      "grad_norm": 1.6122941970825195,
      "learning_rate": 1e-05,
      "loss": 0.5114,
      "mean_token_accuracy": 0.8376173377037048,
      "num_tokens": 298660517.0,
      "step": 1875
    },
    {
      "epoch": 0.9542217700915565,
      "grad_norm": 1.0106393098831177,
      "learning_rate": 1e-05,
      "loss": 0.5264,
      "mean_token_accuracy": 0.8369991779327393,
      "num_tokens": 298813928.0,
      "step": 1876
    },
    {
      "epoch": 0.9547304170905392,
      "grad_norm": 0.9368502497673035,
      "learning_rate": 1e-05,
      "loss": 0.4926,
      "mean_token_accuracy": 0.8439090251922607,
      "num_tokens": 298982892.0,
      "step": 1877
    },
    {
      "epoch": 0.9552390640895219,
      "grad_norm": 0.9868039488792419,
      "learning_rate": 1e-05,
      "loss": 0.4975,
      "mean_token_accuracy": 0.8426105976104736,
      "num_tokens": 299145054.0,
      "step": 1878
    },
    {
      "epoch": 0.9557477110885045,
      "grad_norm": 0.9505625367164612,
      "learning_rate": 1e-05,
      "loss": 0.4993,
      "mean_token_accuracy": 0.8433289527893066,
      "num_tokens": 299310640.0,
      "step": 1879
    },
    {
      "epoch": 0.9562563580874873,
      "grad_norm": 1.135102391242981,
      "learning_rate": 1e-05,
      "loss": 0.5154,
      "mean_token_accuracy": 0.8381885290145874,
      "num_tokens": 299484449.0,
      "step": 1880
    },
    {
      "epoch": 0.95676500508647,
      "grad_norm": 1.0819809436798096,
      "learning_rate": 1e-05,
      "loss": 0.5274,
      "mean_token_accuracy": 0.8357242345809937,
      "num_tokens": 299657556.0,
      "step": 1881
    },
    {
      "epoch": 0.9572736520854527,
      "grad_norm": 1.0308226346969604,
      "learning_rate": 1e-05,
      "loss": 0.5189,
      "mean_token_accuracy": 0.8379658460617065,
      "num_tokens": 299821866.0,
      "step": 1882
    },
    {
      "epoch": 0.9577822990844354,
      "grad_norm": 1.0907171964645386,
      "learning_rate": 1e-05,
      "loss": 0.5252,
      "mean_token_accuracy": 0.8378491401672363,
      "num_tokens": 299975575.0,
      "step": 1883
    },
    {
      "epoch": 0.9582909460834181,
      "grad_norm": 1.1639477014541626,
      "learning_rate": 1e-05,
      "loss": 0.5145,
      "mean_token_accuracy": 0.8402724862098694,
      "num_tokens": 300122149.0,
      "step": 1884
    },
    {
      "epoch": 0.9587995930824008,
      "grad_norm": 0.9936185479164124,
      "learning_rate": 1e-05,
      "loss": 0.5247,
      "mean_token_accuracy": 0.8354524374008179,
      "num_tokens": 300275629.0,
      "step": 1885
    },
    {
      "epoch": 0.9593082400813835,
      "grad_norm": 1.031510829925537,
      "learning_rate": 1e-05,
      "loss": 0.5151,
      "mean_token_accuracy": 0.8391316533088684,
      "num_tokens": 300434761.0,
      "step": 1886
    },
    {
      "epoch": 0.9598168870803663,
      "grad_norm": 1.049988865852356,
      "learning_rate": 1e-05,
      "loss": 0.5134,
      "mean_token_accuracy": 0.8404117822647095,
      "num_tokens": 300595753.0,
      "step": 1887
    },
    {
      "epoch": 0.960325534079349,
      "grad_norm": 1.060204029083252,
      "learning_rate": 1e-05,
      "loss": 0.5242,
      "mean_token_accuracy": 0.8364053964614868,
      "num_tokens": 300761476.0,
      "step": 1888
    },
    {
      "epoch": 0.9608341810783316,
      "grad_norm": 1.0777702331542969,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.8494613170623779,
      "num_tokens": 300907255.0,
      "step": 1889
    },
    {
      "epoch": 0.9613428280773143,
      "grad_norm": 0.9640669822692871,
      "learning_rate": 1e-05,
      "loss": 0.5083,
      "mean_token_accuracy": 0.8406792283058167,
      "num_tokens": 301078714.0,
      "step": 1890
    },
    {
      "epoch": 0.9618514750762971,
      "grad_norm": 1.027019739151001,
      "learning_rate": 1e-05,
      "loss": 0.5121,
      "mean_token_accuracy": 0.8419984579086304,
      "num_tokens": 301235054.0,
      "step": 1891
    },
    {
      "epoch": 0.9623601220752798,
      "grad_norm": 1.0344923734664917,
      "learning_rate": 1e-05,
      "loss": 0.5137,
      "mean_token_accuracy": 0.8401339054107666,
      "num_tokens": 301390887.0,
      "step": 1892
    },
    {
      "epoch": 0.9628687690742624,
      "grad_norm": 0.9816353917121887,
      "learning_rate": 1e-05,
      "loss": 0.526,
      "mean_token_accuracy": 0.8368819952011108,
      "num_tokens": 301553051.0,
      "step": 1893
    },
    {
      "epoch": 0.9633774160732451,
      "grad_norm": 1.041377067565918,
      "learning_rate": 1e-05,
      "loss": 0.5005,
      "mean_token_accuracy": 0.8439712524414062,
      "num_tokens": 301709240.0,
      "step": 1894
    },
    {
      "epoch": 0.9638860630722279,
      "grad_norm": 0.9798972010612488,
      "learning_rate": 1e-05,
      "loss": 0.5152,
      "mean_token_accuracy": 0.8388742804527283,
      "num_tokens": 301873555.0,
      "step": 1895
    },
    {
      "epoch": 0.9643947100712106,
      "grad_norm": 1.0527175664901733,
      "learning_rate": 1e-05,
      "loss": 0.519,
      "mean_token_accuracy": 0.8380733728408813,
      "num_tokens": 302020968.0,
      "step": 1896
    },
    {
      "epoch": 0.9649033570701933,
      "grad_norm": 1.01810622215271,
      "learning_rate": 1e-05,
      "loss": 0.5486,
      "mean_token_accuracy": 0.8303346037864685,
      "num_tokens": 302178645.0,
      "step": 1897
    },
    {
      "epoch": 0.965412004069176,
      "grad_norm": 1.0569944381713867,
      "learning_rate": 1e-05,
      "loss": 0.5532,
      "mean_token_accuracy": 0.8283469676971436,
      "num_tokens": 302349196.0,
      "step": 1898
    },
    {
      "epoch": 0.9659206510681587,
      "grad_norm": 1.0374387502670288,
      "learning_rate": 1e-05,
      "loss": 0.5273,
      "mean_token_accuracy": 0.8351933360099792,
      "num_tokens": 302506292.0,
      "step": 1899
    },
    {
      "epoch": 0.9664292980671414,
      "grad_norm": 1.0482120513916016,
      "learning_rate": 1e-05,
      "loss": 0.5322,
      "mean_token_accuracy": 0.8335620164871216,
      "num_tokens": 302661848.0,
      "step": 1900
    },
    {
      "epoch": 0.9669379450661241,
      "grad_norm": 0.978890597820282,
      "learning_rate": 1e-05,
      "loss": 0.5473,
      "mean_token_accuracy": 0.8288494944572449,
      "num_tokens": 302822413.0,
      "step": 1901
    },
    {
      "epoch": 0.9674465920651069,
      "grad_norm": 0.9532626867294312,
      "learning_rate": 1e-05,
      "loss": 0.4979,
      "mean_token_accuracy": 0.8441656827926636,
      "num_tokens": 302985335.0,
      "step": 1902
    },
    {
      "epoch": 0.9679552390640895,
      "grad_norm": 1.0545532703399658,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.843116044998169,
      "num_tokens": 303149395.0,
      "step": 1903
    },
    {
      "epoch": 0.9684638860630722,
      "grad_norm": 1.0217972993850708,
      "learning_rate": 1e-05,
      "loss": 0.4876,
      "mean_token_accuracy": 0.8471866846084595,
      "num_tokens": 303312885.0,
      "step": 1904
    },
    {
      "epoch": 0.9689725330620549,
      "grad_norm": 0.9480841159820557,
      "learning_rate": 1e-05,
      "loss": 0.5242,
      "mean_token_accuracy": 0.8366263508796692,
      "num_tokens": 303478385.0,
      "step": 1905
    },
    {
      "epoch": 0.9694811800610377,
      "grad_norm": 1.0249661207199097,
      "learning_rate": 1e-05,
      "loss": 0.5149,
      "mean_token_accuracy": 0.8401471972465515,
      "num_tokens": 303632397.0,
      "step": 1906
    },
    {
      "epoch": 0.9699898270600203,
      "grad_norm": 0.9777758121490479,
      "learning_rate": 1e-05,
      "loss": 0.5155,
      "mean_token_accuracy": 0.8376993536949158,
      "num_tokens": 303793145.0,
      "step": 1907
    },
    {
      "epoch": 0.970498474059003,
      "grad_norm": 1.0128408670425415,
      "learning_rate": 1e-05,
      "loss": 0.5355,
      "mean_token_accuracy": 0.833324670791626,
      "num_tokens": 303954402.0,
      "step": 1908
    },
    {
      "epoch": 0.9710071210579858,
      "grad_norm": 1.0025849342346191,
      "learning_rate": 1e-05,
      "loss": 0.519,
      "mean_token_accuracy": 0.837595522403717,
      "num_tokens": 304115485.0,
      "step": 1909
    },
    {
      "epoch": 0.9715157680569685,
      "grad_norm": 0.9397594332695007,
      "learning_rate": 1e-05,
      "loss": 0.5176,
      "mean_token_accuracy": 0.8387343287467957,
      "num_tokens": 304285608.0,
      "step": 1910
    },
    {
      "epoch": 0.9720244150559512,
      "grad_norm": 0.9547067284584045,
      "learning_rate": 1e-05,
      "loss": 0.4658,
      "mean_token_accuracy": 0.8515326976776123,
      "num_tokens": 304444429.0,
      "step": 1911
    },
    {
      "epoch": 0.9725330620549338,
      "grad_norm": 1.0898922681808472,
      "learning_rate": 1e-05,
      "loss": 0.4982,
      "mean_token_accuracy": 0.8440953493118286,
      "num_tokens": 304604968.0,
      "step": 1912
    },
    {
      "epoch": 0.9730417090539166,
      "grad_norm": 1.0754295587539673,
      "learning_rate": 1e-05,
      "loss": 0.4895,
      "mean_token_accuracy": 0.8468303084373474,
      "num_tokens": 304762895.0,
      "step": 1913
    },
    {
      "epoch": 0.9735503560528993,
      "grad_norm": 1.0889835357666016,
      "learning_rate": 1e-05,
      "loss": 0.5106,
      "mean_token_accuracy": 0.8393197059631348,
      "num_tokens": 304919739.0,
      "step": 1914
    },
    {
      "epoch": 0.974059003051882,
      "grad_norm": 0.9590617418289185,
      "learning_rate": 1e-05,
      "loss": 0.5234,
      "mean_token_accuracy": 0.8376962542533875,
      "num_tokens": 305084650.0,
      "step": 1915
    },
    {
      "epoch": 0.9745676500508647,
      "grad_norm": 0.9880885481834412,
      "learning_rate": 1e-05,
      "loss": 0.5042,
      "mean_token_accuracy": 0.8409290313720703,
      "num_tokens": 305245025.0,
      "step": 1916
    },
    {
      "epoch": 0.9750762970498474,
      "grad_norm": 0.996501088142395,
      "learning_rate": 1e-05,
      "loss": 0.5005,
      "mean_token_accuracy": 0.8432765007019043,
      "num_tokens": 305403475.0,
      "step": 1917
    },
    {
      "epoch": 0.9755849440488301,
      "grad_norm": 0.9220758080482483,
      "learning_rate": 1e-05,
      "loss": 0.5154,
      "mean_token_accuracy": 0.8385381102561951,
      "num_tokens": 305554045.0,
      "step": 1918
    },
    {
      "epoch": 0.9760935910478128,
      "grad_norm": 1.0082985162734985,
      "learning_rate": 1e-05,
      "loss": 0.4907,
      "mean_token_accuracy": 0.846260666847229,
      "num_tokens": 305712579.0,
      "step": 1919
    },
    {
      "epoch": 0.9766022380467956,
      "grad_norm": 0.987065851688385,
      "learning_rate": 1e-05,
      "loss": 0.4949,
      "mean_token_accuracy": 0.8460931777954102,
      "num_tokens": 305880937.0,
      "step": 1920
    },
    {
      "epoch": 0.9771108850457783,
      "grad_norm": 0.9880823493003845,
      "learning_rate": 1e-05,
      "loss": 0.4994,
      "mean_token_accuracy": 0.8439615964889526,
      "num_tokens": 306037487.0,
      "step": 1921
    },
    {
      "epoch": 0.9776195320447609,
      "grad_norm": 0.9675670862197876,
      "learning_rate": 1e-05,
      "loss": 0.5101,
      "mean_token_accuracy": 0.8417237401008606,
      "num_tokens": 306205585.0,
      "step": 1922
    },
    {
      "epoch": 0.9781281790437436,
      "grad_norm": 1.1072001457214355,
      "learning_rate": 1e-05,
      "loss": 0.5216,
      "mean_token_accuracy": 0.8380455374717712,
      "num_tokens": 306363764.0,
      "step": 1923
    },
    {
      "epoch": 0.9786368260427264,
      "grad_norm": 0.98995441198349,
      "learning_rate": 1e-05,
      "loss": 0.501,
      "mean_token_accuracy": 0.843612551689148,
      "num_tokens": 306521564.0,
      "step": 1924
    },
    {
      "epoch": 0.9791454730417091,
      "grad_norm": 1.060228705406189,
      "learning_rate": 1e-05,
      "loss": 0.5014,
      "mean_token_accuracy": 0.8430065512657166,
      "num_tokens": 306674429.0,
      "step": 1925
    },
    {
      "epoch": 0.9796541200406917,
      "grad_norm": 1.018702507019043,
      "learning_rate": 1e-05,
      "loss": 0.5409,
      "mean_token_accuracy": 0.8309513330459595,
      "num_tokens": 306832682.0,
      "step": 1926
    },
    {
      "epoch": 0.9801627670396744,
      "grad_norm": 1.13094961643219,
      "learning_rate": 1e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8488995432853699,
      "num_tokens": 306990015.0,
      "step": 1927
    },
    {
      "epoch": 0.9806714140386572,
      "grad_norm": 0.9446847438812256,
      "learning_rate": 1e-05,
      "loss": 0.5033,
      "mean_token_accuracy": 0.8416002988815308,
      "num_tokens": 307155638.0,
      "step": 1928
    },
    {
      "epoch": 0.9811800610376399,
      "grad_norm": 1.0495901107788086,
      "learning_rate": 1e-05,
      "loss": 0.4828,
      "mean_token_accuracy": 0.8481249809265137,
      "num_tokens": 307317052.0,
      "step": 1929
    },
    {
      "epoch": 0.9816887080366226,
      "grad_norm": 0.997715950012207,
      "learning_rate": 1e-05,
      "loss": 0.5117,
      "mean_token_accuracy": 0.8404562473297119,
      "num_tokens": 307461975.0,
      "step": 1930
    },
    {
      "epoch": 0.9821973550356052,
      "grad_norm": 1.0938962697982788,
      "learning_rate": 1e-05,
      "loss": 0.5479,
      "mean_token_accuracy": 0.8287118673324585,
      "num_tokens": 307618995.0,
      "step": 1931
    },
    {
      "epoch": 0.982706002034588,
      "grad_norm": 1.0079737901687622,
      "learning_rate": 1e-05,
      "loss": 0.5047,
      "mean_token_accuracy": 0.8424063920974731,
      "num_tokens": 307787106.0,
      "step": 1932
    },
    {
      "epoch": 0.9832146490335707,
      "grad_norm": 0.9513643980026245,
      "learning_rate": 1e-05,
      "loss": 0.5141,
      "mean_token_accuracy": 0.8406564593315125,
      "num_tokens": 307957991.0,
      "step": 1933
    },
    {
      "epoch": 0.9837232960325534,
      "grad_norm": 1.1360417604446411,
      "learning_rate": 1e-05,
      "loss": 0.531,
      "mean_token_accuracy": 0.8353432416915894,
      "num_tokens": 308114670.0,
      "step": 1934
    },
    {
      "epoch": 0.9842319430315362,
      "grad_norm": 1.0051788091659546,
      "learning_rate": 1e-05,
      "loss": 0.5346,
      "mean_token_accuracy": 0.8331654071807861,
      "num_tokens": 308273482.0,
      "step": 1935
    },
    {
      "epoch": 0.9847405900305188,
      "grad_norm": 0.9172512292861938,
      "learning_rate": 1e-05,
      "loss": 0.5107,
      "mean_token_accuracy": 0.8405485153198242,
      "num_tokens": 308438717.0,
      "step": 1936
    },
    {
      "epoch": 0.9852492370295015,
      "grad_norm": 1.0240017175674438,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8513554334640503,
      "num_tokens": 308606239.0,
      "step": 1937
    },
    {
      "epoch": 0.9857578840284842,
      "grad_norm": 0.9732988476753235,
      "learning_rate": 1e-05,
      "loss": 0.513,
      "mean_token_accuracy": 0.8397548198699951,
      "num_tokens": 308770389.0,
      "step": 1938
    },
    {
      "epoch": 0.986266531027467,
      "grad_norm": 1.0476912260055542,
      "learning_rate": 1e-05,
      "loss": 0.5123,
      "mean_token_accuracy": 0.8402565121650696,
      "num_tokens": 308938986.0,
      "step": 1939
    },
    {
      "epoch": 0.9867751780264497,
      "grad_norm": 0.9261658787727356,
      "learning_rate": 1e-05,
      "loss": 0.4896,
      "mean_token_accuracy": 0.8465872406959534,
      "num_tokens": 309098467.0,
      "step": 1940
    },
    {
      "epoch": 0.9872838250254323,
      "grad_norm": 1.0257370471954346,
      "learning_rate": 1e-05,
      "loss": 0.5159,
      "mean_token_accuracy": 0.837329626083374,
      "num_tokens": 309271767.0,
      "step": 1941
    },
    {
      "epoch": 0.987792472024415,
      "grad_norm": 0.9949601292610168,
      "learning_rate": 1e-05,
      "loss": 0.5321,
      "mean_token_accuracy": 0.8342413902282715,
      "num_tokens": 309430767.0,
      "step": 1942
    },
    {
      "epoch": 0.9883011190233978,
      "grad_norm": 1.031502604484558,
      "learning_rate": 1e-05,
      "loss": 0.5133,
      "mean_token_accuracy": 0.8384732604026794,
      "num_tokens": 309588008.0,
      "step": 1943
    },
    {
      "epoch": 0.9888097660223805,
      "grad_norm": 0.976768434047699,
      "learning_rate": 1e-05,
      "loss": 0.5165,
      "mean_token_accuracy": 0.8383283615112305,
      "num_tokens": 309743337.0,
      "step": 1944
    },
    {
      "epoch": 0.9893184130213631,
      "grad_norm": 1.033813714981079,
      "learning_rate": 1e-05,
      "loss": 0.5188,
      "mean_token_accuracy": 0.8372408747673035,
      "num_tokens": 309895383.0,
      "step": 1945
    },
    {
      "epoch": 0.9898270600203459,
      "grad_norm": 1.203722596168518,
      "learning_rate": 1e-05,
      "loss": 0.5355,
      "mean_token_accuracy": 0.8340418338775635,
      "num_tokens": 310051593.0,
      "step": 1946
    },
    {
      "epoch": 0.9903357070193286,
      "grad_norm": 1.0881632566452026,
      "learning_rate": 1e-05,
      "loss": 0.5241,
      "mean_token_accuracy": 0.8362948894500732,
      "num_tokens": 310222144.0,
      "step": 1947
    },
    {
      "epoch": 0.9908443540183113,
      "grad_norm": 0.967003345489502,
      "learning_rate": 1e-05,
      "loss": 0.487,
      "mean_token_accuracy": 0.8465834259986877,
      "num_tokens": 310376369.0,
      "step": 1948
    },
    {
      "epoch": 0.991353001017294,
      "grad_norm": 0.9813733696937561,
      "learning_rate": 1e-05,
      "loss": 0.5329,
      "mean_token_accuracy": 0.8346635103225708,
      "num_tokens": 310540773.0,
      "step": 1949
    },
    {
      "epoch": 0.9918616480162767,
      "grad_norm": 1.0921562910079956,
      "learning_rate": 1e-05,
      "loss": 0.5189,
      "mean_token_accuracy": 0.8375669121742249,
      "num_tokens": 310705795.0,
      "step": 1950
    },
    {
      "epoch": 0.9923702950152594,
      "grad_norm": 0.9453521966934204,
      "learning_rate": 1e-05,
      "loss": 0.4979,
      "mean_token_accuracy": 0.8441818356513977,
      "num_tokens": 310870435.0,
      "step": 1951
    },
    {
      "epoch": 0.9928789420142421,
      "grad_norm": 0.9923408031463623,
      "learning_rate": 1e-05,
      "loss": 0.5326,
      "mean_token_accuracy": 0.833539605140686,
      "num_tokens": 311023220.0,
      "step": 1952
    },
    {
      "epoch": 0.9933875890132248,
      "grad_norm": 0.971986711025238,
      "learning_rate": 1e-05,
      "loss": 0.5149,
      "mean_token_accuracy": 0.8401809334754944,
      "num_tokens": 311192559.0,
      "step": 1953
    },
    {
      "epoch": 0.9938962360122076,
      "grad_norm": 1.054033637046814,
      "learning_rate": 1e-05,
      "loss": 0.5368,
      "mean_token_accuracy": 0.8325077891349792,
      "num_tokens": 311361294.0,
      "step": 1954
    },
    {
      "epoch": 0.9944048830111902,
      "grad_norm": 1.0171631574630737,
      "learning_rate": 1e-05,
      "loss": 0.5425,
      "mean_token_accuracy": 0.830352783203125,
      "num_tokens": 311513624.0,
      "step": 1955
    },
    {
      "epoch": 0.9949135300101729,
      "grad_norm": 1.076941967010498,
      "learning_rate": 1e-05,
      "loss": 0.5534,
      "mean_token_accuracy": 0.8284592032432556,
      "num_tokens": 311682933.0,
      "step": 1956
    },
    {
      "epoch": 0.9954221770091557,
      "grad_norm": 1.0492786169052124,
      "learning_rate": 1e-05,
      "loss": 0.5291,
      "mean_token_accuracy": 0.8351242542266846,
      "num_tokens": 311836701.0,
      "step": 1957
    },
    {
      "epoch": 0.9959308240081384,
      "grad_norm": 0.9567775726318359,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.8453713655471802,
      "num_tokens": 311997757.0,
      "step": 1958
    },
    {
      "epoch": 0.996439471007121,
      "grad_norm": 1.0214710235595703,
      "learning_rate": 1e-05,
      "loss": 0.5105,
      "mean_token_accuracy": 0.840237021446228,
      "num_tokens": 312141716.0,
      "step": 1959
    },
    {
      "epoch": 0.9969481180061037,
      "grad_norm": 0.9785928130149841,
      "learning_rate": 1e-05,
      "loss": 0.5123,
      "mean_token_accuracy": 0.8398464918136597,
      "num_tokens": 312300232.0,
      "step": 1960
    },
    {
      "epoch": 0.9974567650050865,
      "grad_norm": 0.9771664142608643,
      "learning_rate": 1e-05,
      "loss": 0.5236,
      "mean_token_accuracy": 0.8375827670097351,
      "num_tokens": 312460325.0,
      "step": 1961
    },
    {
      "epoch": 0.9979654120040692,
      "grad_norm": 0.9888442754745483,
      "learning_rate": 1e-05,
      "loss": 0.5002,
      "mean_token_accuracy": 0.844666600227356,
      "num_tokens": 312617272.0,
      "step": 1962
    },
    {
      "epoch": 0.9984740590030519,
      "grad_norm": 0.9254450798034668,
      "learning_rate": 1e-05,
      "loss": 0.5047,
      "mean_token_accuracy": 0.8417128324508667,
      "num_tokens": 312782970.0,
      "step": 1963
    },
    {
      "epoch": 0.9989827060020345,
      "grad_norm": 1.0160105228424072,
      "learning_rate": 1e-05,
      "loss": 0.5225,
      "mean_token_accuracy": 0.8384478092193604,
      "num_tokens": 312940535.0,
      "step": 1964
    },
    {
      "epoch": 0.9994913530010173,
      "grad_norm": 0.9914231896400452,
      "learning_rate": 1e-05,
      "loss": 0.5035,
      "mean_token_accuracy": 0.8427067995071411,
      "num_tokens": 313094817.0,
      "step": 1965
    },
    {
      "epoch": 1.0,
      "grad_norm": 1.0151771306991577,
      "learning_rate": 1e-05,
      "loss": 0.5262,
      "mean_token_accuracy": 0.8351953625679016,
      "num_tokens": 313255150.0,
      "step": 1966
    },
    {
      "epoch": 1.0005086469989828,
      "grad_norm": 1.1044137477874756,
      "learning_rate": 1e-05,
      "loss": 0.4927,
      "mean_token_accuracy": 0.8446985483169556,
      "num_tokens": 313416474.0,
      "step": 1967
    },
    {
      "epoch": 1.0010172939979654,
      "grad_norm": 1.0383646488189697,
      "learning_rate": 1e-05,
      "loss": 0.4672,
      "mean_token_accuracy": 0.8528542518615723,
      "num_tokens": 313576480.0,
      "step": 1968
    },
    {
      "epoch": 1.0015259409969481,
      "grad_norm": 1.0455737113952637,
      "learning_rate": 1e-05,
      "loss": 0.5156,
      "mean_token_accuracy": 0.8373713493347168,
      "num_tokens": 313725157.0,
      "step": 1969
    },
    {
      "epoch": 1.002034587995931,
      "grad_norm": 1.0510598421096802,
      "learning_rate": 1e-05,
      "loss": 0.5082,
      "mean_token_accuracy": 0.8405560851097107,
      "num_tokens": 313894362.0,
      "step": 1970
    },
    {
      "epoch": 1.0025432349949135,
      "grad_norm": 1.1204897165298462,
      "learning_rate": 1e-05,
      "loss": 0.506,
      "mean_token_accuracy": 0.840548038482666,
      "num_tokens": 314046204.0,
      "step": 1971
    },
    {
      "epoch": 1.0030518819938963,
      "grad_norm": 1.0423732995986938,
      "learning_rate": 1e-05,
      "loss": 0.471,
      "mean_token_accuracy": 0.849092960357666,
      "num_tokens": 314199033.0,
      "step": 1972
    },
    {
      "epoch": 1.0035605289928788,
      "grad_norm": 0.9673753380775452,
      "learning_rate": 1e-05,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8502670526504517,
      "num_tokens": 314357461.0,
      "step": 1973
    },
    {
      "epoch": 1.0040691759918616,
      "grad_norm": 1.0575178861618042,
      "learning_rate": 1e-05,
      "loss": 0.4874,
      "mean_token_accuracy": 0.8461350202560425,
      "num_tokens": 314529806.0,
      "step": 1974
    },
    {
      "epoch": 1.0045778229908444,
      "grad_norm": 1.0000046491622925,
      "learning_rate": 1e-05,
      "loss": 0.4759,
      "mean_token_accuracy": 0.8510136008262634,
      "num_tokens": 314685706.0,
      "step": 1975
    },
    {
      "epoch": 1.005086469989827,
      "grad_norm": 0.9816222786903381,
      "learning_rate": 1e-05,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8437051773071289,
      "num_tokens": 314836980.0,
      "step": 1976
    },
    {
      "epoch": 1.0055951169888098,
      "grad_norm": 0.9830356240272522,
      "learning_rate": 1e-05,
      "loss": 0.507,
      "mean_token_accuracy": 0.8409254550933838,
      "num_tokens": 315004512.0,
      "step": 1977
    },
    {
      "epoch": 1.0061037639877926,
      "grad_norm": 1.0614138841629028,
      "learning_rate": 1e-05,
      "loss": 0.5242,
      "mean_token_accuracy": 0.8373826742172241,
      "num_tokens": 315155851.0,
      "step": 1978
    },
    {
      "epoch": 1.0066124109867751,
      "grad_norm": 1.0838987827301025,
      "learning_rate": 1e-05,
      "loss": 0.4765,
      "mean_token_accuracy": 0.8484848141670227,
      "num_tokens": 315323264.0,
      "step": 1979
    },
    {
      "epoch": 1.007121057985758,
      "grad_norm": 1.0375914573669434,
      "learning_rate": 1e-05,
      "loss": 0.4774,
      "mean_token_accuracy": 0.8493205904960632,
      "num_tokens": 315477553.0,
      "step": 1980
    },
    {
      "epoch": 1.0076297049847407,
      "grad_norm": 1.0500863790512085,
      "learning_rate": 1e-05,
      "loss": 0.5053,
      "mean_token_accuracy": 0.8408976793289185,
      "num_tokens": 315636869.0,
      "step": 1981
    },
    {
      "epoch": 1.0081383519837233,
      "grad_norm": 1.0597950220108032,
      "learning_rate": 1e-05,
      "loss": 0.4785,
      "mean_token_accuracy": 0.8486466407775879,
      "num_tokens": 315789205.0,
      "step": 1982
    },
    {
      "epoch": 1.008646998982706,
      "grad_norm": 1.0866880416870117,
      "learning_rate": 1e-05,
      "loss": 0.4856,
      "mean_token_accuracy": 0.8461682200431824,
      "num_tokens": 315949798.0,
      "step": 1983
    },
    {
      "epoch": 1.0091556459816886,
      "grad_norm": 1.0247963666915894,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.843558132648468,
      "num_tokens": 316108041.0,
      "step": 1984
    },
    {
      "epoch": 1.0096642929806714,
      "grad_norm": 1.1197168827056885,
      "learning_rate": 1e-05,
      "loss": 0.4658,
      "mean_token_accuracy": 0.850692629814148,
      "num_tokens": 316271100.0,
      "step": 1985
    },
    {
      "epoch": 1.0101729399796542,
      "grad_norm": 1.0855802297592163,
      "learning_rate": 1e-05,
      "loss": 0.51,
      "mean_token_accuracy": 0.840994656085968,
      "num_tokens": 316433915.0,
      "step": 1986
    },
    {
      "epoch": 1.0106815869786367,
      "grad_norm": 1.0583574771881104,
      "learning_rate": 1e-05,
      "loss": 0.5189,
      "mean_token_accuracy": 0.8364496231079102,
      "num_tokens": 316609023.0,
      "step": 1987
    },
    {
      "epoch": 1.0111902339776195,
      "grad_norm": 0.9981748461723328,
      "learning_rate": 1e-05,
      "loss": 0.4573,
      "mean_token_accuracy": 0.8547645211219788,
      "num_tokens": 316764032.0,
      "step": 1988
    },
    {
      "epoch": 1.0116988809766023,
      "grad_norm": 0.9980987906455994,
      "learning_rate": 1e-05,
      "loss": 0.5112,
      "mean_token_accuracy": 0.8400788307189941,
      "num_tokens": 316924785.0,
      "step": 1989
    },
    {
      "epoch": 1.0122075279755849,
      "grad_norm": 1.0092408657073975,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8480200171470642,
      "num_tokens": 317068693.0,
      "step": 1990
    },
    {
      "epoch": 1.0127161749745677,
      "grad_norm": 1.148630142211914,
      "learning_rate": 1e-05,
      "loss": 0.4868,
      "mean_token_accuracy": 0.846847653388977,
      "num_tokens": 317231900.0,
      "step": 1991
    },
    {
      "epoch": 1.0132248219735505,
      "grad_norm": 1.0388243198394775,
      "learning_rate": 1e-05,
      "loss": 0.4764,
      "mean_token_accuracy": 0.8494249582290649,
      "num_tokens": 317399513.0,
      "step": 1992
    },
    {
      "epoch": 1.013733468972533,
      "grad_norm": 0.9582164883613586,
      "learning_rate": 1e-05,
      "loss": 0.4889,
      "mean_token_accuracy": 0.8472070693969727,
      "num_tokens": 317561644.0,
      "step": 1993
    },
    {
      "epoch": 1.0142421159715158,
      "grad_norm": 0.9729287028312683,
      "learning_rate": 1e-05,
      "loss": 0.4728,
      "mean_token_accuracy": 0.8492978811264038,
      "num_tokens": 317722640.0,
      "step": 1994
    },
    {
      "epoch": 1.0147507629704984,
      "grad_norm": 1.0420944690704346,
      "learning_rate": 1e-05,
      "loss": 0.4729,
      "mean_token_accuracy": 0.8504926562309265,
      "num_tokens": 317881374.0,
      "step": 1995
    },
    {
      "epoch": 1.0152594099694812,
      "grad_norm": 1.0152589082717896,
      "learning_rate": 1e-05,
      "loss": 0.4817,
      "mean_token_accuracy": 0.8495712280273438,
      "num_tokens": 318033418.0,
      "step": 1996
    },
    {
      "epoch": 1.015768056968464,
      "grad_norm": 1.0669312477111816,
      "learning_rate": 1e-05,
      "loss": 0.5092,
      "mean_token_accuracy": 0.8412264585494995,
      "num_tokens": 318183443.0,
      "step": 1997
    },
    {
      "epoch": 1.0162767039674465,
      "grad_norm": 1.1268723011016846,
      "learning_rate": 1e-05,
      "loss": 0.4913,
      "mean_token_accuracy": 0.845348596572876,
      "num_tokens": 318339258.0,
      "step": 1998
    },
    {
      "epoch": 1.0167853509664293,
      "grad_norm": 1.0353676080703735,
      "learning_rate": 1e-05,
      "loss": 0.4586,
      "mean_token_accuracy": 0.8535833954811096,
      "num_tokens": 318495267.0,
      "step": 1999
    },
    {
      "epoch": 1.017293997965412,
      "grad_norm": 1.0712075233459473,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.848924994468689,
      "num_tokens": 318648663.0,
      "step": 2000
    },
    {
      "epoch": 1.0178026449643947,
      "grad_norm": 1.197359323501587,
      "learning_rate": 1e-05,
      "loss": 0.4963,
      "mean_token_accuracy": 0.8441604375839233,
      "num_tokens": 318803431.0,
      "step": 2001
    },
    {
      "epoch": 1.0183112919633774,
      "grad_norm": 1.0463975667953491,
      "learning_rate": 1e-05,
      "loss": 0.4914,
      "mean_token_accuracy": 0.8444896936416626,
      "num_tokens": 318960576.0,
      "step": 2002
    },
    {
      "epoch": 1.0188199389623602,
      "grad_norm": 1.0998681783676147,
      "learning_rate": 1e-05,
      "loss": 0.469,
      "mean_token_accuracy": 0.8531366586685181,
      "num_tokens": 319111546.0,
      "step": 2003
    },
    {
      "epoch": 1.0193285859613428,
      "grad_norm": 1.059656023979187,
      "learning_rate": 1e-05,
      "loss": 0.4952,
      "mean_token_accuracy": 0.8430195450782776,
      "num_tokens": 319276087.0,
      "step": 2004
    },
    {
      "epoch": 1.0198372329603256,
      "grad_norm": 1.1603281497955322,
      "learning_rate": 1e-05,
      "loss": 0.5113,
      "mean_token_accuracy": 0.8404607772827148,
      "num_tokens": 319423273.0,
      "step": 2005
    },
    {
      "epoch": 1.0203458799593081,
      "grad_norm": 1.0294604301452637,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8559034466743469,
      "num_tokens": 319576787.0,
      "step": 2006
    },
    {
      "epoch": 1.020854526958291,
      "grad_norm": 1.0385469198226929,
      "learning_rate": 1e-05,
      "loss": 0.4849,
      "mean_token_accuracy": 0.8464286923408508,
      "num_tokens": 319740271.0,
      "step": 2007
    },
    {
      "epoch": 1.0213631739572737,
      "grad_norm": 1.0036464929580688,
      "learning_rate": 1e-05,
      "loss": 0.4896,
      "mean_token_accuracy": 0.8454139828681946,
      "num_tokens": 319902904.0,
      "step": 2008
    },
    {
      "epoch": 1.0218718209562563,
      "grad_norm": 1.1799579858779907,
      "learning_rate": 1e-05,
      "loss": 0.5066,
      "mean_token_accuracy": 0.8418562412261963,
      "num_tokens": 320057699.0,
      "step": 2009
    },
    {
      "epoch": 1.022380467955239,
      "grad_norm": 1.1496026515960693,
      "learning_rate": 1e-05,
      "loss": 0.4769,
      "mean_token_accuracy": 0.8499530553817749,
      "num_tokens": 320212284.0,
      "step": 2010
    },
    {
      "epoch": 1.0228891149542219,
      "grad_norm": 0.9787399768829346,
      "learning_rate": 1e-05,
      "loss": 0.4947,
      "mean_token_accuracy": 0.8445889949798584,
      "num_tokens": 320374261.0,
      "step": 2011
    },
    {
      "epoch": 1.0233977619532044,
      "grad_norm": 1.1572648286819458,
      "learning_rate": 1e-05,
      "loss": 0.5078,
      "mean_token_accuracy": 0.8388811349868774,
      "num_tokens": 320529125.0,
      "step": 2012
    },
    {
      "epoch": 1.0239064089521872,
      "grad_norm": 1.0809093713760376,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8499780893325806,
      "num_tokens": 320672706.0,
      "step": 2013
    },
    {
      "epoch": 1.02441505595117,
      "grad_norm": 1.005975604057312,
      "learning_rate": 1e-05,
      "loss": 0.486,
      "mean_token_accuracy": 0.8479955196380615,
      "num_tokens": 320834351.0,
      "step": 2014
    },
    {
      "epoch": 1.0249237029501526,
      "grad_norm": 1.150314450263977,
      "learning_rate": 1e-05,
      "loss": 0.4988,
      "mean_token_accuracy": 0.841154158115387,
      "num_tokens": 321001419.0,
      "step": 2015
    },
    {
      "epoch": 1.0254323499491353,
      "grad_norm": 1.0990020036697388,
      "learning_rate": 1e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8472510576248169,
      "num_tokens": 321155407.0,
      "step": 2016
    },
    {
      "epoch": 1.025940996948118,
      "grad_norm": 1.1457676887512207,
      "learning_rate": 1e-05,
      "loss": 0.4997,
      "mean_token_accuracy": 0.8403406739234924,
      "num_tokens": 321311640.0,
      "step": 2017
    },
    {
      "epoch": 1.0264496439471007,
      "grad_norm": 1.0399061441421509,
      "learning_rate": 1e-05,
      "loss": 0.5103,
      "mean_token_accuracy": 0.8391156792640686,
      "num_tokens": 321466786.0,
      "step": 2018
    },
    {
      "epoch": 1.0269582909460835,
      "grad_norm": 1.0857439041137695,
      "learning_rate": 1e-05,
      "loss": 0.4786,
      "mean_token_accuracy": 0.8480024933815002,
      "num_tokens": 321630638.0,
      "step": 2019
    },
    {
      "epoch": 1.027466937945066,
      "grad_norm": 1.072295069694519,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8517623543739319,
      "num_tokens": 321794738.0,
      "step": 2020
    },
    {
      "epoch": 1.0279755849440488,
      "grad_norm": 1.0555133819580078,
      "learning_rate": 1e-05,
      "loss": 0.5279,
      "mean_token_accuracy": 0.8357231616973877,
      "num_tokens": 321953968.0,
      "step": 2021
    },
    {
      "epoch": 1.0284842319430316,
      "grad_norm": 1.181735873222351,
      "learning_rate": 1e-05,
      "loss": 0.4526,
      "mean_token_accuracy": 0.8548371195793152,
      "num_tokens": 322105270.0,
      "step": 2022
    },
    {
      "epoch": 1.0289928789420142,
      "grad_norm": 0.9960180521011353,
      "learning_rate": 1e-05,
      "loss": 0.4867,
      "mean_token_accuracy": 0.8455308079719543,
      "num_tokens": 322267010.0,
      "step": 2023
    },
    {
      "epoch": 1.029501525940997,
      "grad_norm": 1.0796343088150024,
      "learning_rate": 1e-05,
      "loss": 0.4888,
      "mean_token_accuracy": 0.8429710865020752,
      "num_tokens": 322430631.0,
      "step": 2024
    },
    {
      "epoch": 1.0300101729399798,
      "grad_norm": 1.0349819660186768,
      "learning_rate": 1e-05,
      "loss": 0.4803,
      "mean_token_accuracy": 0.8490273952484131,
      "num_tokens": 322593551.0,
      "step": 2025
    },
    {
      "epoch": 1.0305188199389623,
      "grad_norm": 0.9780552387237549,
      "learning_rate": 1e-05,
      "loss": 0.4559,
      "mean_token_accuracy": 0.8544877171516418,
      "num_tokens": 322762321.0,
      "step": 2026
    },
    {
      "epoch": 1.0310274669379451,
      "grad_norm": 1.1404180526733398,
      "learning_rate": 1e-05,
      "loss": 0.4898,
      "mean_token_accuracy": 0.8458755016326904,
      "num_tokens": 322922120.0,
      "step": 2027
    },
    {
      "epoch": 1.0315361139369277,
      "grad_norm": 0.9819681644439697,
      "learning_rate": 1e-05,
      "loss": 0.5019,
      "mean_token_accuracy": 0.8418173789978027,
      "num_tokens": 323079150.0,
      "step": 2028
    },
    {
      "epoch": 1.0320447609359105,
      "grad_norm": 0.976266622543335,
      "learning_rate": 1e-05,
      "loss": 0.4823,
      "mean_token_accuracy": 0.8466558456420898,
      "num_tokens": 323238505.0,
      "step": 2029
    },
    {
      "epoch": 1.0325534079348933,
      "grad_norm": 0.9896778464317322,
      "learning_rate": 1e-05,
      "loss": 0.5024,
      "mean_token_accuracy": 0.8411461710929871,
      "num_tokens": 323409267.0,
      "step": 2030
    },
    {
      "epoch": 1.0330620549338758,
      "grad_norm": 1.007405161857605,
      "learning_rate": 1e-05,
      "loss": 0.5161,
      "mean_token_accuracy": 0.8370198011398315,
      "num_tokens": 323574623.0,
      "step": 2031
    },
    {
      "epoch": 1.0335707019328586,
      "grad_norm": 1.0376770496368408,
      "learning_rate": 1e-05,
      "loss": 0.4955,
      "mean_token_accuracy": 0.8449573516845703,
      "num_tokens": 323717438.0,
      "step": 2032
    },
    {
      "epoch": 1.0340793489318414,
      "grad_norm": 0.9660647511482239,
      "learning_rate": 1e-05,
      "loss": 0.4681,
      "mean_token_accuracy": 0.851511538028717,
      "num_tokens": 323876742.0,
      "step": 2033
    },
    {
      "epoch": 1.034587995930824,
      "grad_norm": 1.0799311399459839,
      "learning_rate": 1e-05,
      "loss": 0.5258,
      "mean_token_accuracy": 0.835211455821991,
      "num_tokens": 324033814.0,
      "step": 2034
    },
    {
      "epoch": 1.0350966429298067,
      "grad_norm": 1.0707277059555054,
      "learning_rate": 1e-05,
      "loss": 0.511,
      "mean_token_accuracy": 0.8382892608642578,
      "num_tokens": 324183869.0,
      "step": 2035
    },
    {
      "epoch": 1.0356052899287893,
      "grad_norm": 1.0144026279449463,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8489046096801758,
      "num_tokens": 324339938.0,
      "step": 2036
    },
    {
      "epoch": 1.036113936927772,
      "grad_norm": 0.952110230922699,
      "learning_rate": 1e-05,
      "loss": 0.5004,
      "mean_token_accuracy": 0.8435638546943665,
      "num_tokens": 324495300.0,
      "step": 2037
    },
    {
      "epoch": 1.0366225839267549,
      "grad_norm": 1.0803865194320679,
      "learning_rate": 1e-05,
      "loss": 0.485,
      "mean_token_accuracy": 0.8461806774139404,
      "num_tokens": 324649242.0,
      "step": 2038
    },
    {
      "epoch": 1.0371312309257374,
      "grad_norm": 1.0200896263122559,
      "learning_rate": 1e-05,
      "loss": 0.5113,
      "mean_token_accuracy": 0.8410135507583618,
      "num_tokens": 324807416.0,
      "step": 2039
    },
    {
      "epoch": 1.0376398779247202,
      "grad_norm": 1.008554458618164,
      "learning_rate": 1e-05,
      "loss": 0.4786,
      "mean_token_accuracy": 0.8479388952255249,
      "num_tokens": 324973812.0,
      "step": 2040
    },
    {
      "epoch": 1.038148524923703,
      "grad_norm": 0.9462720155715942,
      "learning_rate": 1e-05,
      "loss": 0.4948,
      "mean_token_accuracy": 0.8447370529174805,
      "num_tokens": 325141500.0,
      "step": 2041
    },
    {
      "epoch": 1.0386571719226856,
      "grad_norm": 1.0549455881118774,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.8438306450843811,
      "num_tokens": 325293984.0,
      "step": 2042
    },
    {
      "epoch": 1.0391658189216684,
      "grad_norm": 1.0447968244552612,
      "learning_rate": 1e-05,
      "loss": 0.4903,
      "mean_token_accuracy": 0.8438676595687866,
      "num_tokens": 325446428.0,
      "step": 2043
    },
    {
      "epoch": 1.0396744659206512,
      "grad_norm": 1.0543104410171509,
      "learning_rate": 1e-05,
      "loss": 0.5218,
      "mean_token_accuracy": 0.8362342119216919,
      "num_tokens": 325602970.0,
      "step": 2044
    },
    {
      "epoch": 1.0401831129196337,
      "grad_norm": 1.1367326974868774,
      "learning_rate": 1e-05,
      "loss": 0.5208,
      "mean_token_accuracy": 0.8370205760002136,
      "num_tokens": 325754688.0,
      "step": 2045
    },
    {
      "epoch": 1.0406917599186165,
      "grad_norm": 0.986159086227417,
      "learning_rate": 1e-05,
      "loss": 0.4598,
      "mean_token_accuracy": 0.8532018065452576,
      "num_tokens": 325898114.0,
      "step": 2046
    },
    {
      "epoch": 1.041200406917599,
      "grad_norm": 1.0895652770996094,
      "learning_rate": 1e-05,
      "loss": 0.5051,
      "mean_token_accuracy": 0.8412327766418457,
      "num_tokens": 326050746.0,
      "step": 2047
    },
    {
      "epoch": 1.0417090539165819,
      "grad_norm": 1.020145058631897,
      "learning_rate": 1e-05,
      "loss": 0.4913,
      "mean_token_accuracy": 0.8442976474761963,
      "num_tokens": 326210039.0,
      "step": 2048
    },
    {
      "epoch": 1.0422177009155646,
      "grad_norm": 1.036494493484497,
      "learning_rate": 1e-05,
      "loss": 0.5133,
      "mean_token_accuracy": 0.8386721014976501,
      "num_tokens": 326372235.0,
      "step": 2049
    },
    {
      "epoch": 1.0427263479145472,
      "grad_norm": 1.0419977903366089,
      "learning_rate": 1e-05,
      "loss": 0.5094,
      "mean_token_accuracy": 0.840260922908783,
      "num_tokens": 326539486.0,
      "step": 2050
    },
    {
      "epoch": 1.04323499491353,
      "grad_norm": 1.0354567766189575,
      "learning_rate": 1e-05,
      "loss": 0.4931,
      "mean_token_accuracy": 0.8432134389877319,
      "num_tokens": 326697650.0,
      "step": 2051
    },
    {
      "epoch": 1.0437436419125128,
      "grad_norm": 1.0431313514709473,
      "learning_rate": 1e-05,
      "loss": 0.4811,
      "mean_token_accuracy": 0.8483883142471313,
      "num_tokens": 326855696.0,
      "step": 2052
    },
    {
      "epoch": 1.0442522889114954,
      "grad_norm": 1.0206801891326904,
      "learning_rate": 1e-05,
      "loss": 0.5135,
      "mean_token_accuracy": 0.8393006324768066,
      "num_tokens": 327015415.0,
      "step": 2053
    },
    {
      "epoch": 1.0447609359104781,
      "grad_norm": 1.0079280138015747,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8453682065010071,
      "num_tokens": 327187580.0,
      "step": 2054
    },
    {
      "epoch": 1.045269582909461,
      "grad_norm": 1.0278279781341553,
      "learning_rate": 1e-05,
      "loss": 0.4992,
      "mean_token_accuracy": 0.8421951532363892,
      "num_tokens": 327357070.0,
      "step": 2055
    },
    {
      "epoch": 1.0457782299084435,
      "grad_norm": 1.018744707107544,
      "learning_rate": 1e-05,
      "loss": 0.5024,
      "mean_token_accuracy": 0.8416374921798706,
      "num_tokens": 327507342.0,
      "step": 2056
    },
    {
      "epoch": 1.0462868769074263,
      "grad_norm": 1.0134575366973877,
      "learning_rate": 1e-05,
      "loss": 0.4886,
      "mean_token_accuracy": 0.8452538251876831,
      "num_tokens": 327667011.0,
      "step": 2057
    },
    {
      "epoch": 1.0467955239064088,
      "grad_norm": 1.0265029668807983,
      "learning_rate": 1e-05,
      "loss": 0.508,
      "mean_token_accuracy": 0.8401978015899658,
      "num_tokens": 327836633.0,
      "step": 2058
    },
    {
      "epoch": 1.0473041709053916,
      "grad_norm": 1.0409969091415405,
      "learning_rate": 1e-05,
      "loss": 0.4697,
      "mean_token_accuracy": 0.8507318496704102,
      "num_tokens": 327973733.0,
      "step": 2059
    },
    {
      "epoch": 1.0478128179043744,
      "grad_norm": 1.0191627740859985,
      "learning_rate": 1e-05,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8423019647598267,
      "num_tokens": 328133022.0,
      "step": 2060
    },
    {
      "epoch": 1.048321464903357,
      "grad_norm": 1.214523196220398,
      "learning_rate": 1e-05,
      "loss": 0.5134,
      "mean_token_accuracy": 0.8397959470748901,
      "num_tokens": 328277969.0,
      "step": 2061
    },
    {
      "epoch": 1.0488301119023398,
      "grad_norm": 1.123297095298767,
      "learning_rate": 1e-05,
      "loss": 0.4865,
      "mean_token_accuracy": 0.8451673984527588,
      "num_tokens": 328442910.0,
      "step": 2062
    },
    {
      "epoch": 1.0493387589013226,
      "grad_norm": 1.0316575765609741,
      "learning_rate": 1e-05,
      "loss": 0.4745,
      "mean_token_accuracy": 0.850106954574585,
      "num_tokens": 328593481.0,
      "step": 2063
    },
    {
      "epoch": 1.0498474059003051,
      "grad_norm": 1.0077111721038818,
      "learning_rate": 1e-05,
      "loss": 0.5053,
      "mean_token_accuracy": 0.8421992659568787,
      "num_tokens": 328766960.0,
      "step": 2064
    },
    {
      "epoch": 1.050356052899288,
      "grad_norm": 1.0166810750961304,
      "learning_rate": 1e-05,
      "loss": 0.5074,
      "mean_token_accuracy": 0.8405725955963135,
      "num_tokens": 328919144.0,
      "step": 2065
    },
    {
      "epoch": 1.0508646998982707,
      "grad_norm": 1.0643290281295776,
      "learning_rate": 1e-05,
      "loss": 0.5067,
      "mean_token_accuracy": 0.8404885530471802,
      "num_tokens": 329079757.0,
      "step": 2066
    },
    {
      "epoch": 1.0513733468972533,
      "grad_norm": 0.9662615656852722,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.8407886028289795,
      "num_tokens": 329239039.0,
      "step": 2067
    },
    {
      "epoch": 1.051881993896236,
      "grad_norm": 0.9852247834205627,
      "learning_rate": 1e-05,
      "loss": 0.5069,
      "mean_token_accuracy": 0.8418384194374084,
      "num_tokens": 329396255.0,
      "step": 2068
    },
    {
      "epoch": 1.0523906408952186,
      "grad_norm": 1.0212758779525757,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8585224151611328,
      "num_tokens": 329542301.0,
      "step": 2069
    },
    {
      "epoch": 1.0528992878942014,
      "grad_norm": 1.0381677150726318,
      "learning_rate": 1e-05,
      "loss": 0.4746,
      "mean_token_accuracy": 0.8479477763175964,
      "num_tokens": 329691299.0,
      "step": 2070
    },
    {
      "epoch": 1.0534079348931842,
      "grad_norm": 0.9904256463050842,
      "learning_rate": 1e-05,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8424674272537231,
      "num_tokens": 329860490.0,
      "step": 2071
    },
    {
      "epoch": 1.0539165818921667,
      "grad_norm": 1.053727388381958,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8522511720657349,
      "num_tokens": 330014371.0,
      "step": 2072
    },
    {
      "epoch": 1.0544252288911495,
      "grad_norm": 0.9747759699821472,
      "learning_rate": 1e-05,
      "loss": 0.4756,
      "mean_token_accuracy": 0.8489809036254883,
      "num_tokens": 330193372.0,
      "step": 2073
    },
    {
      "epoch": 1.0549338758901323,
      "grad_norm": 1.0386484861373901,
      "learning_rate": 1e-05,
      "loss": 0.5115,
      "mean_token_accuracy": 0.840887188911438,
      "num_tokens": 330363604.0,
      "step": 2074
    },
    {
      "epoch": 1.0554425228891149,
      "grad_norm": 1.0518840551376343,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8496643304824829,
      "num_tokens": 330524103.0,
      "step": 2075
    },
    {
      "epoch": 1.0559511698880977,
      "grad_norm": 1.0117826461791992,
      "learning_rate": 1e-05,
      "loss": 0.4649,
      "mean_token_accuracy": 0.8528888821601868,
      "num_tokens": 330698211.0,
      "step": 2076
    },
    {
      "epoch": 1.0564598168870805,
      "grad_norm": 1.1235733032226562,
      "learning_rate": 1e-05,
      "loss": 0.5136,
      "mean_token_accuracy": 0.8376957178115845,
      "num_tokens": 330868485.0,
      "step": 2077
    },
    {
      "epoch": 1.056968463886063,
      "grad_norm": 1.1107542514801025,
      "learning_rate": 1e-05,
      "loss": 0.4661,
      "mean_token_accuracy": 0.8518024682998657,
      "num_tokens": 331039546.0,
      "step": 2078
    },
    {
      "epoch": 1.0574771108850458,
      "grad_norm": 1.0027269124984741,
      "learning_rate": 1e-05,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8440948724746704,
      "num_tokens": 331203155.0,
      "step": 2079
    },
    {
      "epoch": 1.0579857578840284,
      "grad_norm": 1.0469239950180054,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8530600070953369,
      "num_tokens": 331349908.0,
      "step": 2080
    },
    {
      "epoch": 1.0584944048830112,
      "grad_norm": 0.9495140910148621,
      "learning_rate": 1e-05,
      "loss": 0.4945,
      "mean_token_accuracy": 0.8430256843566895,
      "num_tokens": 331518516.0,
      "step": 2081
    },
    {
      "epoch": 1.059003051881994,
      "grad_norm": 0.972150444984436,
      "learning_rate": 1e-05,
      "loss": 0.494,
      "mean_token_accuracy": 0.8443804383277893,
      "num_tokens": 331685842.0,
      "step": 2082
    },
    {
      "epoch": 1.0595116988809765,
      "grad_norm": 1.0417505502700806,
      "learning_rate": 1e-05,
      "loss": 0.4905,
      "mean_token_accuracy": 0.8454884886741638,
      "num_tokens": 331844766.0,
      "step": 2083
    },
    {
      "epoch": 1.0600203458799593,
      "grad_norm": 1.0959299802780151,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8524262309074402,
      "num_tokens": 332002207.0,
      "step": 2084
    },
    {
      "epoch": 1.060528992878942,
      "grad_norm": 0.9652280807495117,
      "learning_rate": 1e-05,
      "loss": 0.4804,
      "mean_token_accuracy": 0.8485095500946045,
      "num_tokens": 332156850.0,
      "step": 2085
    },
    {
      "epoch": 1.0610376398779247,
      "grad_norm": 1.0024477243423462,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8554891347885132,
      "num_tokens": 332313613.0,
      "step": 2086
    },
    {
      "epoch": 1.0615462868769074,
      "grad_norm": 1.0409293174743652,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8442890644073486,
      "num_tokens": 332479167.0,
      "step": 2087
    },
    {
      "epoch": 1.0620549338758902,
      "grad_norm": 0.9680423140525818,
      "learning_rate": 1e-05,
      "loss": 0.4787,
      "mean_token_accuracy": 0.8484407663345337,
      "num_tokens": 332638608.0,
      "step": 2088
    },
    {
      "epoch": 1.0625635808748728,
      "grad_norm": 0.9663770794868469,
      "learning_rate": 1e-05,
      "loss": 0.4992,
      "mean_token_accuracy": 0.8420337438583374,
      "num_tokens": 332793827.0,
      "step": 2089
    },
    {
      "epoch": 1.0630722278738556,
      "grad_norm": 0.9979959726333618,
      "learning_rate": 1e-05,
      "loss": 0.4556,
      "mean_token_accuracy": 0.8552628755569458,
      "num_tokens": 332951297.0,
      "step": 2090
    },
    {
      "epoch": 1.0635808748728381,
      "grad_norm": 1.112915277481079,
      "learning_rate": 1e-05,
      "loss": 0.5213,
      "mean_token_accuracy": 0.8363105654716492,
      "num_tokens": 333110022.0,
      "step": 2091
    },
    {
      "epoch": 1.064089521871821,
      "grad_norm": 0.9754927158355713,
      "learning_rate": 1e-05,
      "loss": 0.4827,
      "mean_token_accuracy": 0.8483446836471558,
      "num_tokens": 333249289.0,
      "step": 2092
    },
    {
      "epoch": 1.0645981688708037,
      "grad_norm": 0.9905518889427185,
      "learning_rate": 1e-05,
      "loss": 0.4548,
      "mean_token_accuracy": 0.8558975458145142,
      "num_tokens": 333400414.0,
      "step": 2093
    },
    {
      "epoch": 1.0651068158697863,
      "grad_norm": 0.9375442266464233,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.856006920337677,
      "num_tokens": 333565896.0,
      "step": 2094
    },
    {
      "epoch": 1.065615462868769,
      "grad_norm": 0.9339344501495361,
      "learning_rate": 1e-05,
      "loss": 0.4827,
      "mean_token_accuracy": 0.848110020160675,
      "num_tokens": 333744235.0,
      "step": 2095
    },
    {
      "epoch": 1.0661241098677519,
      "grad_norm": 0.9189639091491699,
      "learning_rate": 1e-05,
      "loss": 0.5094,
      "mean_token_accuracy": 0.8405580520629883,
      "num_tokens": 333906023.0,
      "step": 2096
    },
    {
      "epoch": 1.0666327568667344,
      "grad_norm": 1.0412248373031616,
      "learning_rate": 1e-05,
      "loss": 0.4929,
      "mean_token_accuracy": 0.843265175819397,
      "num_tokens": 334070290.0,
      "step": 2097
    },
    {
      "epoch": 1.0671414038657172,
      "grad_norm": 0.9829212427139282,
      "learning_rate": 1e-05,
      "loss": 0.5228,
      "mean_token_accuracy": 0.836217999458313,
      "num_tokens": 334238798.0,
      "step": 2098
    },
    {
      "epoch": 1.0676500508647,
      "grad_norm": 1.2226958274841309,
      "learning_rate": 1e-05,
      "loss": 0.4986,
      "mean_token_accuracy": 0.8441588878631592,
      "num_tokens": 334405320.0,
      "step": 2099
    },
    {
      "epoch": 1.0681586978636826,
      "grad_norm": 1.1169629096984863,
      "learning_rate": 1e-05,
      "loss": 0.5335,
      "mean_token_accuracy": 0.8330006003379822,
      "num_tokens": 334559804.0,
      "step": 2100
    },
    {
      "epoch": 1.0686673448626653,
      "grad_norm": 1.0243042707443237,
      "learning_rate": 1e-05,
      "loss": 0.4874,
      "mean_token_accuracy": 0.8441094756126404,
      "num_tokens": 334706374.0,
      "step": 2101
    },
    {
      "epoch": 1.069175991861648,
      "grad_norm": 0.9952787160873413,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8492131233215332,
      "num_tokens": 334867946.0,
      "step": 2102
    },
    {
      "epoch": 1.0696846388606307,
      "grad_norm": 1.0360610485076904,
      "learning_rate": 1e-05,
      "loss": 0.4586,
      "mean_token_accuracy": 0.8524994254112244,
      "num_tokens": 335013479.0,
      "step": 2103
    },
    {
      "epoch": 1.0701932858596135,
      "grad_norm": 1.0457143783569336,
      "learning_rate": 1e-05,
      "loss": 0.5183,
      "mean_token_accuracy": 0.8378509283065796,
      "num_tokens": 335181587.0,
      "step": 2104
    },
    {
      "epoch": 1.070701932858596,
      "grad_norm": 1.0518393516540527,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.8486339449882507,
      "num_tokens": 335346515.0,
      "step": 2105
    },
    {
      "epoch": 1.0712105798575788,
      "grad_norm": 1.0435230731964111,
      "learning_rate": 1e-05,
      "loss": 0.5007,
      "mean_token_accuracy": 0.8433654308319092,
      "num_tokens": 335509021.0,
      "step": 2106
    },
    {
      "epoch": 1.0717192268565616,
      "grad_norm": 1.1404730081558228,
      "learning_rate": 1e-05,
      "loss": 0.5184,
      "mean_token_accuracy": 0.8369382619857788,
      "num_tokens": 335659431.0,
      "step": 2107
    },
    {
      "epoch": 1.0722278738555442,
      "grad_norm": 1.0483952760696411,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.8552796840667725,
      "num_tokens": 335809158.0,
      "step": 2108
    },
    {
      "epoch": 1.072736520854527,
      "grad_norm": 1.113566279411316,
      "learning_rate": 1e-05,
      "loss": 0.5126,
      "mean_token_accuracy": 0.8386800289154053,
      "num_tokens": 335959650.0,
      "step": 2109
    },
    {
      "epoch": 1.0732451678535098,
      "grad_norm": 0.9882280826568604,
      "learning_rate": 1e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.8503919839859009,
      "num_tokens": 336119431.0,
      "step": 2110
    },
    {
      "epoch": 1.0737538148524923,
      "grad_norm": 0.9893407225608826,
      "learning_rate": 1e-05,
      "loss": 0.4846,
      "mean_token_accuracy": 0.8462191820144653,
      "num_tokens": 336280307.0,
      "step": 2111
    },
    {
      "epoch": 1.0742624618514751,
      "grad_norm": 1.079674482345581,
      "learning_rate": 1e-05,
      "loss": 0.5072,
      "mean_token_accuracy": 0.8399666547775269,
      "num_tokens": 336442944.0,
      "step": 2112
    },
    {
      "epoch": 1.0747711088504577,
      "grad_norm": 0.9803399443626404,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.8556622862815857,
      "num_tokens": 336583703.0,
      "step": 2113
    },
    {
      "epoch": 1.0752797558494405,
      "grad_norm": 1.0098974704742432,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8502632975578308,
      "num_tokens": 336743733.0,
      "step": 2114
    },
    {
      "epoch": 1.0757884028484233,
      "grad_norm": 0.922788143157959,
      "learning_rate": 1e-05,
      "loss": 0.5002,
      "mean_token_accuracy": 0.8431379199028015,
      "num_tokens": 336923035.0,
      "step": 2115
    },
    {
      "epoch": 1.0762970498474058,
      "grad_norm": 1.0645747184753418,
      "learning_rate": 1e-05,
      "loss": 0.5146,
      "mean_token_accuracy": 0.8396292328834534,
      "num_tokens": 337080544.0,
      "step": 2116
    },
    {
      "epoch": 1.0768056968463886,
      "grad_norm": 1.0139567852020264,
      "learning_rate": 1e-05,
      "loss": 0.5178,
      "mean_token_accuracy": 0.837389349937439,
      "num_tokens": 337234518.0,
      "step": 2117
    },
    {
      "epoch": 1.0773143438453714,
      "grad_norm": 1.058477759361267,
      "learning_rate": 1e-05,
      "loss": 0.4748,
      "mean_token_accuracy": 0.848781406879425,
      "num_tokens": 337392278.0,
      "step": 2118
    },
    {
      "epoch": 1.077822990844354,
      "grad_norm": 0.9809451699256897,
      "learning_rate": 1e-05,
      "loss": 0.457,
      "mean_token_accuracy": 0.8538768291473389,
      "num_tokens": 337552580.0,
      "step": 2119
    },
    {
      "epoch": 1.0783316378433367,
      "grad_norm": 0.9701200723648071,
      "learning_rate": 1e-05,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8502329587936401,
      "num_tokens": 337706603.0,
      "step": 2120
    },
    {
      "epoch": 1.0788402848423195,
      "grad_norm": 1.1232140064239502,
      "learning_rate": 1e-05,
      "loss": 0.4921,
      "mean_token_accuracy": 0.8441417813301086,
      "num_tokens": 337861542.0,
      "step": 2121
    },
    {
      "epoch": 1.079348931841302,
      "grad_norm": 0.986573338508606,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8506069183349609,
      "num_tokens": 338026298.0,
      "step": 2122
    },
    {
      "epoch": 1.0798575788402849,
      "grad_norm": 1.0073679685592651,
      "learning_rate": 1e-05,
      "loss": 0.4842,
      "mean_token_accuracy": 0.8481048345565796,
      "num_tokens": 338196691.0,
      "step": 2123
    },
    {
      "epoch": 1.0803662258392674,
      "grad_norm": 0.989485502243042,
      "learning_rate": 1e-05,
      "loss": 0.4596,
      "mean_token_accuracy": 0.8538649082183838,
      "num_tokens": 338365205.0,
      "step": 2124
    },
    {
      "epoch": 1.0808748728382502,
      "grad_norm": 0.9852368831634521,
      "learning_rate": 1e-05,
      "loss": 0.5001,
      "mean_token_accuracy": 0.8432480096817017,
      "num_tokens": 338521286.0,
      "step": 2125
    },
    {
      "epoch": 1.081383519837233,
      "grad_norm": 1.0921062231063843,
      "learning_rate": 1e-05,
      "loss": 0.4975,
      "mean_token_accuracy": 0.8422955274581909,
      "num_tokens": 338668185.0,
      "step": 2126
    },
    {
      "epoch": 1.0818921668362156,
      "grad_norm": 0.9980047941207886,
      "learning_rate": 1e-05,
      "loss": 0.4877,
      "mean_token_accuracy": 0.8452361822128296,
      "num_tokens": 338835626.0,
      "step": 2127
    },
    {
      "epoch": 1.0824008138351984,
      "grad_norm": 0.9752321839332581,
      "learning_rate": 1e-05,
      "loss": 0.4943,
      "mean_token_accuracy": 0.8446649312973022,
      "num_tokens": 338990507.0,
      "step": 2128
    },
    {
      "epoch": 1.0829094608341812,
      "grad_norm": 1.0377464294433594,
      "learning_rate": 1e-05,
      "loss": 0.4939,
      "mean_token_accuracy": 0.8445258140563965,
      "num_tokens": 339154124.0,
      "step": 2129
    },
    {
      "epoch": 1.0834181078331637,
      "grad_norm": 0.9970061779022217,
      "learning_rate": 1e-05,
      "loss": 0.4862,
      "mean_token_accuracy": 0.8468868136405945,
      "num_tokens": 339326912.0,
      "step": 2130
    },
    {
      "epoch": 1.0839267548321465,
      "grad_norm": 1.0732330083847046,
      "learning_rate": 1e-05,
      "loss": 0.4909,
      "mean_token_accuracy": 0.8445082306861877,
      "num_tokens": 339475749.0,
      "step": 2131
    },
    {
      "epoch": 1.0844354018311293,
      "grad_norm": 1.000402808189392,
      "learning_rate": 1e-05,
      "loss": 0.4994,
      "mean_token_accuracy": 0.8430907726287842,
      "num_tokens": 339637655.0,
      "step": 2132
    },
    {
      "epoch": 1.0849440488301119,
      "grad_norm": 1.0055508613586426,
      "learning_rate": 1e-05,
      "loss": 0.4603,
      "mean_token_accuracy": 0.8536675572395325,
      "num_tokens": 339791196.0,
      "step": 2133
    },
    {
      "epoch": 1.0854526958290946,
      "grad_norm": 0.9924202561378479,
      "learning_rate": 1e-05,
      "loss": 0.4646,
      "mean_token_accuracy": 0.8522224426269531,
      "num_tokens": 339944200.0,
      "step": 2134
    },
    {
      "epoch": 1.0859613428280772,
      "grad_norm": 1.0728756189346313,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.846017599105835,
      "num_tokens": 340097746.0,
      "step": 2135
    },
    {
      "epoch": 1.08646998982706,
      "grad_norm": 0.9821423888206482,
      "learning_rate": 1e-05,
      "loss": 0.5064,
      "mean_token_accuracy": 0.8397558927536011,
      "num_tokens": 340252788.0,
      "step": 2136
    },
    {
      "epoch": 1.0869786368260428,
      "grad_norm": 0.962029755115509,
      "learning_rate": 1e-05,
      "loss": 0.4966,
      "mean_token_accuracy": 0.8433777093887329,
      "num_tokens": 340418368.0,
      "step": 2137
    },
    {
      "epoch": 1.0874872838250254,
      "grad_norm": 0.9940122365951538,
      "learning_rate": 1e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8478418588638306,
      "num_tokens": 340571484.0,
      "step": 2138
    },
    {
      "epoch": 1.0879959308240081,
      "grad_norm": 1.0245951414108276,
      "learning_rate": 1e-05,
      "loss": 0.4901,
      "mean_token_accuracy": 0.8461571931838989,
      "num_tokens": 340720288.0,
      "step": 2139
    },
    {
      "epoch": 1.088504577822991,
      "grad_norm": 0.9604802131652832,
      "learning_rate": 1e-05,
      "loss": 0.4796,
      "mean_token_accuracy": 0.8480315208435059,
      "num_tokens": 340886244.0,
      "step": 2140
    },
    {
      "epoch": 1.0890132248219735,
      "grad_norm": 1.0930918455123901,
      "learning_rate": 1e-05,
      "loss": 0.4966,
      "mean_token_accuracy": 0.8429490327835083,
      "num_tokens": 341039977.0,
      "step": 2141
    },
    {
      "epoch": 1.0895218718209563,
      "grad_norm": 1.1034718751907349,
      "learning_rate": 1e-05,
      "loss": 0.4837,
      "mean_token_accuracy": 0.8489624261856079,
      "num_tokens": 341199202.0,
      "step": 2142
    },
    {
      "epoch": 1.090030518819939,
      "grad_norm": 0.9594838619232178,
      "learning_rate": 1e-05,
      "loss": 0.4877,
      "mean_token_accuracy": 0.8449561595916748,
      "num_tokens": 341365619.0,
      "step": 2143
    },
    {
      "epoch": 1.0905391658189216,
      "grad_norm": 0.9820337295532227,
      "learning_rate": 1e-05,
      "loss": 0.5019,
      "mean_token_accuracy": 0.8415205478668213,
      "num_tokens": 341532738.0,
      "step": 2144
    },
    {
      "epoch": 1.0910478128179044,
      "grad_norm": 0.9692781567573547,
      "learning_rate": 1e-05,
      "loss": 0.49,
      "mean_token_accuracy": 0.8459674715995789,
      "num_tokens": 341683366.0,
      "step": 2145
    },
    {
      "epoch": 1.091556459816887,
      "grad_norm": 1.0075808763504028,
      "learning_rate": 1e-05,
      "loss": 0.4857,
      "mean_token_accuracy": 0.8459229469299316,
      "num_tokens": 341847416.0,
      "step": 2146
    },
    {
      "epoch": 1.0920651068158698,
      "grad_norm": 0.9658797383308411,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8468917608261108,
      "num_tokens": 342011022.0,
      "step": 2147
    },
    {
      "epoch": 1.0925737538148526,
      "grad_norm": 0.9919976592063904,
      "learning_rate": 1e-05,
      "loss": 0.4787,
      "mean_token_accuracy": 0.8484220504760742,
      "num_tokens": 342175374.0,
      "step": 2148
    },
    {
      "epoch": 1.0930824008138351,
      "grad_norm": 0.9361450672149658,
      "learning_rate": 1e-05,
      "loss": 0.5137,
      "mean_token_accuracy": 0.8397796750068665,
      "num_tokens": 342344134.0,
      "step": 2149
    },
    {
      "epoch": 1.093591047812818,
      "grad_norm": 0.961454451084137,
      "learning_rate": 1e-05,
      "loss": 0.5006,
      "mean_token_accuracy": 0.8421031832695007,
      "num_tokens": 342506392.0,
      "step": 2150
    },
    {
      "epoch": 1.0940996948118007,
      "grad_norm": 1.0065735578536987,
      "learning_rate": 1e-05,
      "loss": 0.4779,
      "mean_token_accuracy": 0.8483932018280029,
      "num_tokens": 342671330.0,
      "step": 2151
    },
    {
      "epoch": 1.0946083418107833,
      "grad_norm": 0.9278548955917358,
      "learning_rate": 1e-05,
      "loss": 0.5141,
      "mean_token_accuracy": 0.8382604718208313,
      "num_tokens": 342830993.0,
      "step": 2152
    },
    {
      "epoch": 1.095116988809766,
      "grad_norm": 1.0080245733261108,
      "learning_rate": 1e-05,
      "loss": 0.4944,
      "mean_token_accuracy": 0.843571126461029,
      "num_tokens": 342983161.0,
      "step": 2153
    },
    {
      "epoch": 1.0956256358087488,
      "grad_norm": 1.0011775493621826,
      "learning_rate": 1e-05,
      "loss": 0.4865,
      "mean_token_accuracy": 0.8465944528579712,
      "num_tokens": 343141065.0,
      "step": 2154
    },
    {
      "epoch": 1.0961342828077314,
      "grad_norm": 1.0971297025680542,
      "learning_rate": 1e-05,
      "loss": 0.5232,
      "mean_token_accuracy": 0.8360178470611572,
      "num_tokens": 343293439.0,
      "step": 2155
    },
    {
      "epoch": 1.0966429298067142,
      "grad_norm": 0.9907771944999695,
      "learning_rate": 1e-05,
      "loss": 0.5133,
      "mean_token_accuracy": 0.8384357690811157,
      "num_tokens": 343447392.0,
      "step": 2156
    },
    {
      "epoch": 1.0971515768056967,
      "grad_norm": 1.0306109189987183,
      "learning_rate": 1e-05,
      "loss": 0.4763,
      "mean_token_accuracy": 0.8493988513946533,
      "num_tokens": 343594086.0,
      "step": 2157
    },
    {
      "epoch": 1.0976602238046795,
      "grad_norm": 0.956362783908844,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8483714461326599,
      "num_tokens": 343757619.0,
      "step": 2158
    },
    {
      "epoch": 1.0981688708036623,
      "grad_norm": 0.9461917281150818,
      "learning_rate": 1e-05,
      "loss": 0.4671,
      "mean_token_accuracy": 0.852462887763977,
      "num_tokens": 343918481.0,
      "step": 2159
    },
    {
      "epoch": 1.0986775178026449,
      "grad_norm": 1.0135009288787842,
      "learning_rate": 1e-05,
      "loss": 0.5046,
      "mean_token_accuracy": 0.8400470018386841,
      "num_tokens": 344073814.0,
      "step": 2160
    },
    {
      "epoch": 1.0991861648016277,
      "grad_norm": 0.9641030430793762,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8498222231864929,
      "num_tokens": 344234173.0,
      "step": 2161
    },
    {
      "epoch": 1.0996948118006105,
      "grad_norm": 0.9917462468147278,
      "learning_rate": 1e-05,
      "loss": 0.4966,
      "mean_token_accuracy": 0.8430187702178955,
      "num_tokens": 344391441.0,
      "step": 2162
    },
    {
      "epoch": 1.100203458799593,
      "grad_norm": 0.9825915694236755,
      "learning_rate": 1e-05,
      "loss": 0.4846,
      "mean_token_accuracy": 0.8467402458190918,
      "num_tokens": 344548813.0,
      "step": 2163
    },
    {
      "epoch": 1.1007121057985758,
      "grad_norm": 1.006765365600586,
      "learning_rate": 1e-05,
      "loss": 0.5063,
      "mean_token_accuracy": 0.8426176309585571,
      "num_tokens": 344712984.0,
      "step": 2164
    },
    {
      "epoch": 1.1012207527975586,
      "grad_norm": 0.9520547389984131,
      "learning_rate": 1e-05,
      "loss": 0.4933,
      "mean_token_accuracy": 0.8418510556221008,
      "num_tokens": 344873127.0,
      "step": 2165
    },
    {
      "epoch": 1.1017293997965412,
      "grad_norm": 0.9368497133255005,
      "learning_rate": 1e-05,
      "loss": 0.5136,
      "mean_token_accuracy": 0.8412468433380127,
      "num_tokens": 345036641.0,
      "step": 2166
    },
    {
      "epoch": 1.102238046795524,
      "grad_norm": 1.083411693572998,
      "learning_rate": 1e-05,
      "loss": 0.4849,
      "mean_token_accuracy": 0.8463590145111084,
      "num_tokens": 345205810.0,
      "step": 2167
    },
    {
      "epoch": 1.1027466937945065,
      "grad_norm": 0.988730251789093,
      "learning_rate": 1e-05,
      "loss": 0.491,
      "mean_token_accuracy": 0.8430241942405701,
      "num_tokens": 345369577.0,
      "step": 2168
    },
    {
      "epoch": 1.1032553407934893,
      "grad_norm": 1.005668044090271,
      "learning_rate": 1e-05,
      "loss": 0.4628,
      "mean_token_accuracy": 0.8519200682640076,
      "num_tokens": 345522712.0,
      "step": 2169
    },
    {
      "epoch": 1.103763987792472,
      "grad_norm": 0.9839890599250793,
      "learning_rate": 1e-05,
      "loss": 0.4813,
      "mean_token_accuracy": 0.8477115631103516,
      "num_tokens": 345683367.0,
      "step": 2170
    },
    {
      "epoch": 1.1042726347914547,
      "grad_norm": 1.0047062635421753,
      "learning_rate": 1e-05,
      "loss": 0.4924,
      "mean_token_accuracy": 0.8448252081871033,
      "num_tokens": 345836341.0,
      "step": 2171
    },
    {
      "epoch": 1.1047812817904374,
      "grad_norm": 0.9978728294372559,
      "learning_rate": 1e-05,
      "loss": 0.5228,
      "mean_token_accuracy": 0.8382081985473633,
      "num_tokens": 346003193.0,
      "step": 2172
    },
    {
      "epoch": 1.1052899287894202,
      "grad_norm": 0.9945022463798523,
      "learning_rate": 1e-05,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8414233326911926,
      "num_tokens": 346173134.0,
      "step": 2173
    },
    {
      "epoch": 1.1057985757884028,
      "grad_norm": 0.9274640679359436,
      "learning_rate": 1e-05,
      "loss": 0.4744,
      "mean_token_accuracy": 0.8477193713188171,
      "num_tokens": 346330561.0,
      "step": 2174
    },
    {
      "epoch": 1.1063072227873856,
      "grad_norm": 0.9226838946342468,
      "learning_rate": 1e-05,
      "loss": 0.502,
      "mean_token_accuracy": 0.8429021239280701,
      "num_tokens": 346499119.0,
      "step": 2175
    },
    {
      "epoch": 1.1068158697863684,
      "grad_norm": 0.9797883033752441,
      "learning_rate": 1e-05,
      "loss": 0.4945,
      "mean_token_accuracy": 0.8447971940040588,
      "num_tokens": 346660957.0,
      "step": 2176
    },
    {
      "epoch": 1.107324516785351,
      "grad_norm": 0.9634705781936646,
      "learning_rate": 1e-05,
      "loss": 0.4471,
      "mean_token_accuracy": 0.85663241147995,
      "num_tokens": 346820626.0,
      "step": 2177
    },
    {
      "epoch": 1.1078331637843337,
      "grad_norm": 0.9565395712852478,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8473824858665466,
      "num_tokens": 346990623.0,
      "step": 2178
    },
    {
      "epoch": 1.1083418107833163,
      "grad_norm": 0.9224358797073364,
      "learning_rate": 1e-05,
      "loss": 0.4954,
      "mean_token_accuracy": 0.8449086546897888,
      "num_tokens": 347168826.0,
      "step": 2179
    },
    {
      "epoch": 1.108850457782299,
      "grad_norm": 0.9978235960006714,
      "learning_rate": 1e-05,
      "loss": 0.4954,
      "mean_token_accuracy": 0.8437415361404419,
      "num_tokens": 347324012.0,
      "step": 2180
    },
    {
      "epoch": 1.1093591047812819,
      "grad_norm": 1.0280753374099731,
      "learning_rate": 1e-05,
      "loss": 0.4929,
      "mean_token_accuracy": 0.8424971103668213,
      "num_tokens": 347478544.0,
      "step": 2181
    },
    {
      "epoch": 1.1098677517802644,
      "grad_norm": 0.9893693923950195,
      "learning_rate": 1e-05,
      "loss": 0.4984,
      "mean_token_accuracy": 0.8421764373779297,
      "num_tokens": 347632344.0,
      "step": 2182
    },
    {
      "epoch": 1.1103763987792472,
      "grad_norm": 1.0173699855804443,
      "learning_rate": 1e-05,
      "loss": 0.5012,
      "mean_token_accuracy": 0.8403029441833496,
      "num_tokens": 347788398.0,
      "step": 2183
    },
    {
      "epoch": 1.11088504577823,
      "grad_norm": 1.399958610534668,
      "learning_rate": 1e-05,
      "loss": 0.4959,
      "mean_token_accuracy": 0.840847909450531,
      "num_tokens": 347925739.0,
      "step": 2184
    },
    {
      "epoch": 1.1113936927772126,
      "grad_norm": 1.0244845151901245,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.854158341884613,
      "num_tokens": 348085226.0,
      "step": 2185
    },
    {
      "epoch": 1.1119023397761953,
      "grad_norm": 0.9952741265296936,
      "learning_rate": 1e-05,
      "loss": 0.4837,
      "mean_token_accuracy": 0.8468140959739685,
      "num_tokens": 348240499.0,
      "step": 2186
    },
    {
      "epoch": 1.1124109867751781,
      "grad_norm": 1.0770624876022339,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.8523979783058167,
      "num_tokens": 348394436.0,
      "step": 2187
    },
    {
      "epoch": 1.1129196337741607,
      "grad_norm": 0.97931307554245,
      "learning_rate": 1e-05,
      "loss": 0.4806,
      "mean_token_accuracy": 0.8491225838661194,
      "num_tokens": 348555987.0,
      "step": 2188
    },
    {
      "epoch": 1.1134282807731435,
      "grad_norm": 1.1255407333374023,
      "learning_rate": 1e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8461816310882568,
      "num_tokens": 348708430.0,
      "step": 2189
    },
    {
      "epoch": 1.113936927772126,
      "grad_norm": 1.0181074142456055,
      "learning_rate": 1e-05,
      "loss": 0.4883,
      "mean_token_accuracy": 0.8455284833908081,
      "num_tokens": 348866347.0,
      "step": 2190
    },
    {
      "epoch": 1.1144455747711088,
      "grad_norm": 0.957888662815094,
      "learning_rate": 1e-05,
      "loss": 0.4735,
      "mean_token_accuracy": 0.8496738076210022,
      "num_tokens": 349035863.0,
      "step": 2191
    },
    {
      "epoch": 1.1149542217700916,
      "grad_norm": 1.0182090997695923,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8527451753616333,
      "num_tokens": 349194754.0,
      "step": 2192
    },
    {
      "epoch": 1.1154628687690742,
      "grad_norm": 1.0608559846878052,
      "learning_rate": 1e-05,
      "loss": 0.4881,
      "mean_token_accuracy": 0.8453966975212097,
      "num_tokens": 349350492.0,
      "step": 2193
    },
    {
      "epoch": 1.115971515768057,
      "grad_norm": 1.0267398357391357,
      "learning_rate": 1e-05,
      "loss": 0.471,
      "mean_token_accuracy": 0.8491342067718506,
      "num_tokens": 349501086.0,
      "step": 2194
    },
    {
      "epoch": 1.1164801627670398,
      "grad_norm": 0.9610785245895386,
      "learning_rate": 1e-05,
      "loss": 0.4651,
      "mean_token_accuracy": 0.8535009622573853,
      "num_tokens": 349655643.0,
      "step": 2195
    },
    {
      "epoch": 1.1169888097660223,
      "grad_norm": 0.9949964880943298,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.844301164150238,
      "num_tokens": 349818869.0,
      "step": 2196
    },
    {
      "epoch": 1.1174974567650051,
      "grad_norm": 0.9660967588424683,
      "learning_rate": 1e-05,
      "loss": 0.468,
      "mean_token_accuracy": 0.850468635559082,
      "num_tokens": 349975870.0,
      "step": 2197
    },
    {
      "epoch": 1.118006103763988,
      "grad_norm": 1.0754655599594116,
      "learning_rate": 1e-05,
      "loss": 0.4859,
      "mean_token_accuracy": 0.8464107513427734,
      "num_tokens": 350142280.0,
      "step": 2198
    },
    {
      "epoch": 1.1185147507629705,
      "grad_norm": 1.0086798667907715,
      "learning_rate": 1e-05,
      "loss": 0.4557,
      "mean_token_accuracy": 0.855668842792511,
      "num_tokens": 350299480.0,
      "step": 2199
    },
    {
      "epoch": 1.1190233977619533,
      "grad_norm": 0.9509584307670593,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.8424690365791321,
      "num_tokens": 350469513.0,
      "step": 2200
    },
    {
      "epoch": 1.1195320447609358,
      "grad_norm": 1.24341881275177,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8451713919639587,
      "num_tokens": 350617118.0,
      "step": 2201
    },
    {
      "epoch": 1.1200406917599186,
      "grad_norm": 0.9689223766326904,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8523330688476562,
      "num_tokens": 350784696.0,
      "step": 2202
    },
    {
      "epoch": 1.1205493387589014,
      "grad_norm": 0.9255483746528625,
      "learning_rate": 1e-05,
      "loss": 0.4559,
      "mean_token_accuracy": 0.8544161915779114,
      "num_tokens": 350943249.0,
      "step": 2203
    },
    {
      "epoch": 1.121057985757884,
      "grad_norm": 1.0668855905532837,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8542884588241577,
      "num_tokens": 351109045.0,
      "step": 2204
    },
    {
      "epoch": 1.1215666327568667,
      "grad_norm": 0.9399425387382507,
      "learning_rate": 1e-05,
      "loss": 0.4764,
      "mean_token_accuracy": 0.8503850698471069,
      "num_tokens": 351282282.0,
      "step": 2205
    },
    {
      "epoch": 1.1220752797558495,
      "grad_norm": 1.0401631593704224,
      "learning_rate": 1e-05,
      "loss": 0.4641,
      "mean_token_accuracy": 0.852576494216919,
      "num_tokens": 351433256.0,
      "step": 2206
    },
    {
      "epoch": 1.122583926754832,
      "grad_norm": 1.0747698545455933,
      "learning_rate": 1e-05,
      "loss": 0.511,
      "mean_token_accuracy": 0.8402716517448425,
      "num_tokens": 351581777.0,
      "step": 2207
    },
    {
      "epoch": 1.1230925737538149,
      "grad_norm": 1.0110957622528076,
      "learning_rate": 1e-05,
      "loss": 0.4989,
      "mean_token_accuracy": 0.8414695262908936,
      "num_tokens": 351734672.0,
      "step": 2208
    },
    {
      "epoch": 1.1236012207527977,
      "grad_norm": 1.0141860246658325,
      "learning_rate": 1e-05,
      "loss": 0.4772,
      "mean_token_accuracy": 0.84971022605896,
      "num_tokens": 351891165.0,
      "step": 2209
    },
    {
      "epoch": 1.1241098677517802,
      "grad_norm": 0.9789010882377625,
      "learning_rate": 1e-05,
      "loss": 0.4811,
      "mean_token_accuracy": 0.8458177447319031,
      "num_tokens": 352042249.0,
      "step": 2210
    },
    {
      "epoch": 1.124618514750763,
      "grad_norm": 1.061063528060913,
      "learning_rate": 1e-05,
      "loss": 0.4866,
      "mean_token_accuracy": 0.844889760017395,
      "num_tokens": 352204407.0,
      "step": 2211
    },
    {
      "epoch": 1.1251271617497456,
      "grad_norm": 0.9890654683113098,
      "learning_rate": 1e-05,
      "loss": 0.4935,
      "mean_token_accuracy": 0.8418614268302917,
      "num_tokens": 352358733.0,
      "step": 2212
    },
    {
      "epoch": 1.1256358087487284,
      "grad_norm": 0.9955906867980957,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8459783792495728,
      "num_tokens": 352511222.0,
      "step": 2213
    },
    {
      "epoch": 1.1261444557477112,
      "grad_norm": 0.9302468299865723,
      "learning_rate": 1e-05,
      "loss": 0.493,
      "mean_token_accuracy": 0.8450722694396973,
      "num_tokens": 352667221.0,
      "step": 2214
    },
    {
      "epoch": 1.1266531027466937,
      "grad_norm": 0.9680992960929871,
      "learning_rate": 1e-05,
      "loss": 0.4665,
      "mean_token_accuracy": 0.8513346910476685,
      "num_tokens": 352830778.0,
      "step": 2215
    },
    {
      "epoch": 1.1271617497456765,
      "grad_norm": 0.9598748683929443,
      "learning_rate": 1e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8466798663139343,
      "num_tokens": 352982322.0,
      "step": 2216
    },
    {
      "epoch": 1.1276703967446593,
      "grad_norm": 0.9731796979904175,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8480831980705261,
      "num_tokens": 353143982.0,
      "step": 2217
    },
    {
      "epoch": 1.1281790437436419,
      "grad_norm": 0.9734607338905334,
      "learning_rate": 1e-05,
      "loss": 0.49,
      "mean_token_accuracy": 0.8462680578231812,
      "num_tokens": 353294238.0,
      "step": 2218
    },
    {
      "epoch": 1.1286876907426246,
      "grad_norm": 0.9904110431671143,
      "learning_rate": 1e-05,
      "loss": 0.4869,
      "mean_token_accuracy": 0.8465762138366699,
      "num_tokens": 353451803.0,
      "step": 2219
    },
    {
      "epoch": 1.1291963377416074,
      "grad_norm": 1.043080449104309,
      "learning_rate": 1e-05,
      "loss": 0.5122,
      "mean_token_accuracy": 0.8410126566886902,
      "num_tokens": 353617488.0,
      "step": 2220
    },
    {
      "epoch": 1.12970498474059,
      "grad_norm": 0.9859865307807922,
      "learning_rate": 1e-05,
      "loss": 0.4842,
      "mean_token_accuracy": 0.8460755944252014,
      "num_tokens": 353774242.0,
      "step": 2221
    },
    {
      "epoch": 1.1302136317395728,
      "grad_norm": 1.063538908958435,
      "learning_rate": 1e-05,
      "loss": 0.4824,
      "mean_token_accuracy": 0.8463836908340454,
      "num_tokens": 353942973.0,
      "step": 2222
    },
    {
      "epoch": 1.1307222787385554,
      "grad_norm": 1.027944803237915,
      "learning_rate": 1e-05,
      "loss": 0.5363,
      "mean_token_accuracy": 0.8324622511863708,
      "num_tokens": 354091657.0,
      "step": 2223
    },
    {
      "epoch": 1.1312309257375381,
      "grad_norm": 1.0895392894744873,
      "learning_rate": 1e-05,
      "loss": 0.4823,
      "mean_token_accuracy": 0.8465158939361572,
      "num_tokens": 354241839.0,
      "step": 2224
    },
    {
      "epoch": 1.131739572736521,
      "grad_norm": 1.0151928663253784,
      "learning_rate": 1e-05,
      "loss": 0.4842,
      "mean_token_accuracy": 0.8473328351974487,
      "num_tokens": 354404721.0,
      "step": 2225
    },
    {
      "epoch": 1.1322482197355035,
      "grad_norm": 1.0319733619689941,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.8381280303001404,
      "num_tokens": 354575014.0,
      "step": 2226
    },
    {
      "epoch": 1.1327568667344863,
      "grad_norm": 1.0670835971832275,
      "learning_rate": 1e-05,
      "loss": 0.4944,
      "mean_token_accuracy": 0.8440436124801636,
      "num_tokens": 354737329.0,
      "step": 2227
    },
    {
      "epoch": 1.133265513733469,
      "grad_norm": 0.9908400774002075,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8480251431465149,
      "num_tokens": 354915712.0,
      "step": 2228
    },
    {
      "epoch": 1.1337741607324516,
      "grad_norm": 0.9610514640808105,
      "learning_rate": 1e-05,
      "loss": 0.5001,
      "mean_token_accuracy": 0.8409698605537415,
      "num_tokens": 355070597.0,
      "step": 2229
    },
    {
      "epoch": 1.1342828077314344,
      "grad_norm": 1.01914644241333,
      "learning_rate": 1e-05,
      "loss": 0.5113,
      "mean_token_accuracy": 0.8394330739974976,
      "num_tokens": 355237653.0,
      "step": 2230
    },
    {
      "epoch": 1.1347914547304172,
      "grad_norm": 0.9027879238128662,
      "learning_rate": 1e-05,
      "loss": 0.503,
      "mean_token_accuracy": 0.8422386050224304,
      "num_tokens": 355408779.0,
      "step": 2231
    },
    {
      "epoch": 1.1353001017293998,
      "grad_norm": 0.9325588941574097,
      "learning_rate": 1e-05,
      "loss": 0.4643,
      "mean_token_accuracy": 0.8536386489868164,
      "num_tokens": 355582832.0,
      "step": 2232
    },
    {
      "epoch": 1.1358087487283826,
      "grad_norm": 1.0098017454147339,
      "learning_rate": 1e-05,
      "loss": 0.4743,
      "mean_token_accuracy": 0.8489354252815247,
      "num_tokens": 355724880.0,
      "step": 2233
    },
    {
      "epoch": 1.1363173957273651,
      "grad_norm": 1.0128886699676514,
      "learning_rate": 1e-05,
      "loss": 0.5105,
      "mean_token_accuracy": 0.8378116488456726,
      "num_tokens": 355882190.0,
      "step": 2234
    },
    {
      "epoch": 1.136826042726348,
      "grad_norm": 0.9628869295120239,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8522527813911438,
      "num_tokens": 356027036.0,
      "step": 2235
    },
    {
      "epoch": 1.1373346897253307,
      "grad_norm": 0.9411493539810181,
      "learning_rate": 1e-05,
      "loss": 0.4819,
      "mean_token_accuracy": 0.8461790680885315,
      "num_tokens": 356192221.0,
      "step": 2236
    },
    {
      "epoch": 1.1378433367243133,
      "grad_norm": 0.9630653858184814,
      "learning_rate": 1e-05,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8419363498687744,
      "num_tokens": 356354570.0,
      "step": 2237
    },
    {
      "epoch": 1.138351983723296,
      "grad_norm": 1.036780595779419,
      "learning_rate": 1e-05,
      "loss": 0.5176,
      "mean_token_accuracy": 0.8373992443084717,
      "num_tokens": 356515011.0,
      "step": 2238
    },
    {
      "epoch": 1.1388606307222788,
      "grad_norm": 1.0112197399139404,
      "learning_rate": 1e-05,
      "loss": 0.4697,
      "mean_token_accuracy": 0.8511573076248169,
      "num_tokens": 356676989.0,
      "step": 2239
    },
    {
      "epoch": 1.1393692777212614,
      "grad_norm": 1.0344760417938232,
      "learning_rate": 1e-05,
      "loss": 0.4934,
      "mean_token_accuracy": 0.8441804051399231,
      "num_tokens": 356840072.0,
      "step": 2240
    },
    {
      "epoch": 1.1398779247202442,
      "grad_norm": 0.972714364528656,
      "learning_rate": 1e-05,
      "loss": 0.4925,
      "mean_token_accuracy": 0.8463292121887207,
      "num_tokens": 357003414.0,
      "step": 2241
    },
    {
      "epoch": 1.140386571719227,
      "grad_norm": 0.9765809774398804,
      "learning_rate": 1e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8423345685005188,
      "num_tokens": 357164387.0,
      "step": 2242
    },
    {
      "epoch": 1.1408952187182095,
      "grad_norm": 0.9447953104972839,
      "learning_rate": 1e-05,
      "loss": 0.4613,
      "mean_token_accuracy": 0.8525022864341736,
      "num_tokens": 357323400.0,
      "step": 2243
    },
    {
      "epoch": 1.1414038657171923,
      "grad_norm": 0.946564793586731,
      "learning_rate": 1e-05,
      "loss": 0.4763,
      "mean_token_accuracy": 0.8481206893920898,
      "num_tokens": 357480123.0,
      "step": 2244
    },
    {
      "epoch": 1.1419125127161749,
      "grad_norm": 1.0055655241012573,
      "learning_rate": 1e-05,
      "loss": 0.4601,
      "mean_token_accuracy": 0.8537294268608093,
      "num_tokens": 357642932.0,
      "step": 2245
    },
    {
      "epoch": 1.1424211597151577,
      "grad_norm": 0.9800072908401489,
      "learning_rate": 1e-05,
      "loss": 0.462,
      "mean_token_accuracy": 0.8519518375396729,
      "num_tokens": 357799511.0,
      "step": 2246
    },
    {
      "epoch": 1.1429298067141405,
      "grad_norm": 1.0585764646530151,
      "learning_rate": 1e-05,
      "loss": 0.4837,
      "mean_token_accuracy": 0.8486620187759399,
      "num_tokens": 357950084.0,
      "step": 2247
    },
    {
      "epoch": 1.143438453713123,
      "grad_norm": 1.0663877725601196,
      "learning_rate": 1e-05,
      "loss": 0.5124,
      "mean_token_accuracy": 0.8382563591003418,
      "num_tokens": 358095805.0,
      "step": 2248
    },
    {
      "epoch": 1.1439471007121058,
      "grad_norm": 1.0571013689041138,
      "learning_rate": 1e-05,
      "loss": 0.4997,
      "mean_token_accuracy": 0.8429567217826843,
      "num_tokens": 358247573.0,
      "step": 2249
    },
    {
      "epoch": 1.1444557477110886,
      "grad_norm": 0.9647451639175415,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8561835289001465,
      "num_tokens": 358405765.0,
      "step": 2250
    },
    {
      "epoch": 1.1449643947100712,
      "grad_norm": 0.9498294591903687,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.8418372869491577,
      "num_tokens": 358572605.0,
      "step": 2251
    },
    {
      "epoch": 1.145473041709054,
      "grad_norm": 1.0757380723953247,
      "learning_rate": 1e-05,
      "loss": 0.55,
      "mean_token_accuracy": 0.8287532925605774,
      "num_tokens": 358741998.0,
      "step": 2252
    },
    {
      "epoch": 1.1459816887080367,
      "grad_norm": 1.0514572858810425,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8518444299697876,
      "num_tokens": 358892427.0,
      "step": 2253
    },
    {
      "epoch": 1.1464903357070193,
      "grad_norm": 0.9669085741043091,
      "learning_rate": 1e-05,
      "loss": 0.4671,
      "mean_token_accuracy": 0.854065477848053,
      "num_tokens": 359053024.0,
      "step": 2254
    },
    {
      "epoch": 1.146998982706002,
      "grad_norm": 1.0323379039764404,
      "learning_rate": 1e-05,
      "loss": 0.4773,
      "mean_token_accuracy": 0.8478610515594482,
      "num_tokens": 359211001.0,
      "step": 2255
    },
    {
      "epoch": 1.1475076297049847,
      "grad_norm": 1.0277243852615356,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8460548520088196,
      "num_tokens": 359375453.0,
      "step": 2256
    },
    {
      "epoch": 1.1480162767039674,
      "grad_norm": 1.0732399225234985,
      "learning_rate": 1e-05,
      "loss": 0.461,
      "mean_token_accuracy": 0.8520042300224304,
      "num_tokens": 359521205.0,
      "step": 2257
    },
    {
      "epoch": 1.1485249237029502,
      "grad_norm": 1.0115880966186523,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.8529631495475769,
      "num_tokens": 359692458.0,
      "step": 2258
    },
    {
      "epoch": 1.1490335707019328,
      "grad_norm": 1.0270670652389526,
      "learning_rate": 1e-05,
      "loss": 0.4832,
      "mean_token_accuracy": 0.8482770919799805,
      "num_tokens": 359854056.0,
      "step": 2259
    },
    {
      "epoch": 1.1495422177009156,
      "grad_norm": 1.1131222248077393,
      "learning_rate": 1e-05,
      "loss": 0.5044,
      "mean_token_accuracy": 0.8396096229553223,
      "num_tokens": 360008843.0,
      "step": 2260
    },
    {
      "epoch": 1.1500508646998984,
      "grad_norm": 1.0628095865249634,
      "learning_rate": 1e-05,
      "loss": 0.4878,
      "mean_token_accuracy": 0.8451195955276489,
      "num_tokens": 360171252.0,
      "step": 2261
    },
    {
      "epoch": 1.150559511698881,
      "grad_norm": 0.9063826203346252,
      "learning_rate": 1e-05,
      "loss": 0.4578,
      "mean_token_accuracy": 0.8525300025939941,
      "num_tokens": 360333592.0,
      "step": 2262
    },
    {
      "epoch": 1.1510681586978637,
      "grad_norm": 1.0526858568191528,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8553949594497681,
      "num_tokens": 360485356.0,
      "step": 2263
    },
    {
      "epoch": 1.1515768056968465,
      "grad_norm": 1.6245720386505127,
      "learning_rate": 1e-05,
      "loss": 0.4326,
      "mean_token_accuracy": 0.8621503114700317,
      "num_tokens": 360646033.0,
      "step": 2264
    },
    {
      "epoch": 1.152085452695829,
      "grad_norm": 1.6461989879608154,
      "learning_rate": 1e-05,
      "loss": 0.5018,
      "mean_token_accuracy": 0.8413547277450562,
      "num_tokens": 360822067.0,
      "step": 2265
    },
    {
      "epoch": 1.1525940996948119,
      "grad_norm": 1.0477701425552368,
      "learning_rate": 1e-05,
      "loss": 0.5035,
      "mean_token_accuracy": 0.8408612012863159,
      "num_tokens": 360980106.0,
      "step": 2266
    },
    {
      "epoch": 1.1531027466937944,
      "grad_norm": 0.9780399799346924,
      "learning_rate": 1e-05,
      "loss": 0.477,
      "mean_token_accuracy": 0.8492373824119568,
      "num_tokens": 361141106.0,
      "step": 2267
    },
    {
      "epoch": 1.1536113936927772,
      "grad_norm": 1.042273998260498,
      "learning_rate": 1e-05,
      "loss": 0.4797,
      "mean_token_accuracy": 0.8492919206619263,
      "num_tokens": 361296869.0,
      "step": 2268
    },
    {
      "epoch": 1.15412004069176,
      "grad_norm": 1.0089274644851685,
      "learning_rate": 1e-05,
      "loss": 0.5019,
      "mean_token_accuracy": 0.840201735496521,
      "num_tokens": 361439766.0,
      "step": 2269
    },
    {
      "epoch": 1.1546286876907426,
      "grad_norm": 0.9834834933280945,
      "learning_rate": 1e-05,
      "loss": 0.4702,
      "mean_token_accuracy": 0.850745439529419,
      "num_tokens": 361597141.0,
      "step": 2270
    },
    {
      "epoch": 1.1551373346897253,
      "grad_norm": 1.0707170963287354,
      "learning_rate": 1e-05,
      "loss": 0.4817,
      "mean_token_accuracy": 0.846406102180481,
      "num_tokens": 361760028.0,
      "step": 2271
    },
    {
      "epoch": 1.155645981688708,
      "grad_norm": 1.0368983745574951,
      "learning_rate": 1e-05,
      "loss": 0.4804,
      "mean_token_accuracy": 0.8464101552963257,
      "num_tokens": 361931768.0,
      "step": 2272
    },
    {
      "epoch": 1.1561546286876907,
      "grad_norm": 1.189819574356079,
      "learning_rate": 1e-05,
      "loss": 0.4851,
      "mean_token_accuracy": 0.8495547771453857,
      "num_tokens": 362095781.0,
      "step": 2273
    },
    {
      "epoch": 1.1566632756866735,
      "grad_norm": 1.113652229309082,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8631555438041687,
      "num_tokens": 362246811.0,
      "step": 2274
    },
    {
      "epoch": 1.1571719226856563,
      "grad_norm": 1.0526518821716309,
      "learning_rate": 1e-05,
      "loss": 0.4915,
      "mean_token_accuracy": 0.8442506790161133,
      "num_tokens": 362394108.0,
      "step": 2275
    },
    {
      "epoch": 1.1576805696846388,
      "grad_norm": 1.0336838960647583,
      "learning_rate": 1e-05,
      "loss": 0.4785,
      "mean_token_accuracy": 0.8486695289611816,
      "num_tokens": 362555139.0,
      "step": 2276
    },
    {
      "epoch": 1.1581892166836216,
      "grad_norm": 1.0744441747665405,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8572794198989868,
      "num_tokens": 362718486.0,
      "step": 2277
    },
    {
      "epoch": 1.1586978636826042,
      "grad_norm": 1.03730309009552,
      "learning_rate": 1e-05,
      "loss": 0.4917,
      "mean_token_accuracy": 0.8441042304039001,
      "num_tokens": 362881363.0,
      "step": 2278
    },
    {
      "epoch": 1.159206510681587,
      "grad_norm": 1.021234154701233,
      "learning_rate": 1e-05,
      "loss": 0.5158,
      "mean_token_accuracy": 0.8376511335372925,
      "num_tokens": 363054375.0,
      "step": 2279
    },
    {
      "epoch": 1.1597151576805698,
      "grad_norm": 1.0153952836990356,
      "learning_rate": 1e-05,
      "loss": 0.4688,
      "mean_token_accuracy": 0.8510196805000305,
      "num_tokens": 363207013.0,
      "step": 2280
    },
    {
      "epoch": 1.1602238046795523,
      "grad_norm": 0.9547268748283386,
      "learning_rate": 1e-05,
      "loss": 0.4591,
      "mean_token_accuracy": 0.8534315228462219,
      "num_tokens": 363364310.0,
      "step": 2281
    },
    {
      "epoch": 1.1607324516785351,
      "grad_norm": 1.0788716077804565,
      "learning_rate": 1e-05,
      "loss": 0.4878,
      "mean_token_accuracy": 0.8448123931884766,
      "num_tokens": 363501174.0,
      "step": 2282
    },
    {
      "epoch": 1.1612410986775177,
      "grad_norm": 1.002040982246399,
      "learning_rate": 1e-05,
      "loss": 0.4881,
      "mean_token_accuracy": 0.8461840748786926,
      "num_tokens": 363667951.0,
      "step": 2283
    },
    {
      "epoch": 1.1617497456765005,
      "grad_norm": 0.9879819750785828,
      "learning_rate": 1e-05,
      "loss": 0.478,
      "mean_token_accuracy": 0.848838210105896,
      "num_tokens": 363831350.0,
      "step": 2284
    },
    {
      "epoch": 1.1622583926754833,
      "grad_norm": 0.9468021392822266,
      "learning_rate": 1e-05,
      "loss": 0.5046,
      "mean_token_accuracy": 0.8422623872756958,
      "num_tokens": 364006126.0,
      "step": 2285
    },
    {
      "epoch": 1.162767039674466,
      "grad_norm": 0.9665511846542358,
      "learning_rate": 1e-05,
      "loss": 0.4556,
      "mean_token_accuracy": 0.8548101186752319,
      "num_tokens": 364161537.0,
      "step": 2286
    },
    {
      "epoch": 1.1632756866734486,
      "grad_norm": 0.9804579019546509,
      "learning_rate": 1e-05,
      "loss": 0.4885,
      "mean_token_accuracy": 0.8468865156173706,
      "num_tokens": 364318922.0,
      "step": 2287
    },
    {
      "epoch": 1.1637843336724314,
      "grad_norm": 0.9549255967140198,
      "learning_rate": 1e-05,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8510030508041382,
      "num_tokens": 364468701.0,
      "step": 2288
    },
    {
      "epoch": 1.164292980671414,
      "grad_norm": 0.9666069746017456,
      "learning_rate": 1e-05,
      "loss": 0.5088,
      "mean_token_accuracy": 0.8389880061149597,
      "num_tokens": 364633717.0,
      "step": 2289
    },
    {
      "epoch": 1.1648016276703967,
      "grad_norm": 1.091866135597229,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8548486232757568,
      "num_tokens": 364793011.0,
      "step": 2290
    },
    {
      "epoch": 1.1653102746693795,
      "grad_norm": 1.9256987571716309,
      "learning_rate": 1e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8470572233200073,
      "num_tokens": 364962011.0,
      "step": 2291
    },
    {
      "epoch": 1.165818921668362,
      "grad_norm": 1.0450749397277832,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8515802621841431,
      "num_tokens": 365125068.0,
      "step": 2292
    },
    {
      "epoch": 1.1663275686673449,
      "grad_norm": 0.9731523990631104,
      "learning_rate": 1e-05,
      "loss": 0.5134,
      "mean_token_accuracy": 0.8373537063598633,
      "num_tokens": 365284576.0,
      "step": 2293
    },
    {
      "epoch": 1.1668362156663274,
      "grad_norm": 1.0927114486694336,
      "learning_rate": 1e-05,
      "loss": 0.494,
      "mean_token_accuracy": 0.8448081612586975,
      "num_tokens": 365447819.0,
      "step": 2294
    },
    {
      "epoch": 1.1673448626653102,
      "grad_norm": 0.9757579565048218,
      "learning_rate": 1e-05,
      "loss": 0.478,
      "mean_token_accuracy": 0.8482175469398499,
      "num_tokens": 365608276.0,
      "step": 2295
    },
    {
      "epoch": 1.167853509664293,
      "grad_norm": 0.9663304686546326,
      "learning_rate": 1e-05,
      "loss": 0.5026,
      "mean_token_accuracy": 0.8430571556091309,
      "num_tokens": 365766838.0,
      "step": 2296
    },
    {
      "epoch": 1.1683621566632758,
      "grad_norm": 0.9581319689750671,
      "learning_rate": 1e-05,
      "loss": 0.4884,
      "mean_token_accuracy": 0.8461264371871948,
      "num_tokens": 365938182.0,
      "step": 2297
    },
    {
      "epoch": 1.1688708036622584,
      "grad_norm": 0.9772895574569702,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.8557355403900146,
      "num_tokens": 366101878.0,
      "step": 2298
    },
    {
      "epoch": 1.1693794506612412,
      "grad_norm": 0.9259494543075562,
      "learning_rate": 1e-05,
      "loss": 0.4814,
      "mean_token_accuracy": 0.8471612930297852,
      "num_tokens": 366268381.0,
      "step": 2299
    },
    {
      "epoch": 1.1698880976602237,
      "grad_norm": 1.005831003189087,
      "learning_rate": 1e-05,
      "loss": 0.4718,
      "mean_token_accuracy": 0.85094153881073,
      "num_tokens": 366439188.0,
      "step": 2300
    },
    {
      "epoch": 1.1703967446592065,
      "grad_norm": 0.9507022500038147,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8511804938316345,
      "num_tokens": 366610850.0,
      "step": 2301
    },
    {
      "epoch": 1.1709053916581893,
      "grad_norm": 0.9647862315177917,
      "learning_rate": 1e-05,
      "loss": 0.531,
      "mean_token_accuracy": 0.8360495567321777,
      "num_tokens": 366769572.0,
      "step": 2302
    },
    {
      "epoch": 1.1714140386571719,
      "grad_norm": 0.9812116622924805,
      "learning_rate": 1e-05,
      "loss": 0.4471,
      "mean_token_accuracy": 0.8576605319976807,
      "num_tokens": 366937681.0,
      "step": 2303
    },
    {
      "epoch": 1.1719226856561547,
      "grad_norm": 1.0458122491836548,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8458260297775269,
      "num_tokens": 367097208.0,
      "step": 2304
    },
    {
      "epoch": 1.1724313326551372,
      "grad_norm": 1.0308116674423218,
      "learning_rate": 1e-05,
      "loss": 0.4944,
      "mean_token_accuracy": 0.8454901576042175,
      "num_tokens": 367263406.0,
      "step": 2305
    },
    {
      "epoch": 1.17293997965412,
      "grad_norm": 1.1648168563842773,
      "learning_rate": 1e-05,
      "loss": 0.5273,
      "mean_token_accuracy": 0.8344169855117798,
      "num_tokens": 367424929.0,
      "step": 2306
    },
    {
      "epoch": 1.1734486266531028,
      "grad_norm": 1.046728491783142,
      "learning_rate": 1e-05,
      "loss": 0.4756,
      "mean_token_accuracy": 0.8486387729644775,
      "num_tokens": 367591140.0,
      "step": 2307
    },
    {
      "epoch": 1.1739572736520856,
      "grad_norm": 0.9279728531837463,
      "learning_rate": 1e-05,
      "loss": 0.4859,
      "mean_token_accuracy": 0.8459349870681763,
      "num_tokens": 367753127.0,
      "step": 2308
    },
    {
      "epoch": 1.1744659206510681,
      "grad_norm": 1.1688655614852905,
      "learning_rate": 1e-05,
      "loss": 0.4931,
      "mean_token_accuracy": 0.8448759913444519,
      "num_tokens": 367915582.0,
      "step": 2309
    },
    {
      "epoch": 1.174974567650051,
      "grad_norm": 0.9621530175209045,
      "learning_rate": 1e-05,
      "loss": 0.4815,
      "mean_token_accuracy": 0.8467572331428528,
      "num_tokens": 368073652.0,
      "step": 2310
    },
    {
      "epoch": 1.1754832146490335,
      "grad_norm": 1.200062870979309,
      "learning_rate": 1e-05,
      "loss": 0.4631,
      "mean_token_accuracy": 0.851768970489502,
      "num_tokens": 368224644.0,
      "step": 2311
    },
    {
      "epoch": 1.1759918616480163,
      "grad_norm": 1.0809787511825562,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8461123704910278,
      "num_tokens": 368389034.0,
      "step": 2312
    },
    {
      "epoch": 1.176500508646999,
      "grad_norm": 1.009333848953247,
      "learning_rate": 1e-05,
      "loss": 0.4701,
      "mean_token_accuracy": 0.8512672185897827,
      "num_tokens": 368548528.0,
      "step": 2313
    },
    {
      "epoch": 1.1770091556459816,
      "grad_norm": 1.1124660968780518,
      "learning_rate": 1e-05,
      "loss": 0.5256,
      "mean_token_accuracy": 0.8362528085708618,
      "num_tokens": 368721238.0,
      "step": 2314
    },
    {
      "epoch": 1.1775178026449644,
      "grad_norm": 1.0473283529281616,
      "learning_rate": 1e-05,
      "loss": 0.4848,
      "mean_token_accuracy": 0.8465226888656616,
      "num_tokens": 368875782.0,
      "step": 2315
    },
    {
      "epoch": 1.178026449643947,
      "grad_norm": 1.037091612815857,
      "learning_rate": 1e-05,
      "loss": 0.4949,
      "mean_token_accuracy": 0.8469461798667908,
      "num_tokens": 369039106.0,
      "step": 2316
    },
    {
      "epoch": 1.1785350966429298,
      "grad_norm": 1.1117887496948242,
      "learning_rate": 1e-05,
      "loss": 0.4911,
      "mean_token_accuracy": 0.8467146754264832,
      "num_tokens": 369201771.0,
      "step": 2317
    },
    {
      "epoch": 1.1790437436419126,
      "grad_norm": 1.0993705987930298,
      "learning_rate": 1e-05,
      "loss": 0.4529,
      "mean_token_accuracy": 0.8551526069641113,
      "num_tokens": 369375622.0,
      "step": 2318
    },
    {
      "epoch": 1.1795523906408953,
      "grad_norm": 1.03167724609375,
      "learning_rate": 1e-05,
      "loss": 0.4946,
      "mean_token_accuracy": 0.8446077108383179,
      "num_tokens": 369535466.0,
      "step": 2319
    },
    {
      "epoch": 1.180061037639878,
      "grad_norm": 1.0695133209228516,
      "learning_rate": 1e-05,
      "loss": 0.5084,
      "mean_token_accuracy": 0.840103030204773,
      "num_tokens": 369688789.0,
      "step": 2320
    },
    {
      "epoch": 1.1805696846388607,
      "grad_norm": 0.9820632338523865,
      "learning_rate": 1e-05,
      "loss": 0.4818,
      "mean_token_accuracy": 0.8472625017166138,
      "num_tokens": 369844154.0,
      "step": 2321
    },
    {
      "epoch": 1.1810783316378433,
      "grad_norm": 1.0760818719863892,
      "learning_rate": 1e-05,
      "loss": 0.5241,
      "mean_token_accuracy": 0.8362928628921509,
      "num_tokens": 369998967.0,
      "step": 2322
    },
    {
      "epoch": 1.181586978636826,
      "grad_norm": 1.0259249210357666,
      "learning_rate": 1e-05,
      "loss": 0.499,
      "mean_token_accuracy": 0.8428303599357605,
      "num_tokens": 370168382.0,
      "step": 2323
    },
    {
      "epoch": 1.1820956256358088,
      "grad_norm": 0.956731379032135,
      "learning_rate": 1e-05,
      "loss": 0.4839,
      "mean_token_accuracy": 0.8463218212127686,
      "num_tokens": 370333473.0,
      "step": 2324
    },
    {
      "epoch": 1.1826042726347914,
      "grad_norm": 1.0034064054489136,
      "learning_rate": 1e-05,
      "loss": 0.4796,
      "mean_token_accuracy": 0.8488554954528809,
      "num_tokens": 370507572.0,
      "step": 2325
    },
    {
      "epoch": 1.1831129196337742,
      "grad_norm": 0.9154096841812134,
      "learning_rate": 1e-05,
      "loss": 0.4994,
      "mean_token_accuracy": 0.8431752920150757,
      "num_tokens": 370660466.0,
      "step": 2326
    },
    {
      "epoch": 1.1836215666327567,
      "grad_norm": 0.9893860220909119,
      "learning_rate": 1e-05,
      "loss": 0.5255,
      "mean_token_accuracy": 0.8330589532852173,
      "num_tokens": 370828832.0,
      "step": 2327
    },
    {
      "epoch": 1.1841302136317395,
      "grad_norm": 1.0146458148956299,
      "learning_rate": 1e-05,
      "loss": 0.4863,
      "mean_token_accuracy": 0.8479539155960083,
      "num_tokens": 370998386.0,
      "step": 2328
    },
    {
      "epoch": 1.1846388606307223,
      "grad_norm": 0.9473051428794861,
      "learning_rate": 1e-05,
      "loss": 0.4789,
      "mean_token_accuracy": 0.8470596075057983,
      "num_tokens": 371166920.0,
      "step": 2329
    },
    {
      "epoch": 1.1851475076297049,
      "grad_norm": 0.9573273062705994,
      "learning_rate": 1e-05,
      "loss": 0.4817,
      "mean_token_accuracy": 0.8494331240653992,
      "num_tokens": 371332751.0,
      "step": 2330
    },
    {
      "epoch": 1.1856561546286877,
      "grad_norm": 0.9692023992538452,
      "learning_rate": 1e-05,
      "loss": 0.5028,
      "mean_token_accuracy": 0.8427772521972656,
      "num_tokens": 371495502.0,
      "step": 2331
    },
    {
      "epoch": 1.1861648016276705,
      "grad_norm": 0.9364044666290283,
      "learning_rate": 1e-05,
      "loss": 0.4888,
      "mean_token_accuracy": 0.8459036350250244,
      "num_tokens": 371662573.0,
      "step": 2332
    },
    {
      "epoch": 1.186673448626653,
      "grad_norm": 0.9484080076217651,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8498533964157104,
      "num_tokens": 371825073.0,
      "step": 2333
    },
    {
      "epoch": 1.1871820956256358,
      "grad_norm": 0.9878451228141785,
      "learning_rate": 1e-05,
      "loss": 0.4483,
      "mean_token_accuracy": 0.856770932674408,
      "num_tokens": 371986831.0,
      "step": 2334
    },
    {
      "epoch": 1.1876907426246186,
      "grad_norm": 0.9890164732933044,
      "learning_rate": 1e-05,
      "loss": 0.5071,
      "mean_token_accuracy": 0.8417010307312012,
      "num_tokens": 372140500.0,
      "step": 2335
    },
    {
      "epoch": 1.1881993896236012,
      "grad_norm": 0.9335831999778748,
      "learning_rate": 1e-05,
      "loss": 0.5248,
      "mean_token_accuracy": 0.8363795876502991,
      "num_tokens": 372315615.0,
      "step": 2336
    },
    {
      "epoch": 1.188708036622584,
      "grad_norm": 1.0428636074066162,
      "learning_rate": 1e-05,
      "loss": 0.4701,
      "mean_token_accuracy": 0.8501526117324829,
      "num_tokens": 372464286.0,
      "step": 2337
    },
    {
      "epoch": 1.1892166836215665,
      "grad_norm": 0.9321058988571167,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.844826877117157,
      "num_tokens": 372638897.0,
      "step": 2338
    },
    {
      "epoch": 1.1897253306205493,
      "grad_norm": 1.0857921838760376,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8441271781921387,
      "num_tokens": 372795195.0,
      "step": 2339
    },
    {
      "epoch": 1.190233977619532,
      "grad_norm": 1.0063886642456055,
      "learning_rate": 1e-05,
      "loss": 0.4696,
      "mean_token_accuracy": 0.8519972562789917,
      "num_tokens": 372956645.0,
      "step": 2340
    },
    {
      "epoch": 1.1907426246185147,
      "grad_norm": 0.969150722026825,
      "learning_rate": 1e-05,
      "loss": 0.4692,
      "mean_token_accuracy": 0.8504284024238586,
      "num_tokens": 373118539.0,
      "step": 2341
    },
    {
      "epoch": 1.1912512716174974,
      "grad_norm": 0.9527782201766968,
      "learning_rate": 1e-05,
      "loss": 0.4593,
      "mean_token_accuracy": 0.8541048765182495,
      "num_tokens": 373265980.0,
      "step": 2342
    },
    {
      "epoch": 1.1917599186164802,
      "grad_norm": 0.9804020524024963,
      "learning_rate": 1e-05,
      "loss": 0.521,
      "mean_token_accuracy": 0.8361110687255859,
      "num_tokens": 373433225.0,
      "step": 2343
    },
    {
      "epoch": 1.1922685656154628,
      "grad_norm": 1.0810366868972778,
      "learning_rate": 1e-05,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8493367433547974,
      "num_tokens": 373575839.0,
      "step": 2344
    },
    {
      "epoch": 1.1927772126144456,
      "grad_norm": 1.061769962310791,
      "learning_rate": 1e-05,
      "loss": 0.4745,
      "mean_token_accuracy": 0.8492453098297119,
      "num_tokens": 373725507.0,
      "step": 2345
    },
    {
      "epoch": 1.1932858596134284,
      "grad_norm": 1.0141011476516724,
      "learning_rate": 1e-05,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8460336923599243,
      "num_tokens": 373896838.0,
      "step": 2346
    },
    {
      "epoch": 1.193794506612411,
      "grad_norm": 0.9880650639533997,
      "learning_rate": 1e-05,
      "loss": 0.4847,
      "mean_token_accuracy": 0.8462492227554321,
      "num_tokens": 374061213.0,
      "step": 2347
    },
    {
      "epoch": 1.1943031536113937,
      "grad_norm": 1.1626392602920532,
      "learning_rate": 1e-05,
      "loss": 0.4729,
      "mean_token_accuracy": 0.850334882736206,
      "num_tokens": 374210420.0,
      "step": 2348
    },
    {
      "epoch": 1.1948118006103763,
      "grad_norm": 1.1458953619003296,
      "learning_rate": 1e-05,
      "loss": 0.4472,
      "mean_token_accuracy": 0.8559433221817017,
      "num_tokens": 374372042.0,
      "step": 2349
    },
    {
      "epoch": 1.195320447609359,
      "grad_norm": 1.1105952262878418,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.8556099534034729,
      "num_tokens": 374539265.0,
      "step": 2350
    },
    {
      "epoch": 1.1958290946083419,
      "grad_norm": 1.0160725116729736,
      "learning_rate": 1e-05,
      "loss": 0.487,
      "mean_token_accuracy": 0.8456284999847412,
      "num_tokens": 374697019.0,
      "step": 2351
    },
    {
      "epoch": 1.1963377416073244,
      "grad_norm": 0.9500008821487427,
      "learning_rate": 1e-05,
      "loss": 0.4688,
      "mean_token_accuracy": 0.8500959873199463,
      "num_tokens": 374846358.0,
      "step": 2352
    },
    {
      "epoch": 1.1968463886063072,
      "grad_norm": 1.1950064897537231,
      "learning_rate": 1e-05,
      "loss": 0.4595,
      "mean_token_accuracy": 0.8534548282623291,
      "num_tokens": 375007223.0,
      "step": 2353
    },
    {
      "epoch": 1.19735503560529,
      "grad_norm": 1.174160122871399,
      "learning_rate": 1e-05,
      "loss": 0.4704,
      "mean_token_accuracy": 0.8509343266487122,
      "num_tokens": 375157542.0,
      "step": 2354
    },
    {
      "epoch": 1.1978636826042726,
      "grad_norm": 1.091294765472412,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8505253791809082,
      "num_tokens": 375315585.0,
      "step": 2355
    },
    {
      "epoch": 1.1983723296032553,
      "grad_norm": 0.9935697913169861,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8474482893943787,
      "num_tokens": 375472279.0,
      "step": 2356
    },
    {
      "epoch": 1.1988809766022381,
      "grad_norm": 1.0689009428024292,
      "learning_rate": 1e-05,
      "loss": 0.4832,
      "mean_token_accuracy": 0.8478238582611084,
      "num_tokens": 375630662.0,
      "step": 2357
    },
    {
      "epoch": 1.1993896236012207,
      "grad_norm": 1.1201637983322144,
      "learning_rate": 1e-05,
      "loss": 0.4646,
      "mean_token_accuracy": 0.8541030883789062,
      "num_tokens": 375785251.0,
      "step": 2358
    },
    {
      "epoch": 1.1998982706002035,
      "grad_norm": 0.9344379901885986,
      "learning_rate": 1e-05,
      "loss": 0.4822,
      "mean_token_accuracy": 0.8468241691589355,
      "num_tokens": 375945570.0,
      "step": 2359
    },
    {
      "epoch": 1.200406917599186,
      "grad_norm": 1.052104115486145,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.857330858707428,
      "num_tokens": 376099013.0,
      "step": 2360
    },
    {
      "epoch": 1.2009155645981688,
      "grad_norm": 0.9729582071304321,
      "learning_rate": 1e-05,
      "loss": 0.4897,
      "mean_token_accuracy": 0.8475154638290405,
      "num_tokens": 376260578.0,
      "step": 2361
    },
    {
      "epoch": 1.2014242115971516,
      "grad_norm": 0.9846155047416687,
      "learning_rate": 1e-05,
      "loss": 0.4777,
      "mean_token_accuracy": 0.848920464515686,
      "num_tokens": 376414415.0,
      "step": 2362
    },
    {
      "epoch": 1.2019328585961342,
      "grad_norm": 0.9831588864326477,
      "learning_rate": 1e-05,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8465712070465088,
      "num_tokens": 376570177.0,
      "step": 2363
    },
    {
      "epoch": 1.202441505595117,
      "grad_norm": 0.9749376773834229,
      "learning_rate": 1e-05,
      "loss": 0.5013,
      "mean_token_accuracy": 0.8418674468994141,
      "num_tokens": 376731071.0,
      "step": 2364
    },
    {
      "epoch": 1.2029501525940998,
      "grad_norm": 1.0017136335372925,
      "learning_rate": 1e-05,
      "loss": 0.4833,
      "mean_token_accuracy": 0.8458402752876282,
      "num_tokens": 376900019.0,
      "step": 2365
    },
    {
      "epoch": 1.2034587995930823,
      "grad_norm": 0.9792733788490295,
      "learning_rate": 1e-05,
      "loss": 0.5065,
      "mean_token_accuracy": 0.8399918079376221,
      "num_tokens": 377061505.0,
      "step": 2366
    },
    {
      "epoch": 1.2039674465920651,
      "grad_norm": 1.000549554824829,
      "learning_rate": 1e-05,
      "loss": 0.4922,
      "mean_token_accuracy": 0.8457726836204529,
      "num_tokens": 377219211.0,
      "step": 2367
    },
    {
      "epoch": 1.204476093591048,
      "grad_norm": 1.007737398147583,
      "learning_rate": 1e-05,
      "loss": 0.5006,
      "mean_token_accuracy": 0.8432072401046753,
      "num_tokens": 377377739.0,
      "step": 2368
    },
    {
      "epoch": 1.2049847405900305,
      "grad_norm": 0.9702564477920532,
      "learning_rate": 1e-05,
      "loss": 0.526,
      "mean_token_accuracy": 0.8351813554763794,
      "num_tokens": 377541932.0,
      "step": 2369
    },
    {
      "epoch": 1.2054933875890133,
      "grad_norm": 1.081968903541565,
      "learning_rate": 1e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.8512855172157288,
      "num_tokens": 377694342.0,
      "step": 2370
    },
    {
      "epoch": 1.2060020345879958,
      "grad_norm": 0.9188628196716309,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.858360230922699,
      "num_tokens": 377858941.0,
      "step": 2371
    },
    {
      "epoch": 1.2065106815869786,
      "grad_norm": 1.0161525011062622,
      "learning_rate": 1e-05,
      "loss": 0.5053,
      "mean_token_accuracy": 0.8422462940216064,
      "num_tokens": 378013523.0,
      "step": 2372
    },
    {
      "epoch": 1.2070193285859614,
      "grad_norm": 1.1344497203826904,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8525291681289673,
      "num_tokens": 378174442.0,
      "step": 2373
    },
    {
      "epoch": 1.207527975584944,
      "grad_norm": 1.032597303390503,
      "learning_rate": 1e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8476738333702087,
      "num_tokens": 378315219.0,
      "step": 2374
    },
    {
      "epoch": 1.2080366225839267,
      "grad_norm": 1.0668175220489502,
      "learning_rate": 1e-05,
      "loss": 0.4805,
      "mean_token_accuracy": 0.8466565608978271,
      "num_tokens": 378458710.0,
      "step": 2375
    },
    {
      "epoch": 1.2085452695829095,
      "grad_norm": 0.9233808517456055,
      "learning_rate": 1e-05,
      "loss": 0.4621,
      "mean_token_accuracy": 0.855979859828949,
      "num_tokens": 378622003.0,
      "step": 2376
    },
    {
      "epoch": 1.209053916581892,
      "grad_norm": 1.073775053024292,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8557116985321045,
      "num_tokens": 378784837.0,
      "step": 2377
    },
    {
      "epoch": 1.2095625635808749,
      "grad_norm": 1.0039360523223877,
      "learning_rate": 1e-05,
      "loss": 0.4823,
      "mean_token_accuracy": 0.8465985059738159,
      "num_tokens": 378932580.0,
      "step": 2378
    },
    {
      "epoch": 1.2100712105798577,
      "grad_norm": 1.0067378282546997,
      "learning_rate": 1e-05,
      "loss": 0.4733,
      "mean_token_accuracy": 0.8507084846496582,
      "num_tokens": 379088195.0,
      "step": 2379
    },
    {
      "epoch": 1.2105798575788402,
      "grad_norm": 1.0960066318511963,
      "learning_rate": 1e-05,
      "loss": 0.523,
      "mean_token_accuracy": 0.8363871574401855,
      "num_tokens": 379261815.0,
      "step": 2380
    },
    {
      "epoch": 1.211088504577823,
      "grad_norm": 0.9723873734474182,
      "learning_rate": 1e-05,
      "loss": 0.4883,
      "mean_token_accuracy": 0.8440775871276855,
      "num_tokens": 379421440.0,
      "step": 2381
    },
    {
      "epoch": 1.2115971515768056,
      "grad_norm": 1.0995724201202393,
      "learning_rate": 1e-05,
      "loss": 0.4903,
      "mean_token_accuracy": 0.8438718318939209,
      "num_tokens": 379589270.0,
      "step": 2382
    },
    {
      "epoch": 1.2121057985757884,
      "grad_norm": 1.0866910219192505,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.8414846658706665,
      "num_tokens": 379744264.0,
      "step": 2383
    },
    {
      "epoch": 1.2126144455747712,
      "grad_norm": 1.0899282693862915,
      "learning_rate": 1e-05,
      "loss": 0.4717,
      "mean_token_accuracy": 0.8495341539382935,
      "num_tokens": 379905958.0,
      "step": 2384
    },
    {
      "epoch": 1.2131230925737537,
      "grad_norm": 1.0874824523925781,
      "learning_rate": 1e-05,
      "loss": 0.4888,
      "mean_token_accuracy": 0.8447736501693726,
      "num_tokens": 380056524.0,
      "step": 2385
    },
    {
      "epoch": 1.2136317395727365,
      "grad_norm": 1.0588302612304688,
      "learning_rate": 1e-05,
      "loss": 0.4844,
      "mean_token_accuracy": 0.8479455709457397,
      "num_tokens": 380215696.0,
      "step": 2386
    },
    {
      "epoch": 1.2141403865717193,
      "grad_norm": 1.008784294128418,
      "learning_rate": 1e-05,
      "loss": 0.4769,
      "mean_token_accuracy": 0.8480744957923889,
      "num_tokens": 380386760.0,
      "step": 2387
    },
    {
      "epoch": 1.2146490335707019,
      "grad_norm": 1.0145485401153564,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.8556555509567261,
      "num_tokens": 380552451.0,
      "step": 2388
    },
    {
      "epoch": 1.2151576805696847,
      "grad_norm": 1.0312669277191162,
      "learning_rate": 1e-05,
      "loss": 0.4869,
      "mean_token_accuracy": 0.8460147380828857,
      "num_tokens": 380699985.0,
      "step": 2389
    },
    {
      "epoch": 1.2156663275686674,
      "grad_norm": 0.9727829694747925,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8518386483192444,
      "num_tokens": 380862532.0,
      "step": 2390
    },
    {
      "epoch": 1.21617497456765,
      "grad_norm": 0.9736183881759644,
      "learning_rate": 1e-05,
      "loss": 0.4602,
      "mean_token_accuracy": 0.8531427383422852,
      "num_tokens": 381023201.0,
      "step": 2391
    },
    {
      "epoch": 1.2166836215666328,
      "grad_norm": 1.0878318548202515,
      "learning_rate": 1e-05,
      "loss": 0.4953,
      "mean_token_accuracy": 0.842427134513855,
      "num_tokens": 381181297.0,
      "step": 2392
    },
    {
      "epoch": 1.2171922685656154,
      "grad_norm": 0.9806477427482605,
      "learning_rate": 1e-05,
      "loss": 0.534,
      "mean_token_accuracy": 0.8337027430534363,
      "num_tokens": 381335556.0,
      "step": 2393
    },
    {
      "epoch": 1.2177009155645981,
      "grad_norm": 1.0319772958755493,
      "learning_rate": 1e-05,
      "loss": 0.5084,
      "mean_token_accuracy": 0.8385846614837646,
      "num_tokens": 381492918.0,
      "step": 2394
    },
    {
      "epoch": 1.218209562563581,
      "grad_norm": 1.013101577758789,
      "learning_rate": 1e-05,
      "loss": 0.4856,
      "mean_token_accuracy": 0.8462007641792297,
      "num_tokens": 381650827.0,
      "step": 2395
    },
    {
      "epoch": 1.2187182095625635,
      "grad_norm": 0.9341661334037781,
      "learning_rate": 1e-05,
      "loss": 0.5025,
      "mean_token_accuracy": 0.8424181938171387,
      "num_tokens": 381810610.0,
      "step": 2396
    },
    {
      "epoch": 1.2192268565615463,
      "grad_norm": 1.0404284000396729,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "mean_token_accuracy": 0.8567433953285217,
      "num_tokens": 381974046.0,
      "step": 2397
    },
    {
      "epoch": 1.219735503560529,
      "grad_norm": 1.1205888986587524,
      "learning_rate": 1e-05,
      "loss": 0.4949,
      "mean_token_accuracy": 0.8454039096832275,
      "num_tokens": 382127627.0,
      "step": 2398
    },
    {
      "epoch": 1.2202441505595116,
      "grad_norm": 1.0638031959533691,
      "learning_rate": 1e-05,
      "loss": 0.4833,
      "mean_token_accuracy": 0.8463817238807678,
      "num_tokens": 382289172.0,
      "step": 2399
    },
    {
      "epoch": 1.2207527975584944,
      "grad_norm": 1.1557694673538208,
      "learning_rate": 1e-05,
      "loss": 0.4869,
      "mean_token_accuracy": 0.8461000323295593,
      "num_tokens": 382453657.0,
      "step": 2400
    },
    {
      "epoch": 1.2212614445574772,
      "grad_norm": 1.068440556526184,
      "learning_rate": 1e-05,
      "loss": 0.4782,
      "mean_token_accuracy": 0.8476709127426147,
      "num_tokens": 382617744.0,
      "step": 2401
    },
    {
      "epoch": 1.2217700915564598,
      "grad_norm": 0.9085849523544312,
      "learning_rate": 1e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.851020336151123,
      "num_tokens": 382781691.0,
      "step": 2402
    },
    {
      "epoch": 1.2222787385554426,
      "grad_norm": 1.0901038646697998,
      "learning_rate": 1e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8521700501441956,
      "num_tokens": 382944190.0,
      "step": 2403
    },
    {
      "epoch": 1.2227873855544251,
      "grad_norm": 0.984150767326355,
      "learning_rate": 1e-05,
      "loss": 0.4909,
      "mean_token_accuracy": 0.8450653553009033,
      "num_tokens": 383107927.0,
      "step": 2404
    },
    {
      "epoch": 1.223296032553408,
      "grad_norm": 1.288863182067871,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8521753549575806,
      "num_tokens": 383270582.0,
      "step": 2405
    },
    {
      "epoch": 1.2238046795523907,
      "grad_norm": 1.1820831298828125,
      "learning_rate": 1e-05,
      "loss": 0.5097,
      "mean_token_accuracy": 0.8402947783470154,
      "num_tokens": 383429967.0,
      "step": 2406
    },
    {
      "epoch": 1.2243133265513733,
      "grad_norm": 1.016980528831482,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8534340858459473,
      "num_tokens": 383588086.0,
      "step": 2407
    },
    {
      "epoch": 1.224821973550356,
      "grad_norm": 1.2975318431854248,
      "learning_rate": 1e-05,
      "loss": 0.5041,
      "mean_token_accuracy": 0.8402259349822998,
      "num_tokens": 383744521.0,
      "step": 2408
    },
    {
      "epoch": 1.2253306205493388,
      "grad_norm": 1.1058008670806885,
      "learning_rate": 1e-05,
      "loss": 0.4759,
      "mean_token_accuracy": 0.8495995998382568,
      "num_tokens": 383912014.0,
      "step": 2409
    },
    {
      "epoch": 1.2258392675483214,
      "grad_norm": 1.1800510883331299,
      "learning_rate": 1e-05,
      "loss": 0.4881,
      "mean_token_accuracy": 0.8459175825119019,
      "num_tokens": 384070051.0,
      "step": 2410
    },
    {
      "epoch": 1.2263479145473042,
      "grad_norm": 1.0510597229003906,
      "learning_rate": 1e-05,
      "loss": 0.507,
      "mean_token_accuracy": 0.8406599760055542,
      "num_tokens": 384232662.0,
      "step": 2411
    },
    {
      "epoch": 1.226856561546287,
      "grad_norm": 0.9730938076972961,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8466057777404785,
      "num_tokens": 384381225.0,
      "step": 2412
    },
    {
      "epoch": 1.2273652085452695,
      "grad_norm": 0.9850808382034302,
      "learning_rate": 1e-05,
      "loss": 0.4859,
      "mean_token_accuracy": 0.8452000021934509,
      "num_tokens": 384542789.0,
      "step": 2413
    },
    {
      "epoch": 1.2278738555442523,
      "grad_norm": 0.9626375436782837,
      "learning_rate": 1e-05,
      "loss": 0.5083,
      "mean_token_accuracy": 0.8400334119796753,
      "num_tokens": 384710145.0,
      "step": 2414
    },
    {
      "epoch": 1.2283825025432349,
      "grad_norm": 1.0044533014297485,
      "learning_rate": 1e-05,
      "loss": 0.5257,
      "mean_token_accuracy": 0.8348107933998108,
      "num_tokens": 384873444.0,
      "step": 2415
    },
    {
      "epoch": 1.2288911495422177,
      "grad_norm": 1.079222321510315,
      "learning_rate": 1e-05,
      "loss": 0.4924,
      "mean_token_accuracy": 0.8461150527000427,
      "num_tokens": 385030375.0,
      "step": 2416
    },
    {
      "epoch": 1.2293997965412005,
      "grad_norm": 1.0069643259048462,
      "learning_rate": 1e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8482726216316223,
      "num_tokens": 385195808.0,
      "step": 2417
    },
    {
      "epoch": 1.229908443540183,
      "grad_norm": 1.000921368598938,
      "learning_rate": 1e-05,
      "loss": 0.5255,
      "mean_token_accuracy": 0.8357363939285278,
      "num_tokens": 385361429.0,
      "step": 2418
    },
    {
      "epoch": 1.2304170905391658,
      "grad_norm": 0.9638259410858154,
      "learning_rate": 1e-05,
      "loss": 0.4537,
      "mean_token_accuracy": 0.8551135063171387,
      "num_tokens": 385524948.0,
      "step": 2419
    },
    {
      "epoch": 1.2309257375381486,
      "grad_norm": 0.9751047492027283,
      "learning_rate": 1e-05,
      "loss": 0.525,
      "mean_token_accuracy": 0.834947407245636,
      "num_tokens": 385690394.0,
      "step": 2420
    },
    {
      "epoch": 1.2314343845371312,
      "grad_norm": 0.9985384941101074,
      "learning_rate": 1e-05,
      "loss": 0.4961,
      "mean_token_accuracy": 0.8441144227981567,
      "num_tokens": 385849268.0,
      "step": 2421
    },
    {
      "epoch": 1.231943031536114,
      "grad_norm": 0.9478381276130676,
      "learning_rate": 1e-05,
      "loss": 0.4534,
      "mean_token_accuracy": 0.8564085960388184,
      "num_tokens": 386012537.0,
      "step": 2422
    },
    {
      "epoch": 1.2324516785350967,
      "grad_norm": 1.10886812210083,
      "learning_rate": 1e-05,
      "loss": 0.5219,
      "mean_token_accuracy": 0.8358174562454224,
      "num_tokens": 386175987.0,
      "step": 2423
    },
    {
      "epoch": 1.2329603255340793,
      "grad_norm": 0.9503421187400818,
      "learning_rate": 1e-05,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8555249571800232,
      "num_tokens": 386338087.0,
      "step": 2424
    },
    {
      "epoch": 1.233468972533062,
      "grad_norm": 0.9710291624069214,
      "learning_rate": 1e-05,
      "loss": 0.4705,
      "mean_token_accuracy": 0.8473029136657715,
      "num_tokens": 386495175.0,
      "step": 2425
    },
    {
      "epoch": 1.2339776195320447,
      "grad_norm": 1.0143202543258667,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8451850414276123,
      "num_tokens": 386663700.0,
      "step": 2426
    },
    {
      "epoch": 1.2344862665310274,
      "grad_norm": 0.9353373050689697,
      "learning_rate": 1e-05,
      "loss": 0.4714,
      "mean_token_accuracy": 0.852155327796936,
      "num_tokens": 386818763.0,
      "step": 2427
    },
    {
      "epoch": 1.2349949135300102,
      "grad_norm": 0.9928601980209351,
      "learning_rate": 1e-05,
      "loss": 0.499,
      "mean_token_accuracy": 0.842134952545166,
      "num_tokens": 386979867.0,
      "step": 2428
    },
    {
      "epoch": 1.2355035605289928,
      "grad_norm": 0.9530983567237854,
      "learning_rate": 1e-05,
      "loss": 0.4602,
      "mean_token_accuracy": 0.8531316518783569,
      "num_tokens": 387146689.0,
      "step": 2429
    },
    {
      "epoch": 1.2360122075279756,
      "grad_norm": 0.9834848642349243,
      "learning_rate": 1e-05,
      "loss": 0.4913,
      "mean_token_accuracy": 0.8447766304016113,
      "num_tokens": 387308229.0,
      "step": 2430
    },
    {
      "epoch": 1.2365208545269584,
      "grad_norm": 1.017050862312317,
      "learning_rate": 1e-05,
      "loss": 0.5054,
      "mean_token_accuracy": 0.839943528175354,
      "num_tokens": 387464799.0,
      "step": 2431
    },
    {
      "epoch": 1.237029501525941,
      "grad_norm": 1.0119413137435913,
      "learning_rate": 1e-05,
      "loss": 0.5069,
      "mean_token_accuracy": 0.8400273323059082,
      "num_tokens": 387633541.0,
      "step": 2432
    },
    {
      "epoch": 1.2375381485249237,
      "grad_norm": 0.9309226274490356,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8526896834373474,
      "num_tokens": 387791273.0,
      "step": 2433
    },
    {
      "epoch": 1.2380467955239065,
      "grad_norm": 1.1053496599197388,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8502742648124695,
      "num_tokens": 387945149.0,
      "step": 2434
    },
    {
      "epoch": 1.238555442522889,
      "grad_norm": 1.0243339538574219,
      "learning_rate": 1e-05,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8547742962837219,
      "num_tokens": 388104340.0,
      "step": 2435
    },
    {
      "epoch": 1.2390640895218719,
      "grad_norm": 0.9826825857162476,
      "learning_rate": 1e-05,
      "loss": 0.4782,
      "mean_token_accuracy": 0.8481093049049377,
      "num_tokens": 388262139.0,
      "step": 2436
    },
    {
      "epoch": 1.2395727365208544,
      "grad_norm": 0.9026824235916138,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8482067584991455,
      "num_tokens": 388415175.0,
      "step": 2437
    },
    {
      "epoch": 1.2400813835198372,
      "grad_norm": 1.0203884840011597,
      "learning_rate": 1e-05,
      "loss": 0.4813,
      "mean_token_accuracy": 0.8465292453765869,
      "num_tokens": 388567660.0,
      "step": 2438
    },
    {
      "epoch": 1.24059003051882,
      "grad_norm": 1.0784242153167725,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.8525174856185913,
      "num_tokens": 388711146.0,
      "step": 2439
    },
    {
      "epoch": 1.2410986775178026,
      "grad_norm": 1.0541739463806152,
      "learning_rate": 1e-05,
      "loss": 0.4858,
      "mean_token_accuracy": 0.8467655181884766,
      "num_tokens": 388862732.0,
      "step": 2440
    },
    {
      "epoch": 1.2416073245167853,
      "grad_norm": 1.068041443824768,
      "learning_rate": 1e-05,
      "loss": 0.5087,
      "mean_token_accuracy": 0.8395960927009583,
      "num_tokens": 389018134.0,
      "step": 2441
    },
    {
      "epoch": 1.2421159715157681,
      "grad_norm": 0.9338946342468262,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8567076921463013,
      "num_tokens": 389171597.0,
      "step": 2442
    },
    {
      "epoch": 1.2426246185147507,
      "grad_norm": 0.9714147448539734,
      "learning_rate": 1e-05,
      "loss": 0.5077,
      "mean_token_accuracy": 0.8383303880691528,
      "num_tokens": 389316922.0,
      "step": 2443
    },
    {
      "epoch": 1.2431332655137335,
      "grad_norm": 0.9373910427093506,
      "learning_rate": 1e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8496674299240112,
      "num_tokens": 389471070.0,
      "step": 2444
    },
    {
      "epoch": 1.2436419125127163,
      "grad_norm": 0.9623604416847229,
      "learning_rate": 1e-05,
      "loss": 0.4782,
      "mean_token_accuracy": 0.8476322889328003,
      "num_tokens": 389630233.0,
      "step": 2445
    },
    {
      "epoch": 1.2441505595116988,
      "grad_norm": 0.9489457607269287,
      "learning_rate": 1e-05,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8417967557907104,
      "num_tokens": 389778197.0,
      "step": 2446
    },
    {
      "epoch": 1.2446592065106816,
      "grad_norm": 0.8921250700950623,
      "learning_rate": 1e-05,
      "loss": 0.507,
      "mean_token_accuracy": 0.8411307334899902,
      "num_tokens": 389946253.0,
      "step": 2447
    },
    {
      "epoch": 1.2451678535096642,
      "grad_norm": 0.9162742495536804,
      "learning_rate": 1e-05,
      "loss": 0.4587,
      "mean_token_accuracy": 0.8542463779449463,
      "num_tokens": 390102274.0,
      "step": 2448
    },
    {
      "epoch": 1.245676500508647,
      "grad_norm": 1.0814255475997925,
      "learning_rate": 1e-05,
      "loss": 0.4908,
      "mean_token_accuracy": 0.8456716537475586,
      "num_tokens": 390252711.0,
      "step": 2449
    },
    {
      "epoch": 1.2461851475076298,
      "grad_norm": 0.9779260754585266,
      "learning_rate": 1e-05,
      "loss": 0.5072,
      "mean_token_accuracy": 0.8394188284873962,
      "num_tokens": 390422450.0,
      "step": 2450
    },
    {
      "epoch": 1.2466937945066123,
      "grad_norm": 0.9985749125480652,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8499858975410461,
      "num_tokens": 390587459.0,
      "step": 2451
    },
    {
      "epoch": 1.2472024415055951,
      "grad_norm": 1.0046234130859375,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8494718074798584,
      "num_tokens": 390752194.0,
      "step": 2452
    },
    {
      "epoch": 1.247711088504578,
      "grad_norm": 1.045386791229248,
      "learning_rate": 1e-05,
      "loss": 0.4826,
      "mean_token_accuracy": 0.8462944030761719,
      "num_tokens": 390918814.0,
      "step": 2453
    },
    {
      "epoch": 1.2482197355035605,
      "grad_norm": 0.9235852360725403,
      "learning_rate": 1e-05,
      "loss": 0.4547,
      "mean_token_accuracy": 0.8562573194503784,
      "num_tokens": 391076755.0,
      "step": 2454
    },
    {
      "epoch": 1.2487283825025433,
      "grad_norm": 0.9939376711845398,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8503568768501282,
      "num_tokens": 391234191.0,
      "step": 2455
    },
    {
      "epoch": 1.249237029501526,
      "grad_norm": 1.0112658739089966,
      "learning_rate": 1e-05,
      "loss": 0.4806,
      "mean_token_accuracy": 0.847571611404419,
      "num_tokens": 391386429.0,
      "step": 2456
    },
    {
      "epoch": 1.2497456765005086,
      "grad_norm": 0.9470198750495911,
      "learning_rate": 1e-05,
      "loss": 0.4651,
      "mean_token_accuracy": 0.849887490272522,
      "num_tokens": 391554073.0,
      "step": 2457
    },
    {
      "epoch": 1.2502543234994914,
      "grad_norm": 1.2515413761138916,
      "learning_rate": 1e-05,
      "loss": 0.5176,
      "mean_token_accuracy": 0.8373808860778809,
      "num_tokens": 391709893.0,
      "step": 2458
    },
    {
      "epoch": 1.250762970498474,
      "grad_norm": 1.092207908630371,
      "learning_rate": 1e-05,
      "loss": 0.5099,
      "mean_token_accuracy": 0.837600827217102,
      "num_tokens": 391870192.0,
      "step": 2459
    },
    {
      "epoch": 1.2512716174974567,
      "grad_norm": 0.9322696328163147,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8519747257232666,
      "num_tokens": 392024517.0,
      "step": 2460
    },
    {
      "epoch": 1.2517802644964395,
      "grad_norm": 0.991862416267395,
      "learning_rate": 1e-05,
      "loss": 0.4607,
      "mean_token_accuracy": 0.851301908493042,
      "num_tokens": 392177494.0,
      "step": 2461
    },
    {
      "epoch": 1.252288911495422,
      "grad_norm": 1.0980491638183594,
      "learning_rate": 1e-05,
      "loss": 0.4835,
      "mean_token_accuracy": 0.8465778827667236,
      "num_tokens": 392336854.0,
      "step": 2462
    },
    {
      "epoch": 1.2527975584944049,
      "grad_norm": 0.9788417816162109,
      "learning_rate": 1e-05,
      "loss": 0.4925,
      "mean_token_accuracy": 0.8434962630271912,
      "num_tokens": 392496140.0,
      "step": 2463
    },
    {
      "epoch": 1.2533062054933877,
      "grad_norm": 0.9595546722412109,
      "learning_rate": 1e-05,
      "loss": 0.4907,
      "mean_token_accuracy": 0.8437904119491577,
      "num_tokens": 392656687.0,
      "step": 2464
    },
    {
      "epoch": 1.2538148524923702,
      "grad_norm": 0.9926989674568176,
      "learning_rate": 1e-05,
      "loss": 0.5189,
      "mean_token_accuracy": 0.836510419845581,
      "num_tokens": 392812396.0,
      "step": 2465
    },
    {
      "epoch": 1.254323499491353,
      "grad_norm": 1.0025798082351685,
      "learning_rate": 1e-05,
      "loss": 0.5382,
      "mean_token_accuracy": 0.8307243585586548,
      "num_tokens": 392972527.0,
      "step": 2466
    },
    {
      "epoch": 1.2548321464903358,
      "grad_norm": 0.9914138913154602,
      "learning_rate": 1e-05,
      "loss": 0.4762,
      "mean_token_accuracy": 0.8475267291069031,
      "num_tokens": 393135630.0,
      "step": 2467
    },
    {
      "epoch": 1.2553407934893184,
      "grad_norm": 0.9821697473526001,
      "learning_rate": 1e-05,
      "loss": 0.4857,
      "mean_token_accuracy": 0.8450480699539185,
      "num_tokens": 393296349.0,
      "step": 2468
    },
    {
      "epoch": 1.2558494404883012,
      "grad_norm": 0.9580907821655273,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8602911233901978,
      "num_tokens": 393459529.0,
      "step": 2469
    },
    {
      "epoch": 1.2563580874872837,
      "grad_norm": 0.9034733772277832,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8528581857681274,
      "num_tokens": 393622385.0,
      "step": 2470
    },
    {
      "epoch": 1.2568667344862665,
      "grad_norm": 1.0865966081619263,
      "learning_rate": 1e-05,
      "loss": 0.4819,
      "mean_token_accuracy": 0.8464517593383789,
      "num_tokens": 393786264.0,
      "step": 2471
    },
    {
      "epoch": 1.2573753814852493,
      "grad_norm": 0.9148275852203369,
      "learning_rate": 1e-05,
      "loss": 0.4678,
      "mean_token_accuracy": 0.8529778718948364,
      "num_tokens": 393940347.0,
      "step": 2472
    },
    {
      "epoch": 1.2578840284842319,
      "grad_norm": 0.9414737820625305,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8517553210258484,
      "num_tokens": 394104484.0,
      "step": 2473
    },
    {
      "epoch": 1.2583926754832147,
      "grad_norm": 1.1321500539779663,
      "learning_rate": 1e-05,
      "loss": 0.5036,
      "mean_token_accuracy": 0.8403670191764832,
      "num_tokens": 394274909.0,
      "step": 2474
    },
    {
      "epoch": 1.2589013224821974,
      "grad_norm": 1.0167399644851685,
      "learning_rate": 1e-05,
      "loss": 0.4441,
      "mean_token_accuracy": 0.8573471307754517,
      "num_tokens": 394422663.0,
      "step": 2475
    },
    {
      "epoch": 1.25940996948118,
      "grad_norm": 1.066833257675171,
      "learning_rate": 1e-05,
      "loss": 0.4819,
      "mean_token_accuracy": 0.8465922474861145,
      "num_tokens": 394572404.0,
      "step": 2476
    },
    {
      "epoch": 1.2599186164801628,
      "grad_norm": 1.0820730924606323,
      "learning_rate": 1e-05,
      "loss": 0.5148,
      "mean_token_accuracy": 0.8378000855445862,
      "num_tokens": 394741473.0,
      "step": 2477
    },
    {
      "epoch": 1.2604272634791456,
      "grad_norm": 0.9718859791755676,
      "learning_rate": 1e-05,
      "loss": 0.5259,
      "mean_token_accuracy": 0.8359262943267822,
      "num_tokens": 394904157.0,
      "step": 2478
    },
    {
      "epoch": 1.2609359104781281,
      "grad_norm": 1.038017988204956,
      "learning_rate": 1e-05,
      "loss": 0.4661,
      "mean_token_accuracy": 0.8514251112937927,
      "num_tokens": 395046309.0,
      "step": 2479
    },
    {
      "epoch": 1.261444557477111,
      "grad_norm": 1.0430866479873657,
      "learning_rate": 1e-05,
      "loss": 0.4947,
      "mean_token_accuracy": 0.8451223373413086,
      "num_tokens": 395205744.0,
      "step": 2480
    },
    {
      "epoch": 1.2619532044760935,
      "grad_norm": 0.9635177850723267,
      "learning_rate": 1e-05,
      "loss": 0.4728,
      "mean_token_accuracy": 0.8498380184173584,
      "num_tokens": 395370214.0,
      "step": 2481
    },
    {
      "epoch": 1.2624618514750763,
      "grad_norm": 1.0214176177978516,
      "learning_rate": 1e-05,
      "loss": 0.4738,
      "mean_token_accuracy": 0.8482320308685303,
      "num_tokens": 395535591.0,
      "step": 2482
    },
    {
      "epoch": 1.262970498474059,
      "grad_norm": 0.9416934847831726,
      "learning_rate": 1e-05,
      "loss": 0.4763,
      "mean_token_accuracy": 0.8485772609710693,
      "num_tokens": 395720117.0,
      "step": 2483
    },
    {
      "epoch": 1.2634791454730416,
      "grad_norm": 1.0123696327209473,
      "learning_rate": 1e-05,
      "loss": 0.5005,
      "mean_token_accuracy": 0.8418780565261841,
      "num_tokens": 395887109.0,
      "step": 2484
    },
    {
      "epoch": 1.2639877924720244,
      "grad_norm": 0.9469811916351318,
      "learning_rate": 1e-05,
      "loss": 0.513,
      "mean_token_accuracy": 0.8381527662277222,
      "num_tokens": 396064517.0,
      "step": 2485
    },
    {
      "epoch": 1.264496439471007,
      "grad_norm": 1.1035939455032349,
      "learning_rate": 1e-05,
      "loss": 0.4926,
      "mean_token_accuracy": 0.843600869178772,
      "num_tokens": 396215248.0,
      "step": 2486
    },
    {
      "epoch": 1.2650050864699898,
      "grad_norm": 1.066406011581421,
      "learning_rate": 1e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8506541848182678,
      "num_tokens": 396380487.0,
      "step": 2487
    },
    {
      "epoch": 1.2655137334689726,
      "grad_norm": 0.9918091893196106,
      "learning_rate": 1e-05,
      "loss": 0.5373,
      "mean_token_accuracy": 0.8316949605941772,
      "num_tokens": 396538633.0,
      "step": 2488
    },
    {
      "epoch": 1.2660223804679553,
      "grad_norm": 1.0956774950027466,
      "learning_rate": 1e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.848578691482544,
      "num_tokens": 396707894.0,
      "step": 2489
    },
    {
      "epoch": 1.266531027466938,
      "grad_norm": 0.9060829877853394,
      "learning_rate": 1e-05,
      "loss": 0.4735,
      "mean_token_accuracy": 0.8491650819778442,
      "num_tokens": 396867308.0,
      "step": 2490
    },
    {
      "epoch": 1.2670396744659207,
      "grad_norm": 1.0590204000473022,
      "learning_rate": 1e-05,
      "loss": 0.5051,
      "mean_token_accuracy": 0.8396422863006592,
      "num_tokens": 397012467.0,
      "step": 2491
    },
    {
      "epoch": 1.2675483214649033,
      "grad_norm": 0.9565798044204712,
      "learning_rate": 1e-05,
      "loss": 0.4777,
      "mean_token_accuracy": 0.8488767743110657,
      "num_tokens": 397176821.0,
      "step": 2492
    },
    {
      "epoch": 1.268056968463886,
      "grad_norm": 0.990504801273346,
      "learning_rate": 1e-05,
      "loss": 0.4582,
      "mean_token_accuracy": 0.8547040820121765,
      "num_tokens": 397331212.0,
      "step": 2493
    },
    {
      "epoch": 1.2685656154628688,
      "grad_norm": 1.0066111087799072,
      "learning_rate": 1e-05,
      "loss": 0.4664,
      "mean_token_accuracy": 0.852090060710907,
      "num_tokens": 397507718.0,
      "step": 2494
    },
    {
      "epoch": 1.2690742624618514,
      "grad_norm": 0.9095807671546936,
      "learning_rate": 1e-05,
      "loss": 0.4625,
      "mean_token_accuracy": 0.8523800373077393,
      "num_tokens": 397664472.0,
      "step": 2495
    },
    {
      "epoch": 1.2695829094608342,
      "grad_norm": 1.028704047203064,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8492536544799805,
      "num_tokens": 397832455.0,
      "step": 2496
    },
    {
      "epoch": 1.2700915564598168,
      "grad_norm": 0.970306396484375,
      "learning_rate": 1e-05,
      "loss": 0.4922,
      "mean_token_accuracy": 0.8438721895217896,
      "num_tokens": 397993790.0,
      "step": 2497
    },
    {
      "epoch": 1.2706002034587995,
      "grad_norm": 1.730830192565918,
      "learning_rate": 1e-05,
      "loss": 0.4765,
      "mean_token_accuracy": 0.8479042053222656,
      "num_tokens": 398143376.0,
      "step": 2498
    },
    {
      "epoch": 1.2711088504577823,
      "grad_norm": 0.9698526859283447,
      "learning_rate": 1e-05,
      "loss": 0.4615,
      "mean_token_accuracy": 0.8523569703102112,
      "num_tokens": 398307832.0,
      "step": 2499
    },
    {
      "epoch": 1.2716174974567651,
      "grad_norm": 0.9873821139335632,
      "learning_rate": 1e-05,
      "loss": 0.479,
      "mean_token_accuracy": 0.84932541847229,
      "num_tokens": 398461934.0,
      "step": 2500
    },
    {
      "epoch": 1.2721261444557477,
      "grad_norm": 1.029078483581543,
      "learning_rate": 1e-05,
      "loss": 0.4955,
      "mean_token_accuracy": 0.8425160646438599,
      "num_tokens": 398614969.0,
      "step": 2501
    },
    {
      "epoch": 1.2726347914547305,
      "grad_norm": 0.9706425666809082,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8488206267356873,
      "num_tokens": 398775458.0,
      "step": 2502
    },
    {
      "epoch": 1.273143438453713,
      "grad_norm": 1.0119162797927856,
      "learning_rate": 1e-05,
      "loss": 0.5239,
      "mean_token_accuracy": 0.8351165652275085,
      "num_tokens": 398931322.0,
      "step": 2503
    },
    {
      "epoch": 1.2736520854526958,
      "grad_norm": 0.9316058158874512,
      "learning_rate": 1e-05,
      "loss": 0.4926,
      "mean_token_accuracy": 0.8441140055656433,
      "num_tokens": 399094210.0,
      "step": 2504
    },
    {
      "epoch": 1.2741607324516786,
      "grad_norm": 1.0092213153839111,
      "learning_rate": 1e-05,
      "loss": 0.4898,
      "mean_token_accuracy": 0.8453854322433472,
      "num_tokens": 399251980.0,
      "step": 2505
    },
    {
      "epoch": 1.2746693794506612,
      "grad_norm": 0.9519758820533752,
      "learning_rate": 1e-05,
      "loss": 0.4617,
      "mean_token_accuracy": 0.8530810475349426,
      "num_tokens": 399408637.0,
      "step": 2506
    },
    {
      "epoch": 1.275178026449644,
      "grad_norm": 1.0151798725128174,
      "learning_rate": 1e-05,
      "loss": 0.4917,
      "mean_token_accuracy": 0.8444891571998596,
      "num_tokens": 399567395.0,
      "step": 2507
    },
    {
      "epoch": 1.2756866734486265,
      "grad_norm": 1.0232245922088623,
      "learning_rate": 1e-05,
      "loss": 0.4792,
      "mean_token_accuracy": 0.848068356513977,
      "num_tokens": 399726159.0,
      "step": 2508
    },
    {
      "epoch": 1.2761953204476093,
      "grad_norm": 0.9277547597885132,
      "learning_rate": 1e-05,
      "loss": 0.4508,
      "mean_token_accuracy": 0.8560090065002441,
      "num_tokens": 399888825.0,
      "step": 2509
    },
    {
      "epoch": 1.276703967446592,
      "grad_norm": 0.9623938202857971,
      "learning_rate": 1e-05,
      "loss": 0.4949,
      "mean_token_accuracy": 0.8442527651786804,
      "num_tokens": 400058221.0,
      "step": 2510
    },
    {
      "epoch": 1.2772126144455749,
      "grad_norm": 0.9173216819763184,
      "learning_rate": 1e-05,
      "loss": 0.4642,
      "mean_token_accuracy": 0.8523637652397156,
      "num_tokens": 400223047.0,
      "step": 2511
    },
    {
      "epoch": 1.2777212614445574,
      "grad_norm": 0.9462375640869141,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8448326587677002,
      "num_tokens": 400381507.0,
      "step": 2512
    },
    {
      "epoch": 1.2782299084435402,
      "grad_norm": 0.9886204600334167,
      "learning_rate": 1e-05,
      "loss": 0.4859,
      "mean_token_accuracy": 0.846415102481842,
      "num_tokens": 400546494.0,
      "step": 2513
    },
    {
      "epoch": 1.2787385554425228,
      "grad_norm": 1.0517821311950684,
      "learning_rate": 1e-05,
      "loss": 0.4728,
      "mean_token_accuracy": 0.8493285179138184,
      "num_tokens": 400704230.0,
      "step": 2514
    },
    {
      "epoch": 1.2792472024415056,
      "grad_norm": 0.9828445911407471,
      "learning_rate": 1e-05,
      "loss": 0.498,
      "mean_token_accuracy": 0.8424510359764099,
      "num_tokens": 400862305.0,
      "step": 2515
    },
    {
      "epoch": 1.2797558494404884,
      "grad_norm": 0.9465426802635193,
      "learning_rate": 1e-05,
      "loss": 0.4534,
      "mean_token_accuracy": 0.8549497127532959,
      "num_tokens": 401018958.0,
      "step": 2516
    },
    {
      "epoch": 1.280264496439471,
      "grad_norm": 0.9962822794914246,
      "learning_rate": 1e-05,
      "loss": 0.506,
      "mean_token_accuracy": 0.8403021097183228,
      "num_tokens": 401178895.0,
      "step": 2517
    },
    {
      "epoch": 1.2807731434384537,
      "grad_norm": 0.9522839784622192,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8494329452514648,
      "num_tokens": 401325913.0,
      "step": 2518
    },
    {
      "epoch": 1.2812817904374363,
      "grad_norm": 0.9846671223640442,
      "learning_rate": 1e-05,
      "loss": 0.4927,
      "mean_token_accuracy": 0.8454416394233704,
      "num_tokens": 401492299.0,
      "step": 2519
    },
    {
      "epoch": 1.281790437436419,
      "grad_norm": 0.9660159945487976,
      "learning_rate": 1e-05,
      "loss": 0.5084,
      "mean_token_accuracy": 0.838665246963501,
      "num_tokens": 401650213.0,
      "step": 2520
    },
    {
      "epoch": 1.2822990844354019,
      "grad_norm": 1.0401360988616943,
      "learning_rate": 1e-05,
      "loss": 0.4862,
      "mean_token_accuracy": 0.8451883792877197,
      "num_tokens": 401803969.0,
      "step": 2521
    },
    {
      "epoch": 1.2828077314343846,
      "grad_norm": 0.9664472937583923,
      "learning_rate": 1e-05,
      "loss": 0.4705,
      "mean_token_accuracy": 0.8487030267715454,
      "num_tokens": 401952668.0,
      "step": 2522
    },
    {
      "epoch": 1.2833163784333672,
      "grad_norm": 0.9471451640129089,
      "learning_rate": 1e-05,
      "loss": 0.4936,
      "mean_token_accuracy": 0.8442484736442566,
      "num_tokens": 402114443.0,
      "step": 2523
    },
    {
      "epoch": 1.28382502543235,
      "grad_norm": 0.9966514706611633,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.857598066329956,
      "num_tokens": 402266408.0,
      "step": 2524
    },
    {
      "epoch": 1.2843336724313326,
      "grad_norm": 0.9822807908058167,
      "learning_rate": 1e-05,
      "loss": 0.529,
      "mean_token_accuracy": 0.8337392807006836,
      "num_tokens": 402435949.0,
      "step": 2525
    },
    {
      "epoch": 1.2848423194303153,
      "grad_norm": 1.0041320323944092,
      "learning_rate": 1e-05,
      "loss": 0.4955,
      "mean_token_accuracy": 0.8431259989738464,
      "num_tokens": 402595965.0,
      "step": 2526
    },
    {
      "epoch": 1.2853509664292981,
      "grad_norm": 1.0624803304672241,
      "learning_rate": 1e-05,
      "loss": 0.5135,
      "mean_token_accuracy": 0.8371453285217285,
      "num_tokens": 402757273.0,
      "step": 2527
    },
    {
      "epoch": 1.2858596134282807,
      "grad_norm": 1.093480110168457,
      "learning_rate": 1e-05,
      "loss": 0.5032,
      "mean_token_accuracy": 0.8406103849411011,
      "num_tokens": 402919556.0,
      "step": 2528
    },
    {
      "epoch": 1.2863682604272635,
      "grad_norm": 0.9978317618370056,
      "learning_rate": 1e-05,
      "loss": 0.4968,
      "mean_token_accuracy": 0.8422409892082214,
      "num_tokens": 403073153.0,
      "step": 2529
    },
    {
      "epoch": 1.286876907426246,
      "grad_norm": 1.011997103691101,
      "learning_rate": 1e-05,
      "loss": 0.4662,
      "mean_token_accuracy": 0.8520162105560303,
      "num_tokens": 403234129.0,
      "step": 2530
    },
    {
      "epoch": 1.2873855544252288,
      "grad_norm": 0.9961965680122375,
      "learning_rate": 1e-05,
      "loss": 0.4843,
      "mean_token_accuracy": 0.8475301265716553,
      "num_tokens": 403389659.0,
      "step": 2531
    },
    {
      "epoch": 1.2878942014242116,
      "grad_norm": 0.9876997470855713,
      "learning_rate": 1e-05,
      "loss": 0.5039,
      "mean_token_accuracy": 0.8410624265670776,
      "num_tokens": 403545378.0,
      "step": 2532
    },
    {
      "epoch": 1.2884028484231944,
      "grad_norm": 0.9761861562728882,
      "learning_rate": 1e-05,
      "loss": 0.4943,
      "mean_token_accuracy": 0.8429926633834839,
      "num_tokens": 403721714.0,
      "step": 2533
    },
    {
      "epoch": 1.288911495422177,
      "grad_norm": 1.060185194015503,
      "learning_rate": 1e-05,
      "loss": 0.5029,
      "mean_token_accuracy": 0.8406179547309875,
      "num_tokens": 403879895.0,
      "step": 2534
    },
    {
      "epoch": 1.2894201424211598,
      "grad_norm": 0.9180830121040344,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "mean_token_accuracy": 0.857403039932251,
      "num_tokens": 404039422.0,
      "step": 2535
    },
    {
      "epoch": 1.2899287894201423,
      "grad_norm": 0.9527869820594788,
      "learning_rate": 1e-05,
      "loss": 0.4928,
      "mean_token_accuracy": 0.8432321548461914,
      "num_tokens": 404194940.0,
      "step": 2536
    },
    {
      "epoch": 1.2904374364191251,
      "grad_norm": 1.004622459411621,
      "learning_rate": 1e-05,
      "loss": 0.544,
      "mean_token_accuracy": 0.8304028511047363,
      "num_tokens": 404356690.0,
      "step": 2537
    },
    {
      "epoch": 1.290946083418108,
      "grad_norm": 0.9781545400619507,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8512308597564697,
      "num_tokens": 404516058.0,
      "step": 2538
    },
    {
      "epoch": 1.2914547304170905,
      "grad_norm": 1.0009372234344482,
      "learning_rate": 1e-05,
      "loss": 0.5014,
      "mean_token_accuracy": 0.8395940065383911,
      "num_tokens": 404673804.0,
      "step": 2539
    },
    {
      "epoch": 1.2919633774160733,
      "grad_norm": 1.014769196510315,
      "learning_rate": 1e-05,
      "loss": 0.4988,
      "mean_token_accuracy": 0.8452779054641724,
      "num_tokens": 404823463.0,
      "step": 2540
    },
    {
      "epoch": 1.2924720244150558,
      "grad_norm": 1.0117167234420776,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8504893183708191,
      "num_tokens": 404978592.0,
      "step": 2541
    },
    {
      "epoch": 1.2929806714140386,
      "grad_norm": 1.0531048774719238,
      "learning_rate": 1e-05,
      "loss": 0.5031,
      "mean_token_accuracy": 0.8393880128860474,
      "num_tokens": 405133209.0,
      "step": 2542
    },
    {
      "epoch": 1.2934893184130214,
      "grad_norm": 0.986916184425354,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8475457429885864,
      "num_tokens": 405291663.0,
      "step": 2543
    },
    {
      "epoch": 1.2939979654120042,
      "grad_norm": 0.9645692706108093,
      "learning_rate": 1e-05,
      "loss": 0.4935,
      "mean_token_accuracy": 0.8424645662307739,
      "num_tokens": 405447583.0,
      "step": 2544
    },
    {
      "epoch": 1.2945066124109867,
      "grad_norm": 1.0199670791625977,
      "learning_rate": 1e-05,
      "loss": 0.48,
      "mean_token_accuracy": 0.8481231927871704,
      "num_tokens": 405609401.0,
      "step": 2545
    },
    {
      "epoch": 1.2950152594099695,
      "grad_norm": 1.0394608974456787,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.8566803932189941,
      "num_tokens": 405772749.0,
      "step": 2546
    },
    {
      "epoch": 1.295523906408952,
      "grad_norm": 1.018204689025879,
      "learning_rate": 1e-05,
      "loss": 0.465,
      "mean_token_accuracy": 0.8515591025352478,
      "num_tokens": 405944255.0,
      "step": 2547
    },
    {
      "epoch": 1.2960325534079349,
      "grad_norm": 1.1013845205307007,
      "learning_rate": 1e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8453958034515381,
      "num_tokens": 406102126.0,
      "step": 2548
    },
    {
      "epoch": 1.2965412004069177,
      "grad_norm": 0.9380601644515991,
      "learning_rate": 1e-05,
      "loss": 0.4628,
      "mean_token_accuracy": 0.8526260256767273,
      "num_tokens": 406265194.0,
      "step": 2549
    },
    {
      "epoch": 1.2970498474059002,
      "grad_norm": 1.021324634552002,
      "learning_rate": 1e-05,
      "loss": 0.5045,
      "mean_token_accuracy": 0.8427411317825317,
      "num_tokens": 406423086.0,
      "step": 2550
    },
    {
      "epoch": 1.297558494404883,
      "grad_norm": 1.0104503631591797,
      "learning_rate": 1e-05,
      "loss": 0.4555,
      "mean_token_accuracy": 0.8549754619598389,
      "num_tokens": 406587567.0,
      "step": 2551
    },
    {
      "epoch": 1.2980671414038656,
      "grad_norm": 0.9233775734901428,
      "learning_rate": 1e-05,
      "loss": 0.4687,
      "mean_token_accuracy": 0.8530840873718262,
      "num_tokens": 406744290.0,
      "step": 2552
    },
    {
      "epoch": 1.2985757884028484,
      "grad_norm": 1.0981473922729492,
      "learning_rate": 1e-05,
      "loss": 0.4951,
      "mean_token_accuracy": 0.8443585634231567,
      "num_tokens": 406899607.0,
      "step": 2553
    },
    {
      "epoch": 1.2990844354018312,
      "grad_norm": 1.0113608837127686,
      "learning_rate": 1e-05,
      "loss": 0.4793,
      "mean_token_accuracy": 0.8468562960624695,
      "num_tokens": 407056075.0,
      "step": 2554
    },
    {
      "epoch": 1.299593082400814,
      "grad_norm": 0.9467877149581909,
      "learning_rate": 1e-05,
      "loss": 0.4814,
      "mean_token_accuracy": 0.8476945161819458,
      "num_tokens": 407213538.0,
      "step": 2555
    },
    {
      "epoch": 1.3001017293997965,
      "grad_norm": 1.080271601676941,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.841745138168335,
      "num_tokens": 407364405.0,
      "step": 2556
    },
    {
      "epoch": 1.3006103763987793,
      "grad_norm": 1.006981611251831,
      "learning_rate": 1e-05,
      "loss": 0.4803,
      "mean_token_accuracy": 0.8464992642402649,
      "num_tokens": 407530938.0,
      "step": 2557
    },
    {
      "epoch": 1.3011190233977619,
      "grad_norm": 0.956074595451355,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8508355021476746,
      "num_tokens": 407689878.0,
      "step": 2558
    },
    {
      "epoch": 1.3016276703967447,
      "grad_norm": 0.9157811999320984,
      "learning_rate": 1e-05,
      "loss": 0.4618,
      "mean_token_accuracy": 0.8531558513641357,
      "num_tokens": 407842592.0,
      "step": 2559
    },
    {
      "epoch": 1.3021363173957274,
      "grad_norm": 0.9051852822303772,
      "learning_rate": 1e-05,
      "loss": 0.498,
      "mean_token_accuracy": 0.8445741534233093,
      "num_tokens": 408019653.0,
      "step": 2560
    },
    {
      "epoch": 1.30264496439471,
      "grad_norm": 0.8561477661132812,
      "learning_rate": 1e-05,
      "loss": 0.4602,
      "mean_token_accuracy": 0.8530067205429077,
      "num_tokens": 408182639.0,
      "step": 2561
    },
    {
      "epoch": 1.3031536113936928,
      "grad_norm": 0.9158396124839783,
      "learning_rate": 1e-05,
      "loss": 0.4943,
      "mean_token_accuracy": 0.8439619541168213,
      "num_tokens": 408343205.0,
      "step": 2562
    },
    {
      "epoch": 1.3036622583926754,
      "grad_norm": 0.940427839756012,
      "learning_rate": 1e-05,
      "loss": 0.4771,
      "mean_token_accuracy": 0.8503564596176147,
      "num_tokens": 408505934.0,
      "step": 2563
    },
    {
      "epoch": 1.3041709053916581,
      "grad_norm": 0.9462770223617554,
      "learning_rate": 1e-05,
      "loss": 0.4903,
      "mean_token_accuracy": 0.8440700173377991,
      "num_tokens": 408666686.0,
      "step": 2564
    },
    {
      "epoch": 1.304679552390641,
      "grad_norm": 0.9128457903862,
      "learning_rate": 1e-05,
      "loss": 0.4983,
      "mean_token_accuracy": 0.8417966365814209,
      "num_tokens": 408829260.0,
      "step": 2565
    },
    {
      "epoch": 1.3051881993896237,
      "grad_norm": 0.95364910364151,
      "learning_rate": 1e-05,
      "loss": 0.4976,
      "mean_token_accuracy": 0.8429286479949951,
      "num_tokens": 408989130.0,
      "step": 2566
    },
    {
      "epoch": 1.3056968463886063,
      "grad_norm": 1.0803797245025635,
      "learning_rate": 1e-05,
      "loss": 0.4891,
      "mean_token_accuracy": 0.8446954488754272,
      "num_tokens": 409160100.0,
      "step": 2567
    },
    {
      "epoch": 1.306205493387589,
      "grad_norm": 0.9539232850074768,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8547990918159485,
      "num_tokens": 409319697.0,
      "step": 2568
    },
    {
      "epoch": 1.3067141403865716,
      "grad_norm": 1.019270896911621,
      "learning_rate": 1e-05,
      "loss": 0.4858,
      "mean_token_accuracy": 0.8463223576545715,
      "num_tokens": 409473663.0,
      "step": 2569
    },
    {
      "epoch": 1.3072227873855544,
      "grad_norm": 0.9860494136810303,
      "learning_rate": 1e-05,
      "loss": 0.4802,
      "mean_token_accuracy": 0.8477706909179688,
      "num_tokens": 409638569.0,
      "step": 2570
    },
    {
      "epoch": 1.3077314343845372,
      "grad_norm": 0.9673585295677185,
      "learning_rate": 1e-05,
      "loss": 0.4917,
      "mean_token_accuracy": 0.8434401750564575,
      "num_tokens": 409802014.0,
      "step": 2571
    },
    {
      "epoch": 1.3082400813835198,
      "grad_norm": 1.1659724712371826,
      "learning_rate": 1e-05,
      "loss": 0.4864,
      "mean_token_accuracy": 0.8469688892364502,
      "num_tokens": 409963322.0,
      "step": 2572
    },
    {
      "epoch": 1.3087487283825026,
      "grad_norm": 0.9789167642593384,
      "learning_rate": 1e-05,
      "loss": 0.4718,
      "mean_token_accuracy": 0.8485277891159058,
      "num_tokens": 410124722.0,
      "step": 2573
    },
    {
      "epoch": 1.3092573753814851,
      "grad_norm": 0.8991252183914185,
      "learning_rate": 1e-05,
      "loss": 0.4293,
      "mean_token_accuracy": 0.8623579144477844,
      "num_tokens": 410289774.0,
      "step": 2574
    },
    {
      "epoch": 1.309766022380468,
      "grad_norm": 0.9171684980392456,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8557093739509583,
      "num_tokens": 410456209.0,
      "step": 2575
    },
    {
      "epoch": 1.3102746693794507,
      "grad_norm": 1.0491565465927124,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8587504625320435,
      "num_tokens": 410603142.0,
      "step": 2576
    },
    {
      "epoch": 1.3107833163784335,
      "grad_norm": 1.041669487953186,
      "learning_rate": 1e-05,
      "loss": 0.466,
      "mean_token_accuracy": 0.852723240852356,
      "num_tokens": 410770784.0,
      "step": 2577
    },
    {
      "epoch": 1.311291963377416,
      "grad_norm": 0.9561178684234619,
      "learning_rate": 1e-05,
      "loss": 0.5213,
      "mean_token_accuracy": 0.8371676802635193,
      "num_tokens": 410934281.0,
      "step": 2578
    },
    {
      "epoch": 1.3118006103763988,
      "grad_norm": 0.9502045512199402,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.8432497382164001,
      "num_tokens": 411097001.0,
      "step": 2579
    },
    {
      "epoch": 1.3123092573753814,
      "grad_norm": 1.0522953271865845,
      "learning_rate": 1e-05,
      "loss": 0.4945,
      "mean_token_accuracy": 0.8432407975196838,
      "num_tokens": 411262888.0,
      "step": 2580
    },
    {
      "epoch": 1.3128179043743642,
      "grad_norm": 0.9856316447257996,
      "learning_rate": 1e-05,
      "loss": 0.4722,
      "mean_token_accuracy": 0.8499524593353271,
      "num_tokens": 411412576.0,
      "step": 2581
    },
    {
      "epoch": 1.313326551373347,
      "grad_norm": 0.9543809294700623,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8511854410171509,
      "num_tokens": 411567594.0,
      "step": 2582
    },
    {
      "epoch": 1.3138351983723295,
      "grad_norm": 0.9256858229637146,
      "learning_rate": 1e-05,
      "loss": 0.515,
      "mean_token_accuracy": 0.8376957178115845,
      "num_tokens": 411735468.0,
      "step": 2583
    },
    {
      "epoch": 1.3143438453713123,
      "grad_norm": 0.9677025675773621,
      "learning_rate": 1e-05,
      "loss": 0.4797,
      "mean_token_accuracy": 0.8493592143058777,
      "num_tokens": 411893042.0,
      "step": 2584
    },
    {
      "epoch": 1.314852492370295,
      "grad_norm": 0.9563512206077576,
      "learning_rate": 1e-05,
      "loss": 0.4854,
      "mean_token_accuracy": 0.846301794052124,
      "num_tokens": 412061278.0,
      "step": 2585
    },
    {
      "epoch": 1.3153611393692777,
      "grad_norm": 0.9341158270835876,
      "learning_rate": 1e-05,
      "loss": 0.4609,
      "mean_token_accuracy": 0.8514420390129089,
      "num_tokens": 412215931.0,
      "step": 2586
    },
    {
      "epoch": 1.3158697863682605,
      "grad_norm": 0.9857239723205566,
      "learning_rate": 1e-05,
      "loss": 0.485,
      "mean_token_accuracy": 0.8454148173332214,
      "num_tokens": 412390962.0,
      "step": 2587
    },
    {
      "epoch": 1.3163784333672433,
      "grad_norm": 0.9439259171485901,
      "learning_rate": 1e-05,
      "loss": 0.4777,
      "mean_token_accuracy": 0.8475058674812317,
      "num_tokens": 412552715.0,
      "step": 2588
    },
    {
      "epoch": 1.3168870803662258,
      "grad_norm": 0.9452398419380188,
      "learning_rate": 1e-05,
      "loss": 0.4911,
      "mean_token_accuracy": 0.8462110757827759,
      "num_tokens": 412719801.0,
      "step": 2589
    },
    {
      "epoch": 1.3173957273652086,
      "grad_norm": 0.9879827499389648,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.844982385635376,
      "num_tokens": 412869340.0,
      "step": 2590
    },
    {
      "epoch": 1.3179043743641912,
      "grad_norm": 0.9580371975898743,
      "learning_rate": 1e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.8534672260284424,
      "num_tokens": 413025251.0,
      "step": 2591
    },
    {
      "epoch": 1.318413021363174,
      "grad_norm": 0.9868455529212952,
      "learning_rate": 1e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8489521741867065,
      "num_tokens": 413187434.0,
      "step": 2592
    },
    {
      "epoch": 1.3189216683621567,
      "grad_norm": 1.007317304611206,
      "learning_rate": 1e-05,
      "loss": 0.5096,
      "mean_token_accuracy": 0.8397385478019714,
      "num_tokens": 413359026.0,
      "step": 2593
    },
    {
      "epoch": 1.3194303153611393,
      "grad_norm": 0.9267711639404297,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.8407220244407654,
      "num_tokens": 413523255.0,
      "step": 2594
    },
    {
      "epoch": 1.319938962360122,
      "grad_norm": 0.9290441870689392,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.8536738157272339,
      "num_tokens": 413677670.0,
      "step": 2595
    },
    {
      "epoch": 1.3204476093591047,
      "grad_norm": 0.9466893672943115,
      "learning_rate": 1e-05,
      "loss": 0.5042,
      "mean_token_accuracy": 0.8413321375846863,
      "num_tokens": 413836107.0,
      "step": 2596
    },
    {
      "epoch": 1.3209562563580874,
      "grad_norm": 0.8821899890899658,
      "learning_rate": 1e-05,
      "loss": 0.4428,
      "mean_token_accuracy": 0.8579069972038269,
      "num_tokens": 414000734.0,
      "step": 2597
    },
    {
      "epoch": 1.3214649033570702,
      "grad_norm": 0.9268909692764282,
      "learning_rate": 1e-05,
      "loss": 0.4694,
      "mean_token_accuracy": 0.8505789637565613,
      "num_tokens": 414156938.0,
      "step": 2598
    },
    {
      "epoch": 1.321973550356053,
      "grad_norm": 0.9249274730682373,
      "learning_rate": 1e-05,
      "loss": 0.4856,
      "mean_token_accuracy": 0.847085177898407,
      "num_tokens": 414318719.0,
      "step": 2599
    },
    {
      "epoch": 1.3224821973550356,
      "grad_norm": 0.9466115236282349,
      "learning_rate": 1e-05,
      "loss": 0.5139,
      "mean_token_accuracy": 0.8403315544128418,
      "num_tokens": 414482061.0,
      "step": 2600
    },
    {
      "epoch": 1.3229908443540184,
      "grad_norm": 0.9923340082168579,
      "learning_rate": 1e-05,
      "loss": 0.4805,
      "mean_token_accuracy": 0.8463633060455322,
      "num_tokens": 414637332.0,
      "step": 2601
    },
    {
      "epoch": 1.323499491353001,
      "grad_norm": 0.8582419753074646,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8543981313705444,
      "num_tokens": 414806350.0,
      "step": 2602
    },
    {
      "epoch": 1.3240081383519837,
      "grad_norm": 0.9188969135284424,
      "learning_rate": 1e-05,
      "loss": 0.5154,
      "mean_token_accuracy": 0.8376389145851135,
      "num_tokens": 414976420.0,
      "step": 2603
    },
    {
      "epoch": 1.3245167853509665,
      "grad_norm": 0.9199648499488831,
      "learning_rate": 1e-05,
      "loss": 0.5053,
      "mean_token_accuracy": 0.8413670659065247,
      "num_tokens": 415151776.0,
      "step": 2604
    },
    {
      "epoch": 1.325025432349949,
      "grad_norm": 0.9353446960449219,
      "learning_rate": 1e-05,
      "loss": 0.4777,
      "mean_token_accuracy": 0.8477828502655029,
      "num_tokens": 415305958.0,
      "step": 2605
    },
    {
      "epoch": 1.3255340793489319,
      "grad_norm": 0.9904804229736328,
      "learning_rate": 1e-05,
      "loss": 0.4977,
      "mean_token_accuracy": 0.8437171578407288,
      "num_tokens": 415475717.0,
      "step": 2606
    },
    {
      "epoch": 1.3260427263479144,
      "grad_norm": 0.9145898818969727,
      "learning_rate": 1e-05,
      "loss": 0.4738,
      "mean_token_accuracy": 0.8503988981246948,
      "num_tokens": 415628323.0,
      "step": 2607
    },
    {
      "epoch": 1.3265513733468972,
      "grad_norm": 0.9652713537216187,
      "learning_rate": 1e-05,
      "loss": 0.4594,
      "mean_token_accuracy": 0.854482889175415,
      "num_tokens": 415782339.0,
      "step": 2608
    },
    {
      "epoch": 1.32706002034588,
      "grad_norm": 0.9327253103256226,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8490644693374634,
      "num_tokens": 415936748.0,
      "step": 2609
    },
    {
      "epoch": 1.3275686673448628,
      "grad_norm": 0.9666473269462585,
      "learning_rate": 1e-05,
      "loss": 0.4866,
      "mean_token_accuracy": 0.8455109596252441,
      "num_tokens": 416091384.0,
      "step": 2610
    },
    {
      "epoch": 1.3280773143438453,
      "grad_norm": 0.9348263144493103,
      "learning_rate": 1e-05,
      "loss": 0.4822,
      "mean_token_accuracy": 0.8458163738250732,
      "num_tokens": 416245335.0,
      "step": 2611
    },
    {
      "epoch": 1.3285859613428281,
      "grad_norm": 0.9533799290657043,
      "learning_rate": 1e-05,
      "loss": 0.5053,
      "mean_token_accuracy": 0.8398832082748413,
      "num_tokens": 416408034.0,
      "step": 2612
    },
    {
      "epoch": 1.3290946083418107,
      "grad_norm": 0.962467610836029,
      "learning_rate": 1e-05,
      "loss": 0.5113,
      "mean_token_accuracy": 0.8395194411277771,
      "num_tokens": 416567114.0,
      "step": 2613
    },
    {
      "epoch": 1.3296032553407935,
      "grad_norm": 0.9966299533843994,
      "learning_rate": 1e-05,
      "loss": 0.4661,
      "mean_token_accuracy": 0.8517531156539917,
      "num_tokens": 416715317.0,
      "step": 2614
    },
    {
      "epoch": 1.3301119023397763,
      "grad_norm": 0.993205726146698,
      "learning_rate": 1e-05,
      "loss": 0.4863,
      "mean_token_accuracy": 0.8464823365211487,
      "num_tokens": 416862929.0,
      "step": 2615
    },
    {
      "epoch": 1.3306205493387588,
      "grad_norm": 1.0152370929718018,
      "learning_rate": 1e-05,
      "loss": 0.477,
      "mean_token_accuracy": 0.8483762145042419,
      "num_tokens": 417017357.0,
      "step": 2616
    },
    {
      "epoch": 1.3311291963377416,
      "grad_norm": 0.9989106059074402,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.8536456227302551,
      "num_tokens": 417166390.0,
      "step": 2617
    },
    {
      "epoch": 1.3316378433367242,
      "grad_norm": 0.9126229882240295,
      "learning_rate": 1e-05,
      "loss": 0.4658,
      "mean_token_accuracy": 0.8509222269058228,
      "num_tokens": 417324031.0,
      "step": 2618
    },
    {
      "epoch": 1.332146490335707,
      "grad_norm": 0.9760608673095703,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.842339813709259,
      "num_tokens": 417488486.0,
      "step": 2619
    },
    {
      "epoch": 1.3326551373346898,
      "grad_norm": 0.9777554273605347,
      "learning_rate": 1e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.8518586754798889,
      "num_tokens": 417635848.0,
      "step": 2620
    },
    {
      "epoch": 1.3331637843336726,
      "grad_norm": 0.9069948196411133,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8558377027511597,
      "num_tokens": 417804152.0,
      "step": 2621
    },
    {
      "epoch": 1.3336724313326551,
      "grad_norm": 0.9307112693786621,
      "learning_rate": 1e-05,
      "loss": 0.4491,
      "mean_token_accuracy": 0.8567745685577393,
      "num_tokens": 417948063.0,
      "step": 2622
    },
    {
      "epoch": 1.334181078331638,
      "grad_norm": 0.938089907169342,
      "learning_rate": 1e-05,
      "loss": 0.5172,
      "mean_token_accuracy": 0.8359748721122742,
      "num_tokens": 418099041.0,
      "step": 2623
    },
    {
      "epoch": 1.3346897253306205,
      "grad_norm": 0.8834355473518372,
      "learning_rate": 1e-05,
      "loss": 0.5069,
      "mean_token_accuracy": 0.8408173322677612,
      "num_tokens": 418263839.0,
      "step": 2624
    },
    {
      "epoch": 1.3351983723296033,
      "grad_norm": 0.9467998743057251,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8598389625549316,
      "num_tokens": 418422529.0,
      "step": 2625
    },
    {
      "epoch": 1.335707019328586,
      "grad_norm": 0.9922293424606323,
      "learning_rate": 1e-05,
      "loss": 0.4842,
      "mean_token_accuracy": 0.8466914892196655,
      "num_tokens": 418582350.0,
      "step": 2626
    },
    {
      "epoch": 1.3362156663275686,
      "grad_norm": 0.8823071122169495,
      "learning_rate": 1e-05,
      "loss": 0.4842,
      "mean_token_accuracy": 0.8463080525398254,
      "num_tokens": 418747149.0,
      "step": 2627
    },
    {
      "epoch": 1.3367243133265514,
      "grad_norm": 0.908643364906311,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "mean_token_accuracy": 0.8584833741188049,
      "num_tokens": 418893429.0,
      "step": 2628
    },
    {
      "epoch": 1.337232960325534,
      "grad_norm": 0.9817284345626831,
      "learning_rate": 1e-05,
      "loss": 0.4935,
      "mean_token_accuracy": 0.8436906933784485,
      "num_tokens": 419062878.0,
      "step": 2629
    },
    {
      "epoch": 1.3377416073245167,
      "grad_norm": 0.9526888728141785,
      "learning_rate": 1e-05,
      "loss": 0.4648,
      "mean_token_accuracy": 0.8509718775749207,
      "num_tokens": 419228134.0,
      "step": 2630
    },
    {
      "epoch": 1.3382502543234995,
      "grad_norm": 1.013670563697815,
      "learning_rate": 1e-05,
      "loss": 0.5534,
      "mean_token_accuracy": 0.8273931741714478,
      "num_tokens": 419385610.0,
      "step": 2631
    },
    {
      "epoch": 1.3387589013224823,
      "grad_norm": 1.023470163345337,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.8471496105194092,
      "num_tokens": 419546274.0,
      "step": 2632
    },
    {
      "epoch": 1.3392675483214649,
      "grad_norm": 1.1507434844970703,
      "learning_rate": 1e-05,
      "loss": 0.5087,
      "mean_token_accuracy": 0.8395289182662964,
      "num_tokens": 419698007.0,
      "step": 2633
    },
    {
      "epoch": 1.3397761953204477,
      "grad_norm": 0.9994651079177856,
      "learning_rate": 1e-05,
      "loss": 0.4856,
      "mean_token_accuracy": 0.8446428179740906,
      "num_tokens": 419861881.0,
      "step": 2634
    },
    {
      "epoch": 1.3402848423194302,
      "grad_norm": 0.9943557381629944,
      "learning_rate": 1e-05,
      "loss": 0.4999,
      "mean_token_accuracy": 0.8429102897644043,
      "num_tokens": 420021371.0,
      "step": 2635
    },
    {
      "epoch": 1.340793489318413,
      "grad_norm": 1.1529191732406616,
      "learning_rate": 1e-05,
      "loss": 0.4341,
      "mean_token_accuracy": 0.8602442741394043,
      "num_tokens": 420174883.0,
      "step": 2636
    },
    {
      "epoch": 1.3413021363173958,
      "grad_norm": 0.9898726344108582,
      "learning_rate": 1e-05,
      "loss": 0.5235,
      "mean_token_accuracy": 0.8371015787124634,
      "num_tokens": 420329703.0,
      "step": 2637
    },
    {
      "epoch": 1.3418107833163784,
      "grad_norm": 0.980955958366394,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.851169764995575,
      "num_tokens": 420490175.0,
      "step": 2638
    },
    {
      "epoch": 1.3423194303153612,
      "grad_norm": 0.8918364644050598,
      "learning_rate": 1e-05,
      "loss": 0.467,
      "mean_token_accuracy": 0.8519745469093323,
      "num_tokens": 420654883.0,
      "step": 2639
    },
    {
      "epoch": 1.3428280773143437,
      "grad_norm": 1.001033902168274,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8565718531608582,
      "num_tokens": 420818659.0,
      "step": 2640
    },
    {
      "epoch": 1.3433367243133265,
      "grad_norm": 0.9805684089660645,
      "learning_rate": 1e-05,
      "loss": 0.498,
      "mean_token_accuracy": 0.8446347713470459,
      "num_tokens": 420990326.0,
      "step": 2641
    },
    {
      "epoch": 1.3438453713123093,
      "grad_norm": 0.966640055179596,
      "learning_rate": 1e-05,
      "loss": 0.4882,
      "mean_token_accuracy": 0.8449985980987549,
      "num_tokens": 421147515.0,
      "step": 2642
    },
    {
      "epoch": 1.344354018311292,
      "grad_norm": 1.0281821489334106,
      "learning_rate": 1e-05,
      "loss": 0.4496,
      "mean_token_accuracy": 0.8562248945236206,
      "num_tokens": 421305867.0,
      "step": 2643
    },
    {
      "epoch": 1.3448626653102747,
      "grad_norm": 1.0086641311645508,
      "learning_rate": 1e-05,
      "loss": 0.5143,
      "mean_token_accuracy": 0.8380690813064575,
      "num_tokens": 421474359.0,
      "step": 2644
    },
    {
      "epoch": 1.3453713123092574,
      "grad_norm": 0.9449248313903809,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8509700298309326,
      "num_tokens": 421633879.0,
      "step": 2645
    },
    {
      "epoch": 1.34587995930824,
      "grad_norm": 1.2007900476455688,
      "learning_rate": 1e-05,
      "loss": 0.4892,
      "mean_token_accuracy": 0.8451941013336182,
      "num_tokens": 421795778.0,
      "step": 2646
    },
    {
      "epoch": 1.3463886063072228,
      "grad_norm": 1.054432988166809,
      "learning_rate": 1e-05,
      "loss": 0.4985,
      "mean_token_accuracy": 0.8436516523361206,
      "num_tokens": 421962592.0,
      "step": 2647
    },
    {
      "epoch": 1.3468972533062056,
      "grad_norm": 0.9219520092010498,
      "learning_rate": 1e-05,
      "loss": 0.4862,
      "mean_token_accuracy": 0.8458356857299805,
      "num_tokens": 422127600.0,
      "step": 2648
    },
    {
      "epoch": 1.3474059003051881,
      "grad_norm": 1.0039230585098267,
      "learning_rate": 1e-05,
      "loss": 0.4762,
      "mean_token_accuracy": 0.8490136861801147,
      "num_tokens": 422286922.0,
      "step": 2649
    },
    {
      "epoch": 1.347914547304171,
      "grad_norm": 1.0406755208969116,
      "learning_rate": 1e-05,
      "loss": 0.5171,
      "mean_token_accuracy": 0.8369194269180298,
      "num_tokens": 422441520.0,
      "step": 2650
    },
    {
      "epoch": 1.3484231943031535,
      "grad_norm": 0.985836386680603,
      "learning_rate": 1e-05,
      "loss": 0.5037,
      "mean_token_accuracy": 0.8412511348724365,
      "num_tokens": 422613305.0,
      "step": 2651
    },
    {
      "epoch": 1.3489318413021363,
      "grad_norm": 1.060807466506958,
      "learning_rate": 1e-05,
      "loss": 0.5082,
      "mean_token_accuracy": 0.8395723104476929,
      "num_tokens": 422758073.0,
      "step": 2652
    },
    {
      "epoch": 1.349440488301119,
      "grad_norm": 1.0247995853424072,
      "learning_rate": 1e-05,
      "loss": 0.4683,
      "mean_token_accuracy": 0.8498069643974304,
      "num_tokens": 422916224.0,
      "step": 2653
    },
    {
      "epoch": 1.3499491353001019,
      "grad_norm": 0.9895526170730591,
      "learning_rate": 1e-05,
      "loss": 0.4935,
      "mean_token_accuracy": 0.8443385362625122,
      "num_tokens": 423072900.0,
      "step": 2654
    },
    {
      "epoch": 1.3504577822990844,
      "grad_norm": 0.9098997116088867,
      "learning_rate": 1e-05,
      "loss": 0.4692,
      "mean_token_accuracy": 0.8502326607704163,
      "num_tokens": 423239477.0,
      "step": 2655
    },
    {
      "epoch": 1.3509664292980672,
      "grad_norm": 1.019323706626892,
      "learning_rate": 1e-05,
      "loss": 0.4586,
      "mean_token_accuracy": 0.851577639579773,
      "num_tokens": 423399159.0,
      "step": 2656
    },
    {
      "epoch": 1.3514750762970498,
      "grad_norm": 1.067026138305664,
      "learning_rate": 1e-05,
      "loss": 0.511,
      "mean_token_accuracy": 0.8384853601455688,
      "num_tokens": 423550338.0,
      "step": 2657
    },
    {
      "epoch": 1.3519837232960326,
      "grad_norm": 0.9990204572677612,
      "learning_rate": 1e-05,
      "loss": 0.4742,
      "mean_token_accuracy": 0.8494395613670349,
      "num_tokens": 423723160.0,
      "step": 2658
    },
    {
      "epoch": 1.3524923702950153,
      "grad_norm": 0.9186148047447205,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "mean_token_accuracy": 0.8554156422615051,
      "num_tokens": 423884871.0,
      "step": 2659
    },
    {
      "epoch": 1.353001017293998,
      "grad_norm": 0.946681022644043,
      "learning_rate": 1e-05,
      "loss": 0.4214,
      "mean_token_accuracy": 0.8641282916069031,
      "num_tokens": 424042895.0,
      "step": 2660
    },
    {
      "epoch": 1.3535096642929807,
      "grad_norm": 0.9847366809844971,
      "learning_rate": 1e-05,
      "loss": 0.4319,
      "mean_token_accuracy": 0.8606440424919128,
      "num_tokens": 424204670.0,
      "step": 2661
    },
    {
      "epoch": 1.3540183112919633,
      "grad_norm": 1.00178062915802,
      "learning_rate": 1e-05,
      "loss": 0.5076,
      "mean_token_accuracy": 0.8411481976509094,
      "num_tokens": 424359175.0,
      "step": 2662
    },
    {
      "epoch": 1.354526958290946,
      "grad_norm": 1.051777720451355,
      "learning_rate": 1e-05,
      "loss": 0.4765,
      "mean_token_accuracy": 0.8494985103607178,
      "num_tokens": 424520295.0,
      "step": 2663
    },
    {
      "epoch": 1.3550356052899288,
      "grad_norm": 1.0117286443710327,
      "learning_rate": 1e-05,
      "loss": 0.4905,
      "mean_token_accuracy": 0.8441912531852722,
      "num_tokens": 424688740.0,
      "step": 2664
    },
    {
      "epoch": 1.3555442522889116,
      "grad_norm": 0.9656341075897217,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.853062629699707,
      "num_tokens": 424850384.0,
      "step": 2665
    },
    {
      "epoch": 1.3560528992878942,
      "grad_norm": 1.0329508781433105,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8456158638000488,
      "num_tokens": 425017780.0,
      "step": 2666
    },
    {
      "epoch": 1.356561546286877,
      "grad_norm": 0.9582757949829102,
      "learning_rate": 1e-05,
      "loss": 0.486,
      "mean_token_accuracy": 0.8466944098472595,
      "num_tokens": 425188533.0,
      "step": 2667
    },
    {
      "epoch": 1.3570701932858595,
      "grad_norm": 0.9417458772659302,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8438390493392944,
      "num_tokens": 425345218.0,
      "step": 2668
    },
    {
      "epoch": 1.3575788402848423,
      "grad_norm": 0.9999868273735046,
      "learning_rate": 1e-05,
      "loss": 0.4624,
      "mean_token_accuracy": 0.852497935295105,
      "num_tokens": 425502223.0,
      "step": 2669
    },
    {
      "epoch": 1.3580874872838251,
      "grad_norm": 1.0143228769302368,
      "learning_rate": 1e-05,
      "loss": 0.5025,
      "mean_token_accuracy": 0.841201901435852,
      "num_tokens": 425660450.0,
      "step": 2670
    },
    {
      "epoch": 1.3585961342828077,
      "grad_norm": 0.9608351588249207,
      "learning_rate": 1e-05,
      "loss": 0.4843,
      "mean_token_accuracy": 0.8468529582023621,
      "num_tokens": 425816822.0,
      "step": 2671
    },
    {
      "epoch": 1.3591047812817905,
      "grad_norm": 1.0515697002410889,
      "learning_rate": 1e-05,
      "loss": 0.4861,
      "mean_token_accuracy": 0.8458720445632935,
      "num_tokens": 425961495.0,
      "step": 2672
    },
    {
      "epoch": 1.359613428280773,
      "grad_norm": 1.0757752656936646,
      "learning_rate": 1e-05,
      "loss": 0.4844,
      "mean_token_accuracy": 0.8457462787628174,
      "num_tokens": 426120184.0,
      "step": 2673
    },
    {
      "epoch": 1.3601220752797558,
      "grad_norm": 0.9009885787963867,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.8526752591133118,
      "num_tokens": 426290019.0,
      "step": 2674
    },
    {
      "epoch": 1.3606307222787386,
      "grad_norm": 0.9713742733001709,
      "learning_rate": 1e-05,
      "loss": 0.4482,
      "mean_token_accuracy": 0.8567054271697998,
      "num_tokens": 426453175.0,
      "step": 2675
    },
    {
      "epoch": 1.3611393692777214,
      "grad_norm": 1.000468134880066,
      "learning_rate": 1e-05,
      "loss": 0.4714,
      "mean_token_accuracy": 0.8515945672988892,
      "num_tokens": 426600712.0,
      "step": 2676
    },
    {
      "epoch": 1.361648016276704,
      "grad_norm": 1.0043721199035645,
      "learning_rate": 1e-05,
      "loss": 0.4928,
      "mean_token_accuracy": 0.8436155915260315,
      "num_tokens": 426771838.0,
      "step": 2677
    },
    {
      "epoch": 1.3621566632756867,
      "grad_norm": 0.937410295009613,
      "learning_rate": 1e-05,
      "loss": 0.4773,
      "mean_token_accuracy": 0.8478071093559265,
      "num_tokens": 426930227.0,
      "step": 2678
    },
    {
      "epoch": 1.3626653102746693,
      "grad_norm": 0.9257480502128601,
      "learning_rate": 1e-05,
      "loss": 0.4607,
      "mean_token_accuracy": 0.8536421656608582,
      "num_tokens": 427089271.0,
      "step": 2679
    },
    {
      "epoch": 1.363173957273652,
      "grad_norm": 0.9284461140632629,
      "learning_rate": 1e-05,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8504741191864014,
      "num_tokens": 427263929.0,
      "step": 2680
    },
    {
      "epoch": 1.3636826042726349,
      "grad_norm": 0.8936075568199158,
      "learning_rate": 1e-05,
      "loss": 0.5087,
      "mean_token_accuracy": 0.8401069045066833,
      "num_tokens": 427428756.0,
      "step": 2681
    },
    {
      "epoch": 1.3641912512716174,
      "grad_norm": 0.9413860440254211,
      "learning_rate": 1e-05,
      "loss": 0.4682,
      "mean_token_accuracy": 0.8501747250556946,
      "num_tokens": 427581870.0,
      "step": 2682
    },
    {
      "epoch": 1.3646998982706002,
      "grad_norm": 0.9177972078323364,
      "learning_rate": 1e-05,
      "loss": 0.4617,
      "mean_token_accuracy": 0.8546660542488098,
      "num_tokens": 427739317.0,
      "step": 2683
    },
    {
      "epoch": 1.3652085452695828,
      "grad_norm": 0.9576936364173889,
      "learning_rate": 1e-05,
      "loss": 0.4697,
      "mean_token_accuracy": 0.8502979874610901,
      "num_tokens": 427890146.0,
      "step": 2684
    },
    {
      "epoch": 1.3657171922685656,
      "grad_norm": 0.8964239954948425,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.8439350128173828,
      "num_tokens": 428056020.0,
      "step": 2685
    },
    {
      "epoch": 1.3662258392675484,
      "grad_norm": 0.9161334037780762,
      "learning_rate": 1e-05,
      "loss": 0.4696,
      "mean_token_accuracy": 0.8516874313354492,
      "num_tokens": 428210521.0,
      "step": 2686
    },
    {
      "epoch": 1.3667344862665312,
      "grad_norm": 0.9134512543678284,
      "learning_rate": 1e-05,
      "loss": 0.4618,
      "mean_token_accuracy": 0.8511298298835754,
      "num_tokens": 428375323.0,
      "step": 2687
    },
    {
      "epoch": 1.3672431332655137,
      "grad_norm": 1.0040498971939087,
      "learning_rate": 1e-05,
      "loss": 0.5006,
      "mean_token_accuracy": 0.8420723676681519,
      "num_tokens": 428539565.0,
      "step": 2688
    },
    {
      "epoch": 1.3677517802644965,
      "grad_norm": 0.935520589351654,
      "learning_rate": 1e-05,
      "loss": 0.4939,
      "mean_token_accuracy": 0.8450379371643066,
      "num_tokens": 428720975.0,
      "step": 2689
    },
    {
      "epoch": 1.368260427263479,
      "grad_norm": 0.9524109363555908,
      "learning_rate": 1e-05,
      "loss": 0.4965,
      "mean_token_accuracy": 0.8424445986747742,
      "num_tokens": 428868507.0,
      "step": 2690
    },
    {
      "epoch": 1.3687690742624619,
      "grad_norm": 0.9552338123321533,
      "learning_rate": 1e-05,
      "loss": 0.5028,
      "mean_token_accuracy": 0.8432607650756836,
      "num_tokens": 429029080.0,
      "step": 2691
    },
    {
      "epoch": 1.3692777212614446,
      "grad_norm": 0.9479795098304749,
      "learning_rate": 1e-05,
      "loss": 0.4696,
      "mean_token_accuracy": 0.8508580327033997,
      "num_tokens": 429192002.0,
      "step": 2692
    },
    {
      "epoch": 1.3697863682604272,
      "grad_norm": 0.9779879450798035,
      "learning_rate": 1e-05,
      "loss": 0.4772,
      "mean_token_accuracy": 0.8503663539886475,
      "num_tokens": 429340295.0,
      "step": 2693
    },
    {
      "epoch": 1.37029501525941,
      "grad_norm": 0.9469121098518372,
      "learning_rate": 1e-05,
      "loss": 0.4849,
      "mean_token_accuracy": 0.8442733883857727,
      "num_tokens": 429499970.0,
      "step": 2694
    },
    {
      "epoch": 1.3708036622583926,
      "grad_norm": 0.9048606157302856,
      "learning_rate": 1e-05,
      "loss": 0.4583,
      "mean_token_accuracy": 0.8535619378089905,
      "num_tokens": 429658951.0,
      "step": 2695
    },
    {
      "epoch": 1.3713123092573754,
      "grad_norm": 0.9049401879310608,
      "learning_rate": 1e-05,
      "loss": 0.4954,
      "mean_token_accuracy": 0.8439792990684509,
      "num_tokens": 429822566.0,
      "step": 2696
    },
    {
      "epoch": 1.3718209562563581,
      "grad_norm": 0.9330436587333679,
      "learning_rate": 1e-05,
      "loss": 0.4538,
      "mean_token_accuracy": 0.8555499315261841,
      "num_tokens": 429982638.0,
      "step": 2697
    },
    {
      "epoch": 1.372329603255341,
      "grad_norm": 0.9345403909683228,
      "learning_rate": 1e-05,
      "loss": 0.459,
      "mean_token_accuracy": 0.8530346155166626,
      "num_tokens": 430138798.0,
      "step": 2698
    },
    {
      "epoch": 1.3728382502543235,
      "grad_norm": 0.9769986867904663,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.849714994430542,
      "num_tokens": 430294611.0,
      "step": 2699
    },
    {
      "epoch": 1.3733468972533063,
      "grad_norm": 0.9289859533309937,
      "learning_rate": 1e-05,
      "loss": 0.4339,
      "mean_token_accuracy": 0.8610995411872864,
      "num_tokens": 430457184.0,
      "step": 2700
    },
    {
      "epoch": 1.3738555442522888,
      "grad_norm": 0.959317684173584,
      "learning_rate": 1e-05,
      "loss": 0.4856,
      "mean_token_accuracy": 0.8475931882858276,
      "num_tokens": 430616726.0,
      "step": 2701
    },
    {
      "epoch": 1.3743641912512716,
      "grad_norm": 0.9652572870254517,
      "learning_rate": 1e-05,
      "loss": 0.4958,
      "mean_token_accuracy": 0.8435888290405273,
      "num_tokens": 430782101.0,
      "step": 2702
    },
    {
      "epoch": 1.3748728382502544,
      "grad_norm": 0.9083567261695862,
      "learning_rate": 1e-05,
      "loss": 0.4564,
      "mean_token_accuracy": 0.8545889854431152,
      "num_tokens": 430931165.0,
      "step": 2703
    },
    {
      "epoch": 1.375381485249237,
      "grad_norm": 1.0234804153442383,
      "learning_rate": 1e-05,
      "loss": 0.4629,
      "mean_token_accuracy": 0.85181725025177,
      "num_tokens": 431093719.0,
      "step": 2704
    },
    {
      "epoch": 1.3758901322482198,
      "grad_norm": 0.920213520526886,
      "learning_rate": 1e-05,
      "loss": 0.4506,
      "mean_token_accuracy": 0.8569689989089966,
      "num_tokens": 431256289.0,
      "step": 2705
    },
    {
      "epoch": 1.3763987792472023,
      "grad_norm": 0.9725545644760132,
      "learning_rate": 1e-05,
      "loss": 0.5236,
      "mean_token_accuracy": 0.836646318435669,
      "num_tokens": 431420888.0,
      "step": 2706
    },
    {
      "epoch": 1.3769074262461851,
      "grad_norm": 0.9515104293823242,
      "learning_rate": 1e-05,
      "loss": 0.4903,
      "mean_token_accuracy": 0.8439248204231262,
      "num_tokens": 431587261.0,
      "step": 2707
    },
    {
      "epoch": 1.377416073245168,
      "grad_norm": 0.9942727088928223,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8490264415740967,
      "num_tokens": 431747572.0,
      "step": 2708
    },
    {
      "epoch": 1.3779247202441507,
      "grad_norm": 0.9128475785255432,
      "learning_rate": 1e-05,
      "loss": 0.4742,
      "mean_token_accuracy": 0.8506051301956177,
      "num_tokens": 431901013.0,
      "step": 2709
    },
    {
      "epoch": 1.3784333672431333,
      "grad_norm": 1.0360462665557861,
      "learning_rate": 1e-05,
      "loss": 0.5207,
      "mean_token_accuracy": 0.8344942331314087,
      "num_tokens": 432053072.0,
      "step": 2710
    },
    {
      "epoch": 1.378942014242116,
      "grad_norm": 1.0364712476730347,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8600242137908936,
      "num_tokens": 432218359.0,
      "step": 2711
    },
    {
      "epoch": 1.3794506612410986,
      "grad_norm": 1.0053229331970215,
      "learning_rate": 1e-05,
      "loss": 0.4712,
      "mean_token_accuracy": 0.8485782742500305,
      "num_tokens": 432356553.0,
      "step": 2712
    },
    {
      "epoch": 1.3799593082400814,
      "grad_norm": 0.9933727383613586,
      "learning_rate": 1e-05,
      "loss": 0.5019,
      "mean_token_accuracy": 0.841928243637085,
      "num_tokens": 432507222.0,
      "step": 2713
    },
    {
      "epoch": 1.3804679552390642,
      "grad_norm": 0.9446233510971069,
      "learning_rate": 1e-05,
      "loss": 0.4859,
      "mean_token_accuracy": 0.8459019660949707,
      "num_tokens": 432670085.0,
      "step": 2714
    },
    {
      "epoch": 1.3809766022380467,
      "grad_norm": 0.9640635848045349,
      "learning_rate": 1e-05,
      "loss": 0.5024,
      "mean_token_accuracy": 0.8412888050079346,
      "num_tokens": 432833746.0,
      "step": 2715
    },
    {
      "epoch": 1.3814852492370295,
      "grad_norm": 0.9870038032531738,
      "learning_rate": 1e-05,
      "loss": 0.5188,
      "mean_token_accuracy": 0.8366783261299133,
      "num_tokens": 433004996.0,
      "step": 2716
    },
    {
      "epoch": 1.381993896236012,
      "grad_norm": 0.9928244948387146,
      "learning_rate": 1e-05,
      "loss": 0.4501,
      "mean_token_accuracy": 0.8561815023422241,
      "num_tokens": 433157665.0,
      "step": 2717
    },
    {
      "epoch": 1.3825025432349949,
      "grad_norm": 0.9694227576255798,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8482884764671326,
      "num_tokens": 433309340.0,
      "step": 2718
    },
    {
      "epoch": 1.3830111902339777,
      "grad_norm": 0.9639489054679871,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "mean_token_accuracy": 0.8555192947387695,
      "num_tokens": 433484664.0,
      "step": 2719
    },
    {
      "epoch": 1.3835198372329605,
      "grad_norm": 0.9023261666297913,
      "learning_rate": 1e-05,
      "loss": 0.4643,
      "mean_token_accuracy": 0.8527742028236389,
      "num_tokens": 433655158.0,
      "step": 2720
    },
    {
      "epoch": 1.384028484231943,
      "grad_norm": 1.0446405410766602,
      "learning_rate": 1e-05,
      "loss": 0.4675,
      "mean_token_accuracy": 0.851974368095398,
      "num_tokens": 433811771.0,
      "step": 2721
    },
    {
      "epoch": 1.3845371312309258,
      "grad_norm": 0.9548596143722534,
      "learning_rate": 1e-05,
      "loss": 0.4499,
      "mean_token_accuracy": 0.8571314811706543,
      "num_tokens": 433968624.0,
      "step": 2722
    },
    {
      "epoch": 1.3850457782299084,
      "grad_norm": 0.9368325471878052,
      "learning_rate": 1e-05,
      "loss": 0.488,
      "mean_token_accuracy": 0.8435842990875244,
      "num_tokens": 434132817.0,
      "step": 2723
    },
    {
      "epoch": 1.3855544252288912,
      "grad_norm": 0.9053055644035339,
      "learning_rate": 1e-05,
      "loss": 0.4657,
      "mean_token_accuracy": 0.8512722253799438,
      "num_tokens": 434295363.0,
      "step": 2724
    },
    {
      "epoch": 1.386063072227874,
      "grad_norm": 0.9808530807495117,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8470052480697632,
      "num_tokens": 434461451.0,
      "step": 2725
    },
    {
      "epoch": 1.3865717192268565,
      "grad_norm": 0.9754900932312012,
      "learning_rate": 1e-05,
      "loss": 0.4883,
      "mean_token_accuracy": 0.8470308184623718,
      "num_tokens": 434613231.0,
      "step": 2726
    },
    {
      "epoch": 1.3870803662258393,
      "grad_norm": 0.975768506526947,
      "learning_rate": 1e-05,
      "loss": 0.4967,
      "mean_token_accuracy": 0.842704713344574,
      "num_tokens": 434761743.0,
      "step": 2727
    },
    {
      "epoch": 1.3875890132248219,
      "grad_norm": 1.0329039096832275,
      "learning_rate": 1e-05,
      "loss": 0.4867,
      "mean_token_accuracy": 0.8458406329154968,
      "num_tokens": 434917053.0,
      "step": 2728
    },
    {
      "epoch": 1.3880976602238047,
      "grad_norm": 1.0342756509780884,
      "learning_rate": 1e-05,
      "loss": 0.5157,
      "mean_token_accuracy": 0.8370108008384705,
      "num_tokens": 435071996.0,
      "step": 2729
    },
    {
      "epoch": 1.3886063072227874,
      "grad_norm": 1.015746831893921,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.8457257747650146,
      "num_tokens": 435226320.0,
      "step": 2730
    },
    {
      "epoch": 1.3891149542217702,
      "grad_norm": 1.011082410812378,
      "learning_rate": 1e-05,
      "loss": 0.4614,
      "mean_token_accuracy": 0.8516829013824463,
      "num_tokens": 435383004.0,
      "step": 2731
    },
    {
      "epoch": 1.3896236012207528,
      "grad_norm": 1.0925073623657227,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8538006544113159,
      "num_tokens": 435536990.0,
      "step": 2732
    },
    {
      "epoch": 1.3901322482197356,
      "grad_norm": 0.9984236359596252,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8531023859977722,
      "num_tokens": 435694177.0,
      "step": 2733
    },
    {
      "epoch": 1.3906408952187181,
      "grad_norm": 1.0924962759017944,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.8423185348510742,
      "num_tokens": 435847048.0,
      "step": 2734
    },
    {
      "epoch": 1.391149542217701,
      "grad_norm": 1.1045939922332764,
      "learning_rate": 1e-05,
      "loss": 0.5205,
      "mean_token_accuracy": 0.8350080251693726,
      "num_tokens": 436003472.0,
      "step": 2735
    },
    {
      "epoch": 1.3916581892166837,
      "grad_norm": 1.0720640420913696,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8450804352760315,
      "num_tokens": 436161792.0,
      "step": 2736
    },
    {
      "epoch": 1.3921668362156663,
      "grad_norm": 0.997451663017273,
      "learning_rate": 1e-05,
      "loss": 0.5235,
      "mean_token_accuracy": 0.8371497988700867,
      "num_tokens": 436330865.0,
      "step": 2737
    },
    {
      "epoch": 1.392675483214649,
      "grad_norm": 1.0413810014724731,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.8415367007255554,
      "num_tokens": 436497752.0,
      "step": 2738
    },
    {
      "epoch": 1.3931841302136316,
      "grad_norm": 1.0180333852767944,
      "learning_rate": 1e-05,
      "loss": 0.4683,
      "mean_token_accuracy": 0.8512299656867981,
      "num_tokens": 436671181.0,
      "step": 2739
    },
    {
      "epoch": 1.3936927772126144,
      "grad_norm": 0.9152499437332153,
      "learning_rate": 1e-05,
      "loss": 0.4608,
      "mean_token_accuracy": 0.8541569113731384,
      "num_tokens": 436827912.0,
      "step": 2740
    },
    {
      "epoch": 1.3942014242115972,
      "grad_norm": 0.9704434871673584,
      "learning_rate": 1e-05,
      "loss": 0.4664,
      "mean_token_accuracy": 0.8514987230300903,
      "num_tokens": 436994067.0,
      "step": 2741
    },
    {
      "epoch": 1.39471007121058,
      "grad_norm": 1.0042692422866821,
      "learning_rate": 1e-05,
      "loss": 0.4946,
      "mean_token_accuracy": 0.8427151441574097,
      "num_tokens": 437161113.0,
      "step": 2742
    },
    {
      "epoch": 1.3952187182095626,
      "grad_norm": 1.05972158908844,
      "learning_rate": 1e-05,
      "loss": 0.4934,
      "mean_token_accuracy": 0.8417543172836304,
      "num_tokens": 437304648.0,
      "step": 2743
    },
    {
      "epoch": 1.3957273652085453,
      "grad_norm": 0.996813952922821,
      "learning_rate": 1e-05,
      "loss": 0.5008,
      "mean_token_accuracy": 0.8440243005752563,
      "num_tokens": 437453477.0,
      "step": 2744
    },
    {
      "epoch": 1.396236012207528,
      "grad_norm": 0.9492610692977905,
      "learning_rate": 1e-05,
      "loss": 0.492,
      "mean_token_accuracy": 0.8438138961791992,
      "num_tokens": 437593153.0,
      "step": 2745
    },
    {
      "epoch": 1.3967446592065107,
      "grad_norm": 1.0397744178771973,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.842868447303772,
      "num_tokens": 437749330.0,
      "step": 2746
    },
    {
      "epoch": 1.3972533062054935,
      "grad_norm": 0.9676275253295898,
      "learning_rate": 1e-05,
      "loss": 0.4434,
      "mean_token_accuracy": 0.8574696779251099,
      "num_tokens": 437902460.0,
      "step": 2747
    },
    {
      "epoch": 1.397761953204476,
      "grad_norm": 0.9755284190177917,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.850979745388031,
      "num_tokens": 438064905.0,
      "step": 2748
    },
    {
      "epoch": 1.3982706002034588,
      "grad_norm": 0.9887830018997192,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8454263210296631,
      "num_tokens": 438217621.0,
      "step": 2749
    },
    {
      "epoch": 1.3987792472024414,
      "grad_norm": 0.9100257754325867,
      "learning_rate": 1e-05,
      "loss": 0.45,
      "mean_token_accuracy": 0.8556119203567505,
      "num_tokens": 438374191.0,
      "step": 2750
    },
    {
      "epoch": 1.3992878942014242,
      "grad_norm": 1.0211602449417114,
      "learning_rate": 1e-05,
      "loss": 0.475,
      "mean_token_accuracy": 0.847649097442627,
      "num_tokens": 438520298.0,
      "step": 2751
    },
    {
      "epoch": 1.399796541200407,
      "grad_norm": 0.9627428650856018,
      "learning_rate": 1e-05,
      "loss": 0.4474,
      "mean_token_accuracy": 0.8568810224533081,
      "num_tokens": 438672284.0,
      "step": 2752
    },
    {
      "epoch": 1.4003051881993898,
      "grad_norm": 0.966880738735199,
      "learning_rate": 1e-05,
      "loss": 0.5125,
      "mean_token_accuracy": 0.8379854559898376,
      "num_tokens": 438822201.0,
      "step": 2753
    },
    {
      "epoch": 1.4008138351983723,
      "grad_norm": 0.9599084854125977,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8557698726654053,
      "num_tokens": 438980241.0,
      "step": 2754
    },
    {
      "epoch": 1.4013224821973551,
      "grad_norm": 0.9639296531677246,
      "learning_rate": 1e-05,
      "loss": 0.447,
      "mean_token_accuracy": 0.8570913076400757,
      "num_tokens": 439138528.0,
      "step": 2755
    },
    {
      "epoch": 1.4018311291963377,
      "grad_norm": 1.021277666091919,
      "learning_rate": 1e-05,
      "loss": 0.4915,
      "mean_token_accuracy": 0.8457634449005127,
      "num_tokens": 439300154.0,
      "step": 2756
    },
    {
      "epoch": 1.4023397761953205,
      "grad_norm": 0.9952744245529175,
      "learning_rate": 1e-05,
      "loss": 0.4269,
      "mean_token_accuracy": 0.8626008033752441,
      "num_tokens": 439457699.0,
      "step": 2757
    },
    {
      "epoch": 1.4028484231943033,
      "grad_norm": 0.9347396492958069,
      "learning_rate": 1e-05,
      "loss": 0.4666,
      "mean_token_accuracy": 0.851100742816925,
      "num_tokens": 439612805.0,
      "step": 2758
    },
    {
      "epoch": 1.4033570701932858,
      "grad_norm": 0.960252583026886,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8494982719421387,
      "num_tokens": 439772835.0,
      "step": 2759
    },
    {
      "epoch": 1.4038657171922686,
      "grad_norm": 1.1808850765228271,
      "learning_rate": 1e-05,
      "loss": 0.4788,
      "mean_token_accuracy": 0.8483887910842896,
      "num_tokens": 439927068.0,
      "step": 2760
    },
    {
      "epoch": 1.4043743641912512,
      "grad_norm": 1.0060442686080933,
      "learning_rate": 1e-05,
      "loss": 0.4624,
      "mean_token_accuracy": 0.8521533608436584,
      "num_tokens": 440082717.0,
      "step": 2761
    },
    {
      "epoch": 1.404883011190234,
      "grad_norm": 1.1049740314483643,
      "learning_rate": 1e-05,
      "loss": 0.5135,
      "mean_token_accuracy": 0.8387951850891113,
      "num_tokens": 440231502.0,
      "step": 2762
    },
    {
      "epoch": 1.4053916581892167,
      "grad_norm": 1.0528883934020996,
      "learning_rate": 1e-05,
      "loss": 0.4577,
      "mean_token_accuracy": 0.8528305888175964,
      "num_tokens": 440385997.0,
      "step": 2763
    },
    {
      "epoch": 1.4059003051881995,
      "grad_norm": 0.94994056224823,
      "learning_rate": 1e-05,
      "loss": 0.5004,
      "mean_token_accuracy": 0.8418850302696228,
      "num_tokens": 440550613.0,
      "step": 2764
    },
    {
      "epoch": 1.406408952187182,
      "grad_norm": 1.0399938821792603,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8524987697601318,
      "num_tokens": 440724447.0,
      "step": 2765
    },
    {
      "epoch": 1.4069175991861649,
      "grad_norm": 1.128157138824463,
      "learning_rate": 1e-05,
      "loss": 0.4896,
      "mean_token_accuracy": 0.8465457558631897,
      "num_tokens": 440878494.0,
      "step": 2766
    },
    {
      "epoch": 1.4074262461851474,
      "grad_norm": 1.0129573345184326,
      "learning_rate": 1e-05,
      "loss": 0.4749,
      "mean_token_accuracy": 0.8479048609733582,
      "num_tokens": 441034220.0,
      "step": 2767
    },
    {
      "epoch": 1.4079348931841302,
      "grad_norm": 1.0731834173202515,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8481513857841492,
      "num_tokens": 441175256.0,
      "step": 2768
    },
    {
      "epoch": 1.408443540183113,
      "grad_norm": 1.1229848861694336,
      "learning_rate": 1e-05,
      "loss": 0.4729,
      "mean_token_accuracy": 0.8501735925674438,
      "num_tokens": 441330143.0,
      "step": 2769
    },
    {
      "epoch": 1.4089521871820956,
      "grad_norm": 0.996660053730011,
      "learning_rate": 1e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8471207618713379,
      "num_tokens": 441490539.0,
      "step": 2770
    },
    {
      "epoch": 1.4094608341810784,
      "grad_norm": 1.021069049835205,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8487145900726318,
      "num_tokens": 441650753.0,
      "step": 2771
    },
    {
      "epoch": 1.409969481180061,
      "grad_norm": 1.0124939680099487,
      "learning_rate": 1e-05,
      "loss": 0.478,
      "mean_token_accuracy": 0.8463742136955261,
      "num_tokens": 441811314.0,
      "step": 2772
    },
    {
      "epoch": 1.4104781281790437,
      "grad_norm": 0.939225971698761,
      "learning_rate": 1e-05,
      "loss": 0.4578,
      "mean_token_accuracy": 0.8548750281333923,
      "num_tokens": 441966161.0,
      "step": 2773
    },
    {
      "epoch": 1.4109867751780265,
      "grad_norm": 1.051579475402832,
      "learning_rate": 1e-05,
      "loss": 0.4782,
      "mean_token_accuracy": 0.847815752029419,
      "num_tokens": 442127917.0,
      "step": 2774
    },
    {
      "epoch": 1.411495422177009,
      "grad_norm": 1.3779577016830444,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8557332754135132,
      "num_tokens": 442276671.0,
      "step": 2775
    },
    {
      "epoch": 1.4120040691759919,
      "grad_norm": 1.0016024112701416,
      "learning_rate": 1e-05,
      "loss": 0.5314,
      "mean_token_accuracy": 0.8357734680175781,
      "num_tokens": 442435182.0,
      "step": 2776
    },
    {
      "epoch": 1.4125127161749746,
      "grad_norm": 0.9864020347595215,
      "learning_rate": 1e-05,
      "loss": 0.4863,
      "mean_token_accuracy": 0.8437013030052185,
      "num_tokens": 442605940.0,
      "step": 2777
    },
    {
      "epoch": 1.4130213631739572,
      "grad_norm": 1.0168684720993042,
      "learning_rate": 1e-05,
      "loss": 0.4304,
      "mean_token_accuracy": 0.8620567321777344,
      "num_tokens": 442769476.0,
      "step": 2778
    },
    {
      "epoch": 1.41353001017294,
      "grad_norm": 0.9997568130493164,
      "learning_rate": 1e-05,
      "loss": 0.4958,
      "mean_token_accuracy": 0.8428398966789246,
      "num_tokens": 442932405.0,
      "step": 2779
    },
    {
      "epoch": 1.4140386571719228,
      "grad_norm": 1.027672290802002,
      "learning_rate": 1e-05,
      "loss": 0.4648,
      "mean_token_accuracy": 0.8519032001495361,
      "num_tokens": 443108364.0,
      "step": 2780
    },
    {
      "epoch": 1.4145473041709054,
      "grad_norm": 1.060289740562439,
      "learning_rate": 1e-05,
      "loss": 0.4696,
      "mean_token_accuracy": 0.8508599996566772,
      "num_tokens": 443270874.0,
      "step": 2781
    },
    {
      "epoch": 1.4150559511698881,
      "grad_norm": 0.9960584044456482,
      "learning_rate": 1e-05,
      "loss": 0.4639,
      "mean_token_accuracy": 0.8511028289794922,
      "num_tokens": 443434228.0,
      "step": 2782
    },
    {
      "epoch": 1.4155645981688707,
      "grad_norm": 1.0055328607559204,
      "learning_rate": 1e-05,
      "loss": 0.4966,
      "mean_token_accuracy": 0.8413923382759094,
      "num_tokens": 443602526.0,
      "step": 2783
    },
    {
      "epoch": 1.4160732451678535,
      "grad_norm": 1.0092580318450928,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8493560552597046,
      "num_tokens": 443753663.0,
      "step": 2784
    },
    {
      "epoch": 1.4165818921668363,
      "grad_norm": 0.9981493353843689,
      "learning_rate": 1e-05,
      "loss": 0.4607,
      "mean_token_accuracy": 0.8523511290550232,
      "num_tokens": 443913758.0,
      "step": 2785
    },
    {
      "epoch": 1.4170905391658188,
      "grad_norm": 0.988707959651947,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8555818796157837,
      "num_tokens": 444064404.0,
      "step": 2786
    },
    {
      "epoch": 1.4175991861648016,
      "grad_norm": 0.8756354451179504,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.8522771596908569,
      "num_tokens": 444225462.0,
      "step": 2787
    },
    {
      "epoch": 1.4181078331637844,
      "grad_norm": 1.0623990297317505,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.84184730052948,
      "num_tokens": 444385872.0,
      "step": 2788
    },
    {
      "epoch": 1.418616480162767,
      "grad_norm": 1.0685012340545654,
      "learning_rate": 1e-05,
      "loss": 0.4789,
      "mean_token_accuracy": 0.8478412628173828,
      "num_tokens": 444533417.0,
      "step": 2789
    },
    {
      "epoch": 1.4191251271617498,
      "grad_norm": 0.924235999584198,
      "learning_rate": 1e-05,
      "loss": 0.4786,
      "mean_token_accuracy": 0.8480336666107178,
      "num_tokens": 444686912.0,
      "step": 2790
    },
    {
      "epoch": 1.4196337741607326,
      "grad_norm": 0.9503185153007507,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.8519868850708008,
      "num_tokens": 444848516.0,
      "step": 2791
    },
    {
      "epoch": 1.4201424211597151,
      "grad_norm": 0.9894545674324036,
      "learning_rate": 1e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.8499791622161865,
      "num_tokens": 445012426.0,
      "step": 2792
    },
    {
      "epoch": 1.420651068158698,
      "grad_norm": 1.0658549070358276,
      "learning_rate": 1e-05,
      "loss": 0.4923,
      "mean_token_accuracy": 0.8419213891029358,
      "num_tokens": 445177318.0,
      "step": 2793
    },
    {
      "epoch": 1.4211597151576805,
      "grad_norm": 0.8910920023918152,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8462452292442322,
      "num_tokens": 445343179.0,
      "step": 2794
    },
    {
      "epoch": 1.4216683621566633,
      "grad_norm": 0.9932237863540649,
      "learning_rate": 1e-05,
      "loss": 0.4626,
      "mean_token_accuracy": 0.8527805805206299,
      "num_tokens": 445513613.0,
      "step": 2795
    },
    {
      "epoch": 1.422177009155646,
      "grad_norm": 0.9207007884979248,
      "learning_rate": 1e-05,
      "loss": 0.4447,
      "mean_token_accuracy": 0.8579326868057251,
      "num_tokens": 445669852.0,
      "step": 2796
    },
    {
      "epoch": 1.4226856561546286,
      "grad_norm": 1.049477219581604,
      "learning_rate": 1e-05,
      "loss": 0.5077,
      "mean_token_accuracy": 0.8388053178787231,
      "num_tokens": 445837533.0,
      "step": 2797
    },
    {
      "epoch": 1.4231943031536114,
      "grad_norm": 1.0226757526397705,
      "learning_rate": 1e-05,
      "loss": 0.4385,
      "mean_token_accuracy": 0.8591240644454956,
      "num_tokens": 445986396.0,
      "step": 2798
    },
    {
      "epoch": 1.4237029501525942,
      "grad_norm": 0.9482563734054565,
      "learning_rate": 1e-05,
      "loss": 0.4733,
      "mean_token_accuracy": 0.8491603136062622,
      "num_tokens": 446136330.0,
      "step": 2799
    },
    {
      "epoch": 1.4242115971515767,
      "grad_norm": 1.045009970664978,
      "learning_rate": 1e-05,
      "loss": 0.4835,
      "mean_token_accuracy": 0.8467520475387573,
      "num_tokens": 446284635.0,
      "step": 2800
    },
    {
      "epoch": 1.4247202441505595,
      "grad_norm": 0.9708003997802734,
      "learning_rate": 1e-05,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8428059816360474,
      "num_tokens": 446448720.0,
      "step": 2801
    },
    {
      "epoch": 1.4252288911495423,
      "grad_norm": 1.0531318187713623,
      "learning_rate": 1e-05,
      "loss": 0.469,
      "mean_token_accuracy": 0.8497306108474731,
      "num_tokens": 446599329.0,
      "step": 2802
    },
    {
      "epoch": 1.4257375381485249,
      "grad_norm": 0.9104205369949341,
      "learning_rate": 1e-05,
      "loss": 0.4749,
      "mean_token_accuracy": 0.8491522669792175,
      "num_tokens": 446764283.0,
      "step": 2803
    },
    {
      "epoch": 1.4262461851475077,
      "grad_norm": 0.9720317125320435,
      "learning_rate": 1e-05,
      "loss": 0.4796,
      "mean_token_accuracy": 0.8502120971679688,
      "num_tokens": 446921393.0,
      "step": 2804
    },
    {
      "epoch": 1.4267548321464902,
      "grad_norm": 0.9275528192520142,
      "learning_rate": 1e-05,
      "loss": 0.4643,
      "mean_token_accuracy": 0.8525756597518921,
      "num_tokens": 447086304.0,
      "step": 2805
    },
    {
      "epoch": 1.427263479145473,
      "grad_norm": 0.9052826166152954,
      "learning_rate": 1e-05,
      "loss": 0.4782,
      "mean_token_accuracy": 0.8469159603118896,
      "num_tokens": 447257986.0,
      "step": 2806
    },
    {
      "epoch": 1.4277721261444558,
      "grad_norm": 0.965103268623352,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8643838763237,
      "num_tokens": 447408803.0,
      "step": 2807
    },
    {
      "epoch": 1.4282807731434384,
      "grad_norm": 0.9626489877700806,
      "learning_rate": 1e-05,
      "loss": 0.4811,
      "mean_token_accuracy": 0.8474384546279907,
      "num_tokens": 447568033.0,
      "step": 2808
    },
    {
      "epoch": 1.4287894201424212,
      "grad_norm": 0.9533666968345642,
      "learning_rate": 1e-05,
      "loss": 0.4789,
      "mean_token_accuracy": 0.8479937314987183,
      "num_tokens": 447727480.0,
      "step": 2809
    },
    {
      "epoch": 1.4292980671414037,
      "grad_norm": 0.9389809966087341,
      "learning_rate": 1e-05,
      "loss": 0.4553,
      "mean_token_accuracy": 0.8539700508117676,
      "num_tokens": 447885294.0,
      "step": 2810
    },
    {
      "epoch": 1.4298067141403865,
      "grad_norm": 0.95237797498703,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8428686857223511,
      "num_tokens": 448041135.0,
      "step": 2811
    },
    {
      "epoch": 1.4303153611393693,
      "grad_norm": 0.9649375081062317,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8458592295646667,
      "num_tokens": 448201132.0,
      "step": 2812
    },
    {
      "epoch": 1.430824008138352,
      "grad_norm": 0.9626204967498779,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8470381498336792,
      "num_tokens": 448364825.0,
      "step": 2813
    },
    {
      "epoch": 1.4313326551373347,
      "grad_norm": 1.020817518234253,
      "learning_rate": 1e-05,
      "loss": 0.4839,
      "mean_token_accuracy": 0.8466824293136597,
      "num_tokens": 448516015.0,
      "step": 2814
    },
    {
      "epoch": 1.4318413021363174,
      "grad_norm": 1.0039154291152954,
      "learning_rate": 1e-05,
      "loss": 0.4984,
      "mean_token_accuracy": 0.841713011264801,
      "num_tokens": 448680126.0,
      "step": 2815
    },
    {
      "epoch": 1.4323499491353,
      "grad_norm": 0.9738489389419556,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8518779277801514,
      "num_tokens": 448828846.0,
      "step": 2816
    },
    {
      "epoch": 1.4328585961342828,
      "grad_norm": 0.9059901237487793,
      "learning_rate": 1e-05,
      "loss": 0.4671,
      "mean_token_accuracy": 0.8512499332427979,
      "num_tokens": 448987846.0,
      "step": 2817
    },
    {
      "epoch": 1.4333672431332656,
      "grad_norm": 0.9546349048614502,
      "learning_rate": 1e-05,
      "loss": 0.5249,
      "mean_token_accuracy": 0.8342456221580505,
      "num_tokens": 449153221.0,
      "step": 2818
    },
    {
      "epoch": 1.4338758901322481,
      "grad_norm": 0.9544728994369507,
      "learning_rate": 1e-05,
      "loss": 0.4812,
      "mean_token_accuracy": 0.8467628955841064,
      "num_tokens": 449310276.0,
      "step": 2819
    },
    {
      "epoch": 1.434384537131231,
      "grad_norm": 0.9785590767860413,
      "learning_rate": 1e-05,
      "loss": 0.508,
      "mean_token_accuracy": 0.8397389650344849,
      "num_tokens": 449462904.0,
      "step": 2820
    },
    {
      "epoch": 1.4348931841302135,
      "grad_norm": 0.9594732522964478,
      "learning_rate": 1e-05,
      "loss": 0.4715,
      "mean_token_accuracy": 0.8496230244636536,
      "num_tokens": 449625366.0,
      "step": 2821
    },
    {
      "epoch": 1.4354018311291963,
      "grad_norm": 0.9512030482292175,
      "learning_rate": 1e-05,
      "loss": 0.479,
      "mean_token_accuracy": 0.8478878140449524,
      "num_tokens": 449779209.0,
      "step": 2822
    },
    {
      "epoch": 1.435910478128179,
      "grad_norm": 0.9894237518310547,
      "learning_rate": 1e-05,
      "loss": 0.5168,
      "mean_token_accuracy": 0.837615966796875,
      "num_tokens": 449930647.0,
      "step": 2823
    },
    {
      "epoch": 1.4364191251271619,
      "grad_norm": 0.9918603301048279,
      "learning_rate": 1e-05,
      "loss": 0.48,
      "mean_token_accuracy": 0.8468407392501831,
      "num_tokens": 450077259.0,
      "step": 2824
    },
    {
      "epoch": 1.4369277721261444,
      "grad_norm": 1.0565818548202515,
      "learning_rate": 1e-05,
      "loss": 0.4795,
      "mean_token_accuracy": 0.8497186899185181,
      "num_tokens": 450234173.0,
      "step": 2825
    },
    {
      "epoch": 1.4374364191251272,
      "grad_norm": 0.9873266220092773,
      "learning_rate": 1e-05,
      "loss": 0.4826,
      "mean_token_accuracy": 0.8455828428268433,
      "num_tokens": 450408833.0,
      "step": 2826
    },
    {
      "epoch": 1.4379450661241098,
      "grad_norm": 1.0310784578323364,
      "learning_rate": 1e-05,
      "loss": 0.4976,
      "mean_token_accuracy": 0.842021107673645,
      "num_tokens": 450574471.0,
      "step": 2827
    },
    {
      "epoch": 1.4384537131230926,
      "grad_norm": 0.9773332476615906,
      "learning_rate": 1e-05,
      "loss": 0.4595,
      "mean_token_accuracy": 0.8536139130592346,
      "num_tokens": 450736983.0,
      "step": 2828
    },
    {
      "epoch": 1.4389623601220753,
      "grad_norm": 1.030726432800293,
      "learning_rate": 1e-05,
      "loss": 0.4519,
      "mean_token_accuracy": 0.8553815484046936,
      "num_tokens": 450896131.0,
      "step": 2829
    },
    {
      "epoch": 1.439471007121058,
      "grad_norm": 0.9751784801483154,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8510292768478394,
      "num_tokens": 451046227.0,
      "step": 2830
    },
    {
      "epoch": 1.4399796541200407,
      "grad_norm": 1.0584436655044556,
      "learning_rate": 1e-05,
      "loss": 0.4779,
      "mean_token_accuracy": 0.849427342414856,
      "num_tokens": 451194092.0,
      "step": 2831
    },
    {
      "epoch": 1.4404883011190233,
      "grad_norm": 1.0584428310394287,
      "learning_rate": 1e-05,
      "loss": 0.4836,
      "mean_token_accuracy": 0.8473069667816162,
      "num_tokens": 451349176.0,
      "step": 2832
    },
    {
      "epoch": 1.440996948118006,
      "grad_norm": 0.9864810705184937,
      "learning_rate": 1e-05,
      "loss": 0.4852,
      "mean_token_accuracy": 0.8452579379081726,
      "num_tokens": 451503657.0,
      "step": 2833
    },
    {
      "epoch": 1.4415055951169888,
      "grad_norm": 0.9788768291473389,
      "learning_rate": 1e-05,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8486111164093018,
      "num_tokens": 451667384.0,
      "step": 2834
    },
    {
      "epoch": 1.4420142421159716,
      "grad_norm": 0.9842734932899475,
      "learning_rate": 1e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8461251258850098,
      "num_tokens": 451825588.0,
      "step": 2835
    },
    {
      "epoch": 1.4425228891149542,
      "grad_norm": 0.9226424098014832,
      "learning_rate": 1e-05,
      "loss": 0.4933,
      "mean_token_accuracy": 0.8441954255104065,
      "num_tokens": 451978435.0,
      "step": 2836
    },
    {
      "epoch": 1.443031536113937,
      "grad_norm": 1.0472816228866577,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8495010137557983,
      "num_tokens": 452132400.0,
      "step": 2837
    },
    {
      "epoch": 1.4435401831129195,
      "grad_norm": 0.9353339076042175,
      "learning_rate": 1e-05,
      "loss": 0.4927,
      "mean_token_accuracy": 0.8432925939559937,
      "num_tokens": 452287174.0,
      "step": 2838
    },
    {
      "epoch": 1.4440488301119023,
      "grad_norm": 0.919854462146759,
      "learning_rate": 1e-05,
      "loss": 0.479,
      "mean_token_accuracy": 0.8476002216339111,
      "num_tokens": 452446708.0,
      "step": 2839
    },
    {
      "epoch": 1.4445574771108851,
      "grad_norm": 1.0054397583007812,
      "learning_rate": 1e-05,
      "loss": 0.4946,
      "mean_token_accuracy": 0.8426610231399536,
      "num_tokens": 452595925.0,
      "step": 2840
    },
    {
      "epoch": 1.4450661241098677,
      "grad_norm": 0.9773548245429993,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8544117212295532,
      "num_tokens": 452741850.0,
      "step": 2841
    },
    {
      "epoch": 1.4455747711088505,
      "grad_norm": 0.9018303751945496,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.8554398417472839,
      "num_tokens": 452902299.0,
      "step": 2842
    },
    {
      "epoch": 1.446083418107833,
      "grad_norm": 0.9995993375778198,
      "learning_rate": 1e-05,
      "loss": 0.4995,
      "mean_token_accuracy": 0.8423516750335693,
      "num_tokens": 453060469.0,
      "step": 2843
    },
    {
      "epoch": 1.4465920651068158,
      "grad_norm": 1.0751382112503052,
      "learning_rate": 1e-05,
      "loss": 0.476,
      "mean_token_accuracy": 0.8492323160171509,
      "num_tokens": 453219512.0,
      "step": 2844
    },
    {
      "epoch": 1.4471007121057986,
      "grad_norm": 0.9150761961936951,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.8534543514251709,
      "num_tokens": 453373593.0,
      "step": 2845
    },
    {
      "epoch": 1.4476093591047814,
      "grad_norm": 0.9494074583053589,
      "learning_rate": 1e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8472472429275513,
      "num_tokens": 453528683.0,
      "step": 2846
    },
    {
      "epoch": 1.448118006103764,
      "grad_norm": 0.9087952971458435,
      "learning_rate": 1e-05,
      "loss": 0.4866,
      "mean_token_accuracy": 0.8457107543945312,
      "num_tokens": 453691751.0,
      "step": 2847
    },
    {
      "epoch": 1.4486266531027467,
      "grad_norm": 0.9003292322158813,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8518234491348267,
      "num_tokens": 453856172.0,
      "step": 2848
    },
    {
      "epoch": 1.4491353001017293,
      "grad_norm": 0.9344632029533386,
      "learning_rate": 1e-05,
      "loss": 0.4716,
      "mean_token_accuracy": 0.8498226404190063,
      "num_tokens": 454018226.0,
      "step": 2849
    },
    {
      "epoch": 1.449643947100712,
      "grad_norm": 0.9255077242851257,
      "learning_rate": 1e-05,
      "loss": 0.4281,
      "mean_token_accuracy": 0.8616542816162109,
      "num_tokens": 454181489.0,
      "step": 2850
    },
    {
      "epoch": 1.4501525940996949,
      "grad_norm": 0.9817742109298706,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8535650968551636,
      "num_tokens": 454331972.0,
      "step": 2851
    },
    {
      "epoch": 1.4506612410986774,
      "grad_norm": 0.9063531756401062,
      "learning_rate": 1e-05,
      "loss": 0.5027,
      "mean_token_accuracy": 0.8408591747283936,
      "num_tokens": 454494258.0,
      "step": 2852
    },
    {
      "epoch": 1.4511698880976602,
      "grad_norm": 0.9713846445083618,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8623003363609314,
      "num_tokens": 454638672.0,
      "step": 2853
    },
    {
      "epoch": 1.4516785350966428,
      "grad_norm": 0.9173727631568909,
      "learning_rate": 1e-05,
      "loss": 0.4706,
      "mean_token_accuracy": 0.8480606079101562,
      "num_tokens": 454794488.0,
      "step": 2854
    },
    {
      "epoch": 1.4521871820956256,
      "grad_norm": 0.9979090094566345,
      "learning_rate": 1e-05,
      "loss": 0.5043,
      "mean_token_accuracy": 0.8406027555465698,
      "num_tokens": 454949085.0,
      "step": 2855
    },
    {
      "epoch": 1.4526958290946084,
      "grad_norm": 1.0404869318008423,
      "learning_rate": 1e-05,
      "loss": 0.4942,
      "mean_token_accuracy": 0.8417496085166931,
      "num_tokens": 455107150.0,
      "step": 2856
    },
    {
      "epoch": 1.4532044760935912,
      "grad_norm": 0.9395345449447632,
      "learning_rate": 1e-05,
      "loss": 0.4606,
      "mean_token_accuracy": 0.8518513441085815,
      "num_tokens": 455273881.0,
      "step": 2857
    },
    {
      "epoch": 1.4537131230925737,
      "grad_norm": 1.007689356803894,
      "learning_rate": 1e-05,
      "loss": 0.4746,
      "mean_token_accuracy": 0.8480565547943115,
      "num_tokens": 455429921.0,
      "step": 2858
    },
    {
      "epoch": 1.4542217700915565,
      "grad_norm": 0.9877992868423462,
      "learning_rate": 1e-05,
      "loss": 0.4968,
      "mean_token_accuracy": 0.8433884382247925,
      "num_tokens": 455589102.0,
      "step": 2859
    },
    {
      "epoch": 1.454730417090539,
      "grad_norm": 0.8932809829711914,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8504021763801575,
      "num_tokens": 455754141.0,
      "step": 2860
    },
    {
      "epoch": 1.4552390640895219,
      "grad_norm": 0.9541919827461243,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8518562912940979,
      "num_tokens": 455928739.0,
      "step": 2861
    },
    {
      "epoch": 1.4557477110885046,
      "grad_norm": 0.9939335584640503,
      "learning_rate": 1e-05,
      "loss": 0.4828,
      "mean_token_accuracy": 0.8472523093223572,
      "num_tokens": 456076223.0,
      "step": 2862
    },
    {
      "epoch": 1.4562563580874872,
      "grad_norm": 0.9644919037818909,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.8529149293899536,
      "num_tokens": 456244340.0,
      "step": 2863
    },
    {
      "epoch": 1.45676500508647,
      "grad_norm": 0.9466691613197327,
      "learning_rate": 1e-05,
      "loss": 0.4939,
      "mean_token_accuracy": 0.8433713912963867,
      "num_tokens": 456396940.0,
      "step": 2864
    },
    {
      "epoch": 1.4572736520854526,
      "grad_norm": 0.9644611477851868,
      "learning_rate": 1e-05,
      "loss": 0.4386,
      "mean_token_accuracy": 0.8583876490592957,
      "num_tokens": 456544680.0,
      "step": 2865
    },
    {
      "epoch": 1.4577822990844354,
      "grad_norm": 0.9246994256973267,
      "learning_rate": 1e-05,
      "loss": 0.482,
      "mean_token_accuracy": 0.8474558591842651,
      "num_tokens": 456705322.0,
      "step": 2866
    },
    {
      "epoch": 1.4582909460834181,
      "grad_norm": 0.9289038181304932,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8451316356658936,
      "num_tokens": 456868645.0,
      "step": 2867
    },
    {
      "epoch": 1.458799593082401,
      "grad_norm": 1.0041005611419678,
      "learning_rate": 1e-05,
      "loss": 0.4812,
      "mean_token_accuracy": 0.8481215238571167,
      "num_tokens": 457034824.0,
      "step": 2868
    },
    {
      "epoch": 1.4593082400813835,
      "grad_norm": 0.9717512130737305,
      "learning_rate": 1e-05,
      "loss": 0.5092,
      "mean_token_accuracy": 0.8409024477005005,
      "num_tokens": 457207183.0,
      "step": 2869
    },
    {
      "epoch": 1.4598168870803663,
      "grad_norm": 1.0197564363479614,
      "learning_rate": 1e-05,
      "loss": 0.4877,
      "mean_token_accuracy": 0.846825361251831,
      "num_tokens": 457358314.0,
      "step": 2870
    },
    {
      "epoch": 1.4603255340793488,
      "grad_norm": 0.9384062886238098,
      "learning_rate": 1e-05,
      "loss": 0.4532,
      "mean_token_accuracy": 0.8543113470077515,
      "num_tokens": 457523034.0,
      "step": 2871
    },
    {
      "epoch": 1.4608341810783316,
      "grad_norm": 1.0533404350280762,
      "learning_rate": 1e-05,
      "loss": 0.4765,
      "mean_token_accuracy": 0.8486883640289307,
      "num_tokens": 457677702.0,
      "step": 2872
    },
    {
      "epoch": 1.4613428280773144,
      "grad_norm": 0.9609192609786987,
      "learning_rate": 1e-05,
      "loss": 0.4841,
      "mean_token_accuracy": 0.8472496271133423,
      "num_tokens": 457841186.0,
      "step": 2873
    },
    {
      "epoch": 1.461851475076297,
      "grad_norm": 0.913457453250885,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8447855114936829,
      "num_tokens": 458004068.0,
      "step": 2874
    },
    {
      "epoch": 1.4623601220752798,
      "grad_norm": 1.0519827604293823,
      "learning_rate": 1e-05,
      "loss": 0.5015,
      "mean_token_accuracy": 0.8410388827323914,
      "num_tokens": 458171129.0,
      "step": 2875
    },
    {
      "epoch": 1.4628687690742623,
      "grad_norm": 0.8884391784667969,
      "learning_rate": 1e-05,
      "loss": 0.4743,
      "mean_token_accuracy": 0.8498002886772156,
      "num_tokens": 458327986.0,
      "step": 2876
    },
    {
      "epoch": 1.4633774160732451,
      "grad_norm": 0.9776370525360107,
      "learning_rate": 1e-05,
      "loss": 0.4947,
      "mean_token_accuracy": 0.8417224884033203,
      "num_tokens": 458492443.0,
      "step": 2877
    },
    {
      "epoch": 1.463886063072228,
      "grad_norm": 0.906736433506012,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8565185070037842,
      "num_tokens": 458641744.0,
      "step": 2878
    },
    {
      "epoch": 1.4643947100712107,
      "grad_norm": 0.9633562564849854,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.8528835773468018,
      "num_tokens": 458800987.0,
      "step": 2879
    },
    {
      "epoch": 1.4649033570701933,
      "grad_norm": 0.9794289469718933,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8491966724395752,
      "num_tokens": 458963974.0,
      "step": 2880
    },
    {
      "epoch": 1.465412004069176,
      "grad_norm": 0.9550290703773499,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8509371280670166,
      "num_tokens": 459117725.0,
      "step": 2881
    },
    {
      "epoch": 1.4659206510681586,
      "grad_norm": 1.0013341903686523,
      "learning_rate": 1e-05,
      "loss": 0.4812,
      "mean_token_accuracy": 0.8475896120071411,
      "num_tokens": 459273431.0,
      "step": 2882
    },
    {
      "epoch": 1.4664292980671414,
      "grad_norm": 0.9773209691047668,
      "learning_rate": 1e-05,
      "loss": 0.4775,
      "mean_token_accuracy": 0.848754346370697,
      "num_tokens": 459436131.0,
      "step": 2883
    },
    {
      "epoch": 1.4669379450661242,
      "grad_norm": 1.0352989435195923,
      "learning_rate": 1e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.849433422088623,
      "num_tokens": 459590197.0,
      "step": 2884
    },
    {
      "epoch": 1.4674465920651067,
      "grad_norm": 1.086367130279541,
      "learning_rate": 1e-05,
      "loss": 0.4758,
      "mean_token_accuracy": 0.8498767614364624,
      "num_tokens": 459743580.0,
      "step": 2885
    },
    {
      "epoch": 1.4679552390640895,
      "grad_norm": 0.8803079128265381,
      "learning_rate": 1e-05,
      "loss": 0.4811,
      "mean_token_accuracy": 0.8476182222366333,
      "num_tokens": 459898577.0,
      "step": 2886
    },
    {
      "epoch": 1.468463886063072,
      "grad_norm": 1.1584640741348267,
      "learning_rate": 1e-05,
      "loss": 0.4915,
      "mean_token_accuracy": 0.8442654609680176,
      "num_tokens": 460056967.0,
      "step": 2887
    },
    {
      "epoch": 1.4689725330620549,
      "grad_norm": 1.0022369623184204,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8600075244903564,
      "num_tokens": 460209370.0,
      "step": 2888
    },
    {
      "epoch": 1.4694811800610377,
      "grad_norm": 1.0320205688476562,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.859348475933075,
      "num_tokens": 460373485.0,
      "step": 2889
    },
    {
      "epoch": 1.4699898270600205,
      "grad_norm": 1.0857659578323364,
      "learning_rate": 1e-05,
      "loss": 0.4768,
      "mean_token_accuracy": 0.847952127456665,
      "num_tokens": 460538263.0,
      "step": 2890
    },
    {
      "epoch": 1.470498474059003,
      "grad_norm": 1.0369950532913208,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.846710741519928,
      "num_tokens": 460698779.0,
      "step": 2891
    },
    {
      "epoch": 1.4710071210579858,
      "grad_norm": 1.0814622640609741,
      "learning_rate": 1e-05,
      "loss": 0.4706,
      "mean_token_accuracy": 0.8494322299957275,
      "num_tokens": 460866578.0,
      "step": 2892
    },
    {
      "epoch": 1.4715157680569684,
      "grad_norm": 0.9672597646713257,
      "learning_rate": 1e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.8501402139663696,
      "num_tokens": 461012557.0,
      "step": 2893
    },
    {
      "epoch": 1.4720244150559512,
      "grad_norm": 1.0468100309371948,
      "learning_rate": 1e-05,
      "loss": 0.5112,
      "mean_token_accuracy": 0.8384190797805786,
      "num_tokens": 461171630.0,
      "step": 2894
    },
    {
      "epoch": 1.472533062054934,
      "grad_norm": 0.9749298691749573,
      "learning_rate": 1e-05,
      "loss": 0.4624,
      "mean_token_accuracy": 0.852161705493927,
      "num_tokens": 461336078.0,
      "step": 2895
    },
    {
      "epoch": 1.4730417090539165,
      "grad_norm": 0.9439693093299866,
      "learning_rate": 1e-05,
      "loss": 0.4835,
      "mean_token_accuracy": 0.844796359539032,
      "num_tokens": 461506509.0,
      "step": 2896
    },
    {
      "epoch": 1.4735503560528993,
      "grad_norm": 0.9988195896148682,
      "learning_rate": 1e-05,
      "loss": 0.4394,
      "mean_token_accuracy": 0.860859751701355,
      "num_tokens": 461677135.0,
      "step": 2897
    },
    {
      "epoch": 1.4740590030518819,
      "grad_norm": 0.9463707804679871,
      "learning_rate": 1e-05,
      "loss": 0.434,
      "mean_token_accuracy": 0.8603488802909851,
      "num_tokens": 461835821.0,
      "step": 2898
    },
    {
      "epoch": 1.4745676500508647,
      "grad_norm": 1.0607638359069824,
      "learning_rate": 1e-05,
      "loss": 0.459,
      "mean_token_accuracy": 0.8542230725288391,
      "num_tokens": 461982449.0,
      "step": 2899
    },
    {
      "epoch": 1.4750762970498474,
      "grad_norm": 0.9559944868087769,
      "learning_rate": 1e-05,
      "loss": 0.4636,
      "mean_token_accuracy": 0.85247403383255,
      "num_tokens": 462141478.0,
      "step": 2900
    },
    {
      "epoch": 1.4755849440488302,
      "grad_norm": 0.9273171424865723,
      "learning_rate": 1e-05,
      "loss": 0.4469,
      "mean_token_accuracy": 0.8564932346343994,
      "num_tokens": 462308851.0,
      "step": 2901
    },
    {
      "epoch": 1.4760935910478128,
      "grad_norm": 0.932365894317627,
      "learning_rate": 1e-05,
      "loss": 0.4631,
      "mean_token_accuracy": 0.8522019386291504,
      "num_tokens": 462469829.0,
      "step": 2902
    },
    {
      "epoch": 1.4766022380467956,
      "grad_norm": 1.1404035091400146,
      "learning_rate": 1e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.8589218258857727,
      "num_tokens": 462635407.0,
      "step": 2903
    },
    {
      "epoch": 1.4771108850457781,
      "grad_norm": 1.0725706815719604,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8407667875289917,
      "num_tokens": 462805295.0,
      "step": 2904
    },
    {
      "epoch": 1.477619532044761,
      "grad_norm": 0.9216992855072021,
      "learning_rate": 1e-05,
      "loss": 0.4468,
      "mean_token_accuracy": 0.8585999011993408,
      "num_tokens": 462971691.0,
      "step": 2905
    },
    {
      "epoch": 1.4781281790437437,
      "grad_norm": 0.9899900555610657,
      "learning_rate": 1e-05,
      "loss": 0.4804,
      "mean_token_accuracy": 0.8486272096633911,
      "num_tokens": 463137500.0,
      "step": 2906
    },
    {
      "epoch": 1.4786368260427263,
      "grad_norm": 1.0153653621673584,
      "learning_rate": 1e-05,
      "loss": 0.478,
      "mean_token_accuracy": 0.8489232659339905,
      "num_tokens": 463288904.0,
      "step": 2907
    },
    {
      "epoch": 1.479145473041709,
      "grad_norm": 1.4542001485824585,
      "learning_rate": 1e-05,
      "loss": 0.5027,
      "mean_token_accuracy": 0.8426099419593811,
      "num_tokens": 463448939.0,
      "step": 2908
    },
    {
      "epoch": 1.4796541200406916,
      "grad_norm": 1.0369772911071777,
      "learning_rate": 1e-05,
      "loss": 0.4975,
      "mean_token_accuracy": 0.8416335582733154,
      "num_tokens": 463601841.0,
      "step": 2909
    },
    {
      "epoch": 1.4801627670396744,
      "grad_norm": 0.9426633715629578,
      "learning_rate": 1e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8484392166137695,
      "num_tokens": 463759585.0,
      "step": 2910
    },
    {
      "epoch": 1.4806714140386572,
      "grad_norm": 1.0142484903335571,
      "learning_rate": 1e-05,
      "loss": 0.466,
      "mean_token_accuracy": 0.8509640097618103,
      "num_tokens": 463928168.0,
      "step": 2911
    },
    {
      "epoch": 1.48118006103764,
      "grad_norm": 0.9512984752655029,
      "learning_rate": 1e-05,
      "loss": 0.4816,
      "mean_token_accuracy": 0.8469193577766418,
      "num_tokens": 464082388.0,
      "step": 2912
    },
    {
      "epoch": 1.4816887080366226,
      "grad_norm": 0.977675199508667,
      "learning_rate": 1e-05,
      "loss": 0.4683,
      "mean_token_accuracy": 0.8513289093971252,
      "num_tokens": 464242956.0,
      "step": 2913
    },
    {
      "epoch": 1.4821973550356053,
      "grad_norm": 0.9527803659439087,
      "learning_rate": 1e-05,
      "loss": 0.4922,
      "mean_token_accuracy": 0.8448150157928467,
      "num_tokens": 464417202.0,
      "step": 2914
    },
    {
      "epoch": 1.482706002034588,
      "grad_norm": 0.9302776455879211,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8493098020553589,
      "num_tokens": 464570429.0,
      "step": 2915
    },
    {
      "epoch": 1.4832146490335707,
      "grad_norm": 0.9582393765449524,
      "learning_rate": 1e-05,
      "loss": 0.461,
      "mean_token_accuracy": 0.8533623218536377,
      "num_tokens": 464725276.0,
      "step": 2916
    },
    {
      "epoch": 1.4837232960325535,
      "grad_norm": 0.9441304206848145,
      "learning_rate": 1e-05,
      "loss": 0.4636,
      "mean_token_accuracy": 0.8537716865539551,
      "num_tokens": 464893941.0,
      "step": 2917
    },
    {
      "epoch": 1.484231943031536,
      "grad_norm": 0.9462320804595947,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8577545285224915,
      "num_tokens": 465058613.0,
      "step": 2918
    },
    {
      "epoch": 1.4847405900305188,
      "grad_norm": 0.9499002695083618,
      "learning_rate": 1e-05,
      "loss": 0.4862,
      "mean_token_accuracy": 0.8443703651428223,
      "num_tokens": 465218139.0,
      "step": 2919
    },
    {
      "epoch": 1.4852492370295014,
      "grad_norm": 0.958915650844574,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8586746454238892,
      "num_tokens": 465376097.0,
      "step": 2920
    },
    {
      "epoch": 1.4857578840284842,
      "grad_norm": 0.9286035895347595,
      "learning_rate": 1e-05,
      "loss": 0.5144,
      "mean_token_accuracy": 0.8386704921722412,
      "num_tokens": 465532660.0,
      "step": 2921
    },
    {
      "epoch": 1.486266531027467,
      "grad_norm": 0.983659029006958,
      "learning_rate": 1e-05,
      "loss": 0.4876,
      "mean_token_accuracy": 0.8438690900802612,
      "num_tokens": 465702748.0,
      "step": 2922
    },
    {
      "epoch": 1.4867751780264498,
      "grad_norm": 0.9186338782310486,
      "learning_rate": 1e-05,
      "loss": 0.4297,
      "mean_token_accuracy": 0.8618059158325195,
      "num_tokens": 465854774.0,
      "step": 2923
    },
    {
      "epoch": 1.4872838250254323,
      "grad_norm": 1.0261900424957275,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8528866171836853,
      "num_tokens": 466007913.0,
      "step": 2924
    },
    {
      "epoch": 1.4877924720244151,
      "grad_norm": 0.9390912055969238,
      "learning_rate": 1e-05,
      "loss": 0.4954,
      "mean_token_accuracy": 0.8455357551574707,
      "num_tokens": 466164365.0,
      "step": 2925
    },
    {
      "epoch": 1.4883011190233977,
      "grad_norm": 0.8921604156494141,
      "learning_rate": 1e-05,
      "loss": 0.4732,
      "mean_token_accuracy": 0.849597692489624,
      "num_tokens": 466327362.0,
      "step": 2926
    },
    {
      "epoch": 1.4888097660223805,
      "grad_norm": 0.989403247833252,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8581273555755615,
      "num_tokens": 466477193.0,
      "step": 2927
    },
    {
      "epoch": 1.4893184130213633,
      "grad_norm": 0.944860577583313,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8537037372589111,
      "num_tokens": 466638743.0,
      "step": 2928
    },
    {
      "epoch": 1.4898270600203458,
      "grad_norm": 1.0039745569229126,
      "learning_rate": 1e-05,
      "loss": 0.457,
      "mean_token_accuracy": 0.8552755117416382,
      "num_tokens": 466797661.0,
      "step": 2929
    },
    {
      "epoch": 1.4903357070193286,
      "grad_norm": 0.9398649334907532,
      "learning_rate": 1e-05,
      "loss": 0.492,
      "mean_token_accuracy": 0.8440238237380981,
      "num_tokens": 466958501.0,
      "step": 2930
    },
    {
      "epoch": 1.4908443540183112,
      "grad_norm": 1.174691915512085,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8575791120529175,
      "num_tokens": 467110269.0,
      "step": 2931
    },
    {
      "epoch": 1.491353001017294,
      "grad_norm": 1.0185178518295288,
      "learning_rate": 1e-05,
      "loss": 0.5289,
      "mean_token_accuracy": 0.8333423733711243,
      "num_tokens": 467271547.0,
      "step": 2932
    },
    {
      "epoch": 1.4918616480162767,
      "grad_norm": 0.8953125476837158,
      "learning_rate": 1e-05,
      "loss": 0.4392,
      "mean_token_accuracy": 0.8588826656341553,
      "num_tokens": 467440161.0,
      "step": 2933
    },
    {
      "epoch": 1.4923702950152595,
      "grad_norm": 0.9780067801475525,
      "learning_rate": 1e-05,
      "loss": 0.4829,
      "mean_token_accuracy": 0.8456743359565735,
      "num_tokens": 467597696.0,
      "step": 2934
    },
    {
      "epoch": 1.492878942014242,
      "grad_norm": 0.9480140209197998,
      "learning_rate": 1e-05,
      "loss": 0.4682,
      "mean_token_accuracy": 0.8502559661865234,
      "num_tokens": 467763338.0,
      "step": 2935
    },
    {
      "epoch": 1.4933875890132249,
      "grad_norm": 0.9286141395568848,
      "learning_rate": 1e-05,
      "loss": 0.4871,
      "mean_token_accuracy": 0.8461545705795288,
      "num_tokens": 467925187.0,
      "step": 2936
    },
    {
      "epoch": 1.4938962360122074,
      "grad_norm": 1.008232593536377,
      "learning_rate": 1e-05,
      "loss": 0.5066,
      "mean_token_accuracy": 0.8377925157546997,
      "num_tokens": 468077219.0,
      "step": 2937
    },
    {
      "epoch": 1.4944048830111902,
      "grad_norm": 1.0144633054733276,
      "learning_rate": 1e-05,
      "loss": 0.488,
      "mean_token_accuracy": 0.8458861112594604,
      "num_tokens": 468226973.0,
      "step": 2938
    },
    {
      "epoch": 1.494913530010173,
      "grad_norm": 1.2867398262023926,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.8544325828552246,
      "num_tokens": 468386458.0,
      "step": 2939
    },
    {
      "epoch": 1.4954221770091556,
      "grad_norm": 0.9466829299926758,
      "learning_rate": 1e-05,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8576674461364746,
      "num_tokens": 468534657.0,
      "step": 2940
    },
    {
      "epoch": 1.4959308240081384,
      "grad_norm": 0.9843382239341736,
      "learning_rate": 1e-05,
      "loss": 0.4552,
      "mean_token_accuracy": 0.855600118637085,
      "num_tokens": 468699181.0,
      "step": 2941
    },
    {
      "epoch": 1.496439471007121,
      "grad_norm": 0.9357199668884277,
      "learning_rate": 1e-05,
      "loss": 0.5075,
      "mean_token_accuracy": 0.8387920260429382,
      "num_tokens": 468863037.0,
      "step": 2942
    },
    {
      "epoch": 1.4969481180061037,
      "grad_norm": 0.8901887536048889,
      "learning_rate": 1e-05,
      "loss": 0.4801,
      "mean_token_accuracy": 0.8496588468551636,
      "num_tokens": 469024753.0,
      "step": 2943
    },
    {
      "epoch": 1.4974567650050865,
      "grad_norm": 1.0716995000839233,
      "learning_rate": 1e-05,
      "loss": 0.4935,
      "mean_token_accuracy": 0.8449990153312683,
      "num_tokens": 469175522.0,
      "step": 2944
    },
    {
      "epoch": 1.4979654120040693,
      "grad_norm": 0.918582022190094,
      "learning_rate": 1e-05,
      "loss": 0.4658,
      "mean_token_accuracy": 0.8516040444374084,
      "num_tokens": 469335707.0,
      "step": 2945
    },
    {
      "epoch": 1.4984740590030519,
      "grad_norm": 1.4671798944473267,
      "learning_rate": 1e-05,
      "loss": 0.4507,
      "mean_token_accuracy": 0.8577188849449158,
      "num_tokens": 469489877.0,
      "step": 2946
    },
    {
      "epoch": 1.4989827060020346,
      "grad_norm": 1.0488964319229126,
      "learning_rate": 1e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8407468795776367,
      "num_tokens": 469649523.0,
      "step": 2947
    },
    {
      "epoch": 1.4994913530010172,
      "grad_norm": 0.928181529045105,
      "learning_rate": 1e-05,
      "loss": 0.4671,
      "mean_token_accuracy": 0.8499090671539307,
      "num_tokens": 469823003.0,
      "step": 2948
    },
    {
      "epoch": 1.5,
      "grad_norm": 1.0365320444107056,
      "learning_rate": 1e-05,
      "loss": 0.4836,
      "mean_token_accuracy": 0.8448408842086792,
      "num_tokens": 469987692.0,
      "step": 2949
    },
    {
      "epoch": 1.5005086469989828,
      "grad_norm": 1.0148674249649048,
      "learning_rate": 1e-05,
      "loss": 0.4765,
      "mean_token_accuracy": 0.8493233919143677,
      "num_tokens": 470149249.0,
      "step": 2950
    },
    {
      "epoch": 1.5010172939979656,
      "grad_norm": 0.9570903182029724,
      "learning_rate": 1e-05,
      "loss": 0.4921,
      "mean_token_accuracy": 0.844947099685669,
      "num_tokens": 470300640.0,
      "step": 2951
    },
    {
      "epoch": 1.5015259409969481,
      "grad_norm": 1.0189590454101562,
      "learning_rate": 1e-05,
      "loss": 0.5228,
      "mean_token_accuracy": 0.8395256400108337,
      "num_tokens": 470472689.0,
      "step": 2952
    },
    {
      "epoch": 1.5020345879959307,
      "grad_norm": 1.0926448106765747,
      "learning_rate": 1e-05,
      "loss": 0.4706,
      "mean_token_accuracy": 0.8507436513900757,
      "num_tokens": 470632121.0,
      "step": 2953
    },
    {
      "epoch": 1.5025432349949135,
      "grad_norm": 0.9752507209777832,
      "learning_rate": 1e-05,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8419481515884399,
      "num_tokens": 470771648.0,
      "step": 2954
    },
    {
      "epoch": 1.5030518819938963,
      "grad_norm": 0.921377956867218,
      "learning_rate": 1e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.8515791893005371,
      "num_tokens": 470935021.0,
      "step": 2955
    },
    {
      "epoch": 1.503560528992879,
      "grad_norm": 0.9844980835914612,
      "learning_rate": 1e-05,
      "loss": 0.4852,
      "mean_token_accuracy": 0.8456733226776123,
      "num_tokens": 471089333.0,
      "step": 2956
    },
    {
      "epoch": 1.5040691759918616,
      "grad_norm": 1.0702826976776123,
      "learning_rate": 1e-05,
      "loss": 0.4828,
      "mean_token_accuracy": 0.8475288152694702,
      "num_tokens": 471253508.0,
      "step": 2957
    },
    {
      "epoch": 1.5045778229908442,
      "grad_norm": 0.9100438356399536,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8491068482398987,
      "num_tokens": 471422317.0,
      "step": 2958
    },
    {
      "epoch": 1.505086469989827,
      "grad_norm": 1.096585988998413,
      "learning_rate": 1e-05,
      "loss": 0.4363,
      "mean_token_accuracy": 0.8580671548843384,
      "num_tokens": 471571311.0,
      "step": 2959
    },
    {
      "epoch": 1.5055951169888098,
      "grad_norm": 1.007280707359314,
      "learning_rate": 1e-05,
      "loss": 0.5051,
      "mean_token_accuracy": 0.8421881794929504,
      "num_tokens": 471730047.0,
      "step": 2960
    },
    {
      "epoch": 1.5061037639877926,
      "grad_norm": 1.070167899131775,
      "learning_rate": 1e-05,
      "loss": 0.5001,
      "mean_token_accuracy": 0.8414835333824158,
      "num_tokens": 471891342.0,
      "step": 2961
    },
    {
      "epoch": 1.5066124109867753,
      "grad_norm": 1.114966630935669,
      "learning_rate": 1e-05,
      "loss": 0.4803,
      "mean_token_accuracy": 0.8474112749099731,
      "num_tokens": 472056255.0,
      "step": 2962
    },
    {
      "epoch": 1.507121057985758,
      "grad_norm": 0.967980682849884,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8452816009521484,
      "num_tokens": 472225606.0,
      "step": 2963
    },
    {
      "epoch": 1.5076297049847405,
      "grad_norm": 0.9875637292861938,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.853312611579895,
      "num_tokens": 472385581.0,
      "step": 2964
    },
    {
      "epoch": 1.5081383519837233,
      "grad_norm": 1.0010428428649902,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8569035530090332,
      "num_tokens": 472549805.0,
      "step": 2965
    },
    {
      "epoch": 1.508646998982706,
      "grad_norm": 1.1223161220550537,
      "learning_rate": 1e-05,
      "loss": 0.4737,
      "mean_token_accuracy": 0.849111795425415,
      "num_tokens": 472697254.0,
      "step": 2966
    },
    {
      "epoch": 1.5091556459816888,
      "grad_norm": 1.1635527610778809,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8492852449417114,
      "num_tokens": 472853595.0,
      "step": 2967
    },
    {
      "epoch": 1.5096642929806714,
      "grad_norm": 0.9469373822212219,
      "learning_rate": 1e-05,
      "loss": 0.5128,
      "mean_token_accuracy": 0.8389683961868286,
      "num_tokens": 473029719.0,
      "step": 2968
    },
    {
      "epoch": 1.510172939979654,
      "grad_norm": 1.0621305704116821,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8480264544487,
      "num_tokens": 473188682.0,
      "step": 2969
    },
    {
      "epoch": 1.5106815869786367,
      "grad_norm": 0.9598436951637268,
      "learning_rate": 1e-05,
      "loss": 0.4881,
      "mean_token_accuracy": 0.8451364040374756,
      "num_tokens": 473348305.0,
      "step": 2970
    },
    {
      "epoch": 1.5111902339776195,
      "grad_norm": 0.9786617159843445,
      "learning_rate": 1e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.8512166738510132,
      "num_tokens": 473502448.0,
      "step": 2971
    },
    {
      "epoch": 1.5116988809766023,
      "grad_norm": 0.9746079444885254,
      "learning_rate": 1e-05,
      "loss": 0.4894,
      "mean_token_accuracy": 0.8454301953315735,
      "num_tokens": 473666571.0,
      "step": 2972
    },
    {
      "epoch": 1.512207527975585,
      "grad_norm": 0.9722834825515747,
      "learning_rate": 1e-05,
      "loss": 0.4342,
      "mean_token_accuracy": 0.8606178760528564,
      "num_tokens": 473823460.0,
      "step": 2973
    },
    {
      "epoch": 1.5127161749745677,
      "grad_norm": 0.9689220190048218,
      "learning_rate": 1e-05,
      "loss": 0.4749,
      "mean_token_accuracy": 0.8500169515609741,
      "num_tokens": 473976615.0,
      "step": 2974
    },
    {
      "epoch": 1.5132248219735502,
      "grad_norm": 1.01953125,
      "learning_rate": 1e-05,
      "loss": 0.43,
      "mean_token_accuracy": 0.8606699705123901,
      "num_tokens": 474127194.0,
      "step": 2975
    },
    {
      "epoch": 1.513733468972533,
      "grad_norm": 0.9314214587211609,
      "learning_rate": 1e-05,
      "loss": 0.4925,
      "mean_token_accuracy": 0.8451440334320068,
      "num_tokens": 474309906.0,
      "step": 2976
    },
    {
      "epoch": 1.5142421159715158,
      "grad_norm": 0.9425230026245117,
      "learning_rate": 1e-05,
      "loss": 0.4732,
      "mean_token_accuracy": 0.8487437963485718,
      "num_tokens": 474468000.0,
      "step": 2977
    },
    {
      "epoch": 1.5147507629704986,
      "grad_norm": 1.0697709321975708,
      "learning_rate": 1e-05,
      "loss": 0.502,
      "mean_token_accuracy": 0.8403645753860474,
      "num_tokens": 474620054.0,
      "step": 2978
    },
    {
      "epoch": 1.5152594099694812,
      "grad_norm": 0.9415473937988281,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8491325974464417,
      "num_tokens": 474780460.0,
      "step": 2979
    },
    {
      "epoch": 1.5157680569684637,
      "grad_norm": 0.9371607899665833,
      "learning_rate": 1e-05,
      "loss": 0.4792,
      "mean_token_accuracy": 0.8479374647140503,
      "num_tokens": 474940167.0,
      "step": 2980
    },
    {
      "epoch": 1.5162767039674465,
      "grad_norm": 0.9356675148010254,
      "learning_rate": 1e-05,
      "loss": 0.458,
      "mean_token_accuracy": 0.8535172939300537,
      "num_tokens": 475100842.0,
      "step": 2981
    },
    {
      "epoch": 1.5167853509664293,
      "grad_norm": 0.9361054301261902,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8528018593788147,
      "num_tokens": 475258659.0,
      "step": 2982
    },
    {
      "epoch": 1.517293997965412,
      "grad_norm": 0.9554124474525452,
      "learning_rate": 1e-05,
      "loss": 0.4714,
      "mean_token_accuracy": 0.8502927422523499,
      "num_tokens": 475420431.0,
      "step": 2983
    },
    {
      "epoch": 1.5178026449643949,
      "grad_norm": 0.9128940105438232,
      "learning_rate": 1e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.8527088165283203,
      "num_tokens": 475577873.0,
      "step": 2984
    },
    {
      "epoch": 1.5183112919633774,
      "grad_norm": 0.9470569491386414,
      "learning_rate": 1e-05,
      "loss": 0.4894,
      "mean_token_accuracy": 0.843989372253418,
      "num_tokens": 475739554.0,
      "step": 2985
    },
    {
      "epoch": 1.51881993896236,
      "grad_norm": 0.9532779455184937,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8447748422622681,
      "num_tokens": 475909298.0,
      "step": 2986
    },
    {
      "epoch": 1.5193285859613428,
      "grad_norm": 0.9041764736175537,
      "learning_rate": 1e-05,
      "loss": 0.5138,
      "mean_token_accuracy": 0.83664470911026,
      "num_tokens": 476081499.0,
      "step": 2987
    },
    {
      "epoch": 1.5198372329603256,
      "grad_norm": 0.8667990565299988,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.8529576063156128,
      "num_tokens": 476240775.0,
      "step": 2988
    },
    {
      "epoch": 1.5203458799593084,
      "grad_norm": 0.9188169836997986,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8563479781150818,
      "num_tokens": 476404318.0,
      "step": 2989
    },
    {
      "epoch": 1.520854526958291,
      "grad_norm": 1.072871208190918,
      "learning_rate": 1e-05,
      "loss": 0.506,
      "mean_token_accuracy": 0.8398078680038452,
      "num_tokens": 476556580.0,
      "step": 2990
    },
    {
      "epoch": 1.5213631739572735,
      "grad_norm": 0.9862761497497559,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8528329730033875,
      "num_tokens": 476695113.0,
      "step": 2991
    },
    {
      "epoch": 1.5218718209562563,
      "grad_norm": 0.9302590489387512,
      "learning_rate": 1e-05,
      "loss": 0.4677,
      "mean_token_accuracy": 0.8522852659225464,
      "num_tokens": 476855526.0,
      "step": 2992
    },
    {
      "epoch": 1.522380467955239,
      "grad_norm": 1.0184320211410522,
      "learning_rate": 1e-05,
      "loss": 0.4701,
      "mean_token_accuracy": 0.8505998253822327,
      "num_tokens": 477014468.0,
      "step": 2993
    },
    {
      "epoch": 1.5228891149542219,
      "grad_norm": 0.9482423067092896,
      "learning_rate": 1e-05,
      "loss": 0.4907,
      "mean_token_accuracy": 0.8458576202392578,
      "num_tokens": 477177840.0,
      "step": 2994
    },
    {
      "epoch": 1.5233977619532044,
      "grad_norm": 1.9431716203689575,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8519662618637085,
      "num_tokens": 477338479.0,
      "step": 2995
    },
    {
      "epoch": 1.5239064089521872,
      "grad_norm": 0.9888985753059387,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8544604778289795,
      "num_tokens": 477502383.0,
      "step": 2996
    },
    {
      "epoch": 1.5244150559511698,
      "grad_norm": 0.9551851153373718,
      "learning_rate": 1e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8576740026473999,
      "num_tokens": 477664750.0,
      "step": 2997
    },
    {
      "epoch": 1.5249237029501526,
      "grad_norm": 0.9770562648773193,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.8444640636444092,
      "num_tokens": 477826492.0,
      "step": 2998
    },
    {
      "epoch": 1.5254323499491353,
      "grad_norm": 1.3451825380325317,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "mean_token_accuracy": 0.8568999767303467,
      "num_tokens": 477974878.0,
      "step": 2999
    },
    {
      "epoch": 1.5259409969481181,
      "grad_norm": 1.1307803392410278,
      "learning_rate": 1e-05,
      "loss": 0.4546,
      "mean_token_accuracy": 0.8543623685836792,
      "num_tokens": 478124029.0,
      "step": 3000
    },
    {
      "epoch": 1.5264496439471007,
      "grad_norm": 0.9587520360946655,
      "learning_rate": 1e-05,
      "loss": 0.46,
      "mean_token_accuracy": 0.8539354801177979,
      "num_tokens": 478282115.0,
      "step": 3001
    },
    {
      "epoch": 1.5269582909460833,
      "grad_norm": 0.9479132890701294,
      "learning_rate": 1e-05,
      "loss": 0.4697,
      "mean_token_accuracy": 0.8511796593666077,
      "num_tokens": 478451961.0,
      "step": 3002
    },
    {
      "epoch": 1.527466937945066,
      "grad_norm": 1.134979486465454,
      "learning_rate": 1e-05,
      "loss": 0.4863,
      "mean_token_accuracy": 0.8461675643920898,
      "num_tokens": 478608587.0,
      "step": 3003
    },
    {
      "epoch": 1.5279755849440488,
      "grad_norm": 0.8995083570480347,
      "learning_rate": 1e-05,
      "loss": 0.4951,
      "mean_token_accuracy": 0.8425230979919434,
      "num_tokens": 478777570.0,
      "step": 3004
    },
    {
      "epoch": 1.5284842319430316,
      "grad_norm": 1.0487364530563354,
      "learning_rate": 1e-05,
      "loss": 0.4638,
      "mean_token_accuracy": 0.8535329103469849,
      "num_tokens": 478927415.0,
      "step": 3005
    },
    {
      "epoch": 1.5289928789420142,
      "grad_norm": 0.9376885294914246,
      "learning_rate": 1e-05,
      "loss": 0.4691,
      "mean_token_accuracy": 0.8515335917472839,
      "num_tokens": 479091694.0,
      "step": 3006
    },
    {
      "epoch": 1.529501525940997,
      "grad_norm": 1.0092753171920776,
      "learning_rate": 1e-05,
      "loss": 0.4915,
      "mean_token_accuracy": 0.8434138298034668,
      "num_tokens": 479247246.0,
      "step": 3007
    },
    {
      "epoch": 1.5300101729399795,
      "grad_norm": 0.9187234044075012,
      "learning_rate": 1e-05,
      "loss": 0.4664,
      "mean_token_accuracy": 0.8512427806854248,
      "num_tokens": 479395630.0,
      "step": 3008
    },
    {
      "epoch": 1.5305188199389623,
      "grad_norm": 0.9388754367828369,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8556320071220398,
      "num_tokens": 479557400.0,
      "step": 3009
    },
    {
      "epoch": 1.5310274669379451,
      "grad_norm": 0.9462489485740662,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8584042191505432,
      "num_tokens": 479705902.0,
      "step": 3010
    },
    {
      "epoch": 1.531536113936928,
      "grad_norm": 0.9888403415679932,
      "learning_rate": 1e-05,
      "loss": 0.4769,
      "mean_token_accuracy": 0.8489364385604858,
      "num_tokens": 479864999.0,
      "step": 3011
    },
    {
      "epoch": 1.5320447609359105,
      "grad_norm": 0.9582159519195557,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.852144718170166,
      "num_tokens": 480016304.0,
      "step": 3012
    },
    {
      "epoch": 1.532553407934893,
      "grad_norm": 0.9440920352935791,
      "learning_rate": 1e-05,
      "loss": 0.4821,
      "mean_token_accuracy": 0.8469336032867432,
      "num_tokens": 480184076.0,
      "step": 3013
    },
    {
      "epoch": 1.5330620549338758,
      "grad_norm": 0.9551937580108643,
      "learning_rate": 1e-05,
      "loss": 0.4392,
      "mean_token_accuracy": 0.8583161234855652,
      "num_tokens": 480346413.0,
      "step": 3014
    },
    {
      "epoch": 1.5335707019328586,
      "grad_norm": 0.9670236110687256,
      "learning_rate": 1e-05,
      "loss": 0.4395,
      "mean_token_accuracy": 0.8601056337356567,
      "num_tokens": 480496300.0,
      "step": 3015
    },
    {
      "epoch": 1.5340793489318414,
      "grad_norm": 0.961797297000885,
      "learning_rate": 1e-05,
      "loss": 0.499,
      "mean_token_accuracy": 0.841756284236908,
      "num_tokens": 480658234.0,
      "step": 3016
    },
    {
      "epoch": 1.534587995930824,
      "grad_norm": 0.9484275579452515,
      "learning_rate": 1e-05,
      "loss": 0.4755,
      "mean_token_accuracy": 0.8482771515846252,
      "num_tokens": 480815704.0,
      "step": 3017
    },
    {
      "epoch": 1.5350966429298067,
      "grad_norm": 1.072632908821106,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8507258892059326,
      "num_tokens": 480980014.0,
      "step": 3018
    },
    {
      "epoch": 1.5356052899287893,
      "grad_norm": 0.8930185437202454,
      "learning_rate": 1e-05,
      "loss": 0.4989,
      "mean_token_accuracy": 0.8436600565910339,
      "num_tokens": 481147388.0,
      "step": 3019
    },
    {
      "epoch": 1.536113936927772,
      "grad_norm": 0.9993252158164978,
      "learning_rate": 1e-05,
      "loss": 0.512,
      "mean_token_accuracy": 0.8393548727035522,
      "num_tokens": 481310267.0,
      "step": 3020
    },
    {
      "epoch": 1.5366225839267549,
      "grad_norm": 1.0826072692871094,
      "learning_rate": 1e-05,
      "loss": 0.5009,
      "mean_token_accuracy": 0.8405871391296387,
      "num_tokens": 481466862.0,
      "step": 3021
    },
    {
      "epoch": 1.5371312309257377,
      "grad_norm": 0.9848809242248535,
      "learning_rate": 1e-05,
      "loss": 0.5347,
      "mean_token_accuracy": 0.8312636613845825,
      "num_tokens": 481623100.0,
      "step": 3022
    },
    {
      "epoch": 1.5376398779247202,
      "grad_norm": 0.9985467195510864,
      "learning_rate": 1e-05,
      "loss": 0.4824,
      "mean_token_accuracy": 0.8475914001464844,
      "num_tokens": 481786925.0,
      "step": 3023
    },
    {
      "epoch": 1.5381485249237028,
      "grad_norm": 1.0154087543487549,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.8535425662994385,
      "num_tokens": 481941546.0,
      "step": 3024
    },
    {
      "epoch": 1.5386571719226856,
      "grad_norm": 0.9510856866836548,
      "learning_rate": 1e-05,
      "loss": 0.5078,
      "mean_token_accuracy": 0.8398616313934326,
      "num_tokens": 482102028.0,
      "step": 3025
    },
    {
      "epoch": 1.5391658189216684,
      "grad_norm": 0.9930600523948669,
      "learning_rate": 1e-05,
      "loss": 0.5023,
      "mean_token_accuracy": 0.8423943519592285,
      "num_tokens": 482261954.0,
      "step": 3026
    },
    {
      "epoch": 1.5396744659206512,
      "grad_norm": 0.9215907454490662,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8503847122192383,
      "num_tokens": 482424450.0,
      "step": 3027
    },
    {
      "epoch": 1.5401831129196337,
      "grad_norm": 1.0794174671173096,
      "learning_rate": 1e-05,
      "loss": 0.4712,
      "mean_token_accuracy": 0.8509054183959961,
      "num_tokens": 482583861.0,
      "step": 3028
    },
    {
      "epoch": 1.5406917599186165,
      "grad_norm": 0.9406433701515198,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8541445732116699,
      "num_tokens": 482744473.0,
      "step": 3029
    },
    {
      "epoch": 1.541200406917599,
      "grad_norm": 0.896527111530304,
      "learning_rate": 1e-05,
      "loss": 0.4813,
      "mean_token_accuracy": 0.8471190929412842,
      "num_tokens": 482917251.0,
      "step": 3030
    },
    {
      "epoch": 1.5417090539165819,
      "grad_norm": 1.0127754211425781,
      "learning_rate": 1e-05,
      "loss": 0.4708,
      "mean_token_accuracy": 0.8493427634239197,
      "num_tokens": 483088314.0,
      "step": 3031
    },
    {
      "epoch": 1.5422177009155646,
      "grad_norm": 0.9132606387138367,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8510922789573669,
      "num_tokens": 483235136.0,
      "step": 3032
    },
    {
      "epoch": 1.5427263479145474,
      "grad_norm": 1.001651406288147,
      "learning_rate": 1e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.8525159358978271,
      "num_tokens": 483379352.0,
      "step": 3033
    },
    {
      "epoch": 1.54323499491353,
      "grad_norm": 0.9252734184265137,
      "learning_rate": 1e-05,
      "loss": 0.4719,
      "mean_token_accuracy": 0.8498355746269226,
      "num_tokens": 483551953.0,
      "step": 3034
    },
    {
      "epoch": 1.5437436419125126,
      "grad_norm": 0.9206449389457703,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.8521483540534973,
      "num_tokens": 483702864.0,
      "step": 3035
    },
    {
      "epoch": 1.5442522889114954,
      "grad_norm": 1.0111031532287598,
      "learning_rate": 1e-05,
      "loss": 0.4822,
      "mean_token_accuracy": 0.8495075702667236,
      "num_tokens": 483861059.0,
      "step": 3036
    },
    {
      "epoch": 1.5447609359104781,
      "grad_norm": 0.91339111328125,
      "learning_rate": 1e-05,
      "loss": 0.4671,
      "mean_token_accuracy": 0.8509300947189331,
      "num_tokens": 484016392.0,
      "step": 3037
    },
    {
      "epoch": 1.545269582909461,
      "grad_norm": 0.8808000087738037,
      "learning_rate": 1e-05,
      "loss": 0.4525,
      "mean_token_accuracy": 0.8548637628555298,
      "num_tokens": 484173960.0,
      "step": 3038
    },
    {
      "epoch": 1.5457782299084435,
      "grad_norm": 1.042173147201538,
      "learning_rate": 1e-05,
      "loss": 0.4829,
      "mean_token_accuracy": 0.8460304737091064,
      "num_tokens": 484317468.0,
      "step": 3039
    },
    {
      "epoch": 1.5462868769074263,
      "grad_norm": 0.9629754424095154,
      "learning_rate": 1e-05,
      "loss": 0.5107,
      "mean_token_accuracy": 0.8375293016433716,
      "num_tokens": 484486741.0,
      "step": 3040
    },
    {
      "epoch": 1.5467955239064088,
      "grad_norm": 0.9451351761817932,
      "learning_rate": 1e-05,
      "loss": 0.4739,
      "mean_token_accuracy": 0.8494521379470825,
      "num_tokens": 484652684.0,
      "step": 3041
    },
    {
      "epoch": 1.5473041709053916,
      "grad_norm": 0.8708462715148926,
      "learning_rate": 1e-05,
      "loss": 0.4878,
      "mean_token_accuracy": 0.8446558117866516,
      "num_tokens": 484829318.0,
      "step": 3042
    },
    {
      "epoch": 1.5478128179043744,
      "grad_norm": 1.038165807723999,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.848456084728241,
      "num_tokens": 484991769.0,
      "step": 3043
    },
    {
      "epoch": 1.5483214649033572,
      "grad_norm": 0.9187168478965759,
      "learning_rate": 1e-05,
      "loss": 0.4832,
      "mean_token_accuracy": 0.846601128578186,
      "num_tokens": 485146681.0,
      "step": 3044
    },
    {
      "epoch": 1.5488301119023398,
      "grad_norm": 1.0916402339935303,
      "learning_rate": 1e-05,
      "loss": 0.4771,
      "mean_token_accuracy": 0.8475044369697571,
      "num_tokens": 485304764.0,
      "step": 3045
    },
    {
      "epoch": 1.5493387589013223,
      "grad_norm": 0.9968273043632507,
      "learning_rate": 1e-05,
      "loss": 0.4709,
      "mean_token_accuracy": 0.8496679663658142,
      "num_tokens": 485461388.0,
      "step": 3046
    },
    {
      "epoch": 1.5498474059003051,
      "grad_norm": 1.01937735080719,
      "learning_rate": 1e-05,
      "loss": 0.476,
      "mean_token_accuracy": 0.8473126292228699,
      "num_tokens": 485612954.0,
      "step": 3047
    },
    {
      "epoch": 1.550356052899288,
      "grad_norm": 0.9482355713844299,
      "learning_rate": 1e-05,
      "loss": 0.4601,
      "mean_token_accuracy": 0.8525004386901855,
      "num_tokens": 485774911.0,
      "step": 3048
    },
    {
      "epoch": 1.5508646998982707,
      "grad_norm": 0.8762860298156738,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.850465714931488,
      "num_tokens": 485941736.0,
      "step": 3049
    },
    {
      "epoch": 1.5513733468972533,
      "grad_norm": 0.9830577969551086,
      "learning_rate": 1e-05,
      "loss": 0.5025,
      "mean_token_accuracy": 0.8413152098655701,
      "num_tokens": 486099339.0,
      "step": 3050
    },
    {
      "epoch": 1.551881993896236,
      "grad_norm": 0.9780283570289612,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8565871715545654,
      "num_tokens": 486253663.0,
      "step": 3051
    },
    {
      "epoch": 1.5523906408952186,
      "grad_norm": 0.9304409623146057,
      "learning_rate": 1e-05,
      "loss": 0.4832,
      "mean_token_accuracy": 0.8460574150085449,
      "num_tokens": 486412016.0,
      "step": 3052
    },
    {
      "epoch": 1.5528992878942014,
      "grad_norm": 1.0042747259140015,
      "learning_rate": 1e-05,
      "loss": 0.4613,
      "mean_token_accuracy": 0.8515927791595459,
      "num_tokens": 486559498.0,
      "step": 3053
    },
    {
      "epoch": 1.5534079348931842,
      "grad_norm": 0.9880870580673218,
      "learning_rate": 1e-05,
      "loss": 0.4478,
      "mean_token_accuracy": 0.8557184934616089,
      "num_tokens": 486710623.0,
      "step": 3054
    },
    {
      "epoch": 1.553916581892167,
      "grad_norm": 0.9632434844970703,
      "learning_rate": 1e-05,
      "loss": 0.5085,
      "mean_token_accuracy": 0.8388104438781738,
      "num_tokens": 486868761.0,
      "step": 3055
    },
    {
      "epoch": 1.5544252288911495,
      "grad_norm": 0.9700770974159241,
      "learning_rate": 1e-05,
      "loss": 0.4615,
      "mean_token_accuracy": 0.851329505443573,
      "num_tokens": 487025165.0,
      "step": 3056
    },
    {
      "epoch": 1.554933875890132,
      "grad_norm": 0.9787797331809998,
      "learning_rate": 1e-05,
      "loss": 0.465,
      "mean_token_accuracy": 0.8498197197914124,
      "num_tokens": 487184848.0,
      "step": 3057
    },
    {
      "epoch": 1.5554425228891149,
      "grad_norm": 1.0208553075790405,
      "learning_rate": 1e-05,
      "loss": 0.5141,
      "mean_token_accuracy": 0.8400205373764038,
      "num_tokens": 487349819.0,
      "step": 3058
    },
    {
      "epoch": 1.5559511698880977,
      "grad_norm": 1.0137851238250732,
      "learning_rate": 1e-05,
      "loss": 0.5014,
      "mean_token_accuracy": 0.8405393362045288,
      "num_tokens": 487509643.0,
      "step": 3059
    },
    {
      "epoch": 1.5564598168870805,
      "grad_norm": 0.9738591313362122,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8491359353065491,
      "num_tokens": 487662393.0,
      "step": 3060
    },
    {
      "epoch": 1.556968463886063,
      "grad_norm": 0.9042809009552002,
      "learning_rate": 1e-05,
      "loss": 0.501,
      "mean_token_accuracy": 0.8427039384841919,
      "num_tokens": 487826775.0,
      "step": 3061
    },
    {
      "epoch": 1.5574771108850458,
      "grad_norm": 1.1072074174880981,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.8482124209403992,
      "num_tokens": 487985726.0,
      "step": 3062
    },
    {
      "epoch": 1.5579857578840284,
      "grad_norm": 0.9234567880630493,
      "learning_rate": 1e-05,
      "loss": 0.4363,
      "mean_token_accuracy": 0.8603277802467346,
      "num_tokens": 488154603.0,
      "step": 3063
    },
    {
      "epoch": 1.5584944048830112,
      "grad_norm": 0.9251611828804016,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8493567109107971,
      "num_tokens": 488319919.0,
      "step": 3064
    },
    {
      "epoch": 1.559003051881994,
      "grad_norm": 0.9729562401771545,
      "learning_rate": 1e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8612502813339233,
      "num_tokens": 488472353.0,
      "step": 3065
    },
    {
      "epoch": 1.5595116988809767,
      "grad_norm": 0.9432415962219238,
      "learning_rate": 1e-05,
      "loss": 0.5015,
      "mean_token_accuracy": 0.8404766321182251,
      "num_tokens": 488630649.0,
      "step": 3066
    },
    {
      "epoch": 1.5600203458799593,
      "grad_norm": 0.9517757892608643,
      "learning_rate": 1e-05,
      "loss": 0.5058,
      "mean_token_accuracy": 0.8391520380973816,
      "num_tokens": 488786552.0,
      "step": 3067
    },
    {
      "epoch": 1.5605289928789419,
      "grad_norm": 0.9881300330162048,
      "learning_rate": 1e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8583213090896606,
      "num_tokens": 488932053.0,
      "step": 3068
    },
    {
      "epoch": 1.5610376398779247,
      "grad_norm": 1.0126222372055054,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8509069681167603,
      "num_tokens": 489084924.0,
      "step": 3069
    },
    {
      "epoch": 1.5615462868769074,
      "grad_norm": 0.9657802581787109,
      "learning_rate": 1e-05,
      "loss": 0.4812,
      "mean_token_accuracy": 0.8476794958114624,
      "num_tokens": 489250757.0,
      "step": 3070
    },
    {
      "epoch": 1.5620549338758902,
      "grad_norm": 0.9316444396972656,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.851240873336792,
      "num_tokens": 489413258.0,
      "step": 3071
    },
    {
      "epoch": 1.5625635808748728,
      "grad_norm": 1.0126928091049194,
      "learning_rate": 1e-05,
      "loss": 0.5092,
      "mean_token_accuracy": 0.8393445014953613,
      "num_tokens": 489567664.0,
      "step": 3072
    },
    {
      "epoch": 1.5630722278738556,
      "grad_norm": 0.9944961667060852,
      "learning_rate": 1e-05,
      "loss": 0.4904,
      "mean_token_accuracy": 0.8438583612442017,
      "num_tokens": 489723061.0,
      "step": 3073
    },
    {
      "epoch": 1.5635808748728381,
      "grad_norm": 0.9280596971511841,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.8593460321426392,
      "num_tokens": 489874451.0,
      "step": 3074
    },
    {
      "epoch": 1.564089521871821,
      "grad_norm": 1.079498291015625,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8525861501693726,
      "num_tokens": 490036501.0,
      "step": 3075
    },
    {
      "epoch": 1.5645981688708037,
      "grad_norm": 0.9194795489311218,
      "learning_rate": 1e-05,
      "loss": 0.4691,
      "mean_token_accuracy": 0.8513813018798828,
      "num_tokens": 490204909.0,
      "step": 3076
    },
    {
      "epoch": 1.5651068158697865,
      "grad_norm": 1.0592505931854248,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8467689752578735,
      "num_tokens": 490356463.0,
      "step": 3077
    },
    {
      "epoch": 1.565615462868769,
      "grad_norm": 0.9179127216339111,
      "learning_rate": 1e-05,
      "loss": 0.489,
      "mean_token_accuracy": 0.8445532321929932,
      "num_tokens": 490514531.0,
      "step": 3078
    },
    {
      "epoch": 1.5661241098677516,
      "grad_norm": 0.9455628991127014,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8507919311523438,
      "num_tokens": 490685759.0,
      "step": 3079
    },
    {
      "epoch": 1.5666327568667344,
      "grad_norm": 0.891564130783081,
      "learning_rate": 1e-05,
      "loss": 0.4763,
      "mean_token_accuracy": 0.8482155799865723,
      "num_tokens": 490845055.0,
      "step": 3080
    },
    {
      "epoch": 1.5671414038657172,
      "grad_norm": 0.9942660927772522,
      "learning_rate": 1e-05,
      "loss": 0.4928,
      "mean_token_accuracy": 0.8423429727554321,
      "num_tokens": 491001535.0,
      "step": 3081
    },
    {
      "epoch": 1.5676500508647,
      "grad_norm": 0.9702527523040771,
      "learning_rate": 1e-05,
      "loss": 0.5115,
      "mean_token_accuracy": 0.8396128416061401,
      "num_tokens": 491152661.0,
      "step": 3082
    },
    {
      "epoch": 1.5681586978636826,
      "grad_norm": 0.9330050349235535,
      "learning_rate": 1e-05,
      "loss": 0.4752,
      "mean_token_accuracy": 0.8495821356773376,
      "num_tokens": 491305291.0,
      "step": 3083
    },
    {
      "epoch": 1.5686673448626653,
      "grad_norm": 0.9987598061561584,
      "learning_rate": 1e-05,
      "loss": 0.4681,
      "mean_token_accuracy": 0.8500514030456543,
      "num_tokens": 491455713.0,
      "step": 3084
    },
    {
      "epoch": 1.569175991861648,
      "grad_norm": 0.9497177600860596,
      "learning_rate": 1e-05,
      "loss": 0.445,
      "mean_token_accuracy": 0.857146143913269,
      "num_tokens": 491607483.0,
      "step": 3085
    },
    {
      "epoch": 1.5696846388606307,
      "grad_norm": 0.9814060926437378,
      "learning_rate": 1e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.8496817350387573,
      "num_tokens": 491772168.0,
      "step": 3086
    },
    {
      "epoch": 1.5701932858596135,
      "grad_norm": 0.9190930724143982,
      "learning_rate": 1e-05,
      "loss": 0.4644,
      "mean_token_accuracy": 0.8526386618614197,
      "num_tokens": 491931085.0,
      "step": 3087
    },
    {
      "epoch": 1.5707019328585963,
      "grad_norm": 0.9814797639846802,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8529306650161743,
      "num_tokens": 492088615.0,
      "step": 3088
    },
    {
      "epoch": 1.5712105798575788,
      "grad_norm": 0.9221906065940857,
      "learning_rate": 1e-05,
      "loss": 0.4398,
      "mean_token_accuracy": 0.8584678769111633,
      "num_tokens": 492240833.0,
      "step": 3089
    },
    {
      "epoch": 1.5717192268565614,
      "grad_norm": 0.9911572337150574,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8544842004776001,
      "num_tokens": 492397349.0,
      "step": 3090
    },
    {
      "epoch": 1.5722278738555442,
      "grad_norm": 0.9610999226570129,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8520728945732117,
      "num_tokens": 492556148.0,
      "step": 3091
    },
    {
      "epoch": 1.572736520854527,
      "grad_norm": 0.9577346444129944,
      "learning_rate": 1e-05,
      "loss": 0.4795,
      "mean_token_accuracy": 0.8496525287628174,
      "num_tokens": 492713338.0,
      "step": 3092
    },
    {
      "epoch": 1.5732451678535098,
      "grad_norm": 0.9709999561309814,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.848212718963623,
      "num_tokens": 492861776.0,
      "step": 3093
    },
    {
      "epoch": 1.5737538148524923,
      "grad_norm": 0.9600496292114258,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.853322446346283,
      "num_tokens": 493019564.0,
      "step": 3094
    },
    {
      "epoch": 1.5742624618514751,
      "grad_norm": 0.8980912566184998,
      "learning_rate": 1e-05,
      "loss": 0.4642,
      "mean_token_accuracy": 0.8510512113571167,
      "num_tokens": 493172976.0,
      "step": 3095
    },
    {
      "epoch": 1.5747711088504577,
      "grad_norm": 0.896973729133606,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8487134575843811,
      "num_tokens": 493335981.0,
      "step": 3096
    },
    {
      "epoch": 1.5752797558494405,
      "grad_norm": 0.9582679867744446,
      "learning_rate": 1e-05,
      "loss": 0.4489,
      "mean_token_accuracy": 0.8563022613525391,
      "num_tokens": 493497636.0,
      "step": 3097
    },
    {
      "epoch": 1.5757884028484233,
      "grad_norm": 0.9179272055625916,
      "learning_rate": 1e-05,
      "loss": 0.4587,
      "mean_token_accuracy": 0.85282301902771,
      "num_tokens": 493652983.0,
      "step": 3098
    },
    {
      "epoch": 1.576297049847406,
      "grad_norm": 0.9353819489479065,
      "learning_rate": 1e-05,
      "loss": 0.4602,
      "mean_token_accuracy": 0.8537269234657288,
      "num_tokens": 493819335.0,
      "step": 3099
    },
    {
      "epoch": 1.5768056968463886,
      "grad_norm": 1.002805233001709,
      "learning_rate": 1e-05,
      "loss": 0.465,
      "mean_token_accuracy": 0.8521015644073486,
      "num_tokens": 493983683.0,
      "step": 3100
    },
    {
      "epoch": 1.5773143438453712,
      "grad_norm": 1.0276256799697876,
      "learning_rate": 1e-05,
      "loss": 0.4521,
      "mean_token_accuracy": 0.8568700551986694,
      "num_tokens": 494126765.0,
      "step": 3101
    },
    {
      "epoch": 1.577822990844354,
      "grad_norm": 1.151731252670288,
      "learning_rate": 1e-05,
      "loss": 0.5269,
      "mean_token_accuracy": 0.8336443901062012,
      "num_tokens": 494296649.0,
      "step": 3102
    },
    {
      "epoch": 1.5783316378433367,
      "grad_norm": 0.9901059865951538,
      "learning_rate": 1e-05,
      "loss": 0.487,
      "mean_token_accuracy": 0.8451051712036133,
      "num_tokens": 494456286.0,
      "step": 3103
    },
    {
      "epoch": 1.5788402848423195,
      "grad_norm": 0.9622817039489746,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.8511883020401001,
      "num_tokens": 494612800.0,
      "step": 3104
    },
    {
      "epoch": 1.579348931841302,
      "grad_norm": 0.9548298120498657,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8519192337989807,
      "num_tokens": 494766468.0,
      "step": 3105
    },
    {
      "epoch": 1.5798575788402849,
      "grad_norm": 0.979957103729248,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8517515659332275,
      "num_tokens": 494920941.0,
      "step": 3106
    },
    {
      "epoch": 1.5803662258392674,
      "grad_norm": 0.903709352016449,
      "learning_rate": 1e-05,
      "loss": 0.4501,
      "mean_token_accuracy": 0.8568116426467896,
      "num_tokens": 495068008.0,
      "step": 3107
    },
    {
      "epoch": 1.5808748728382502,
      "grad_norm": 0.8897936344146729,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8548612594604492,
      "num_tokens": 495234703.0,
      "step": 3108
    },
    {
      "epoch": 1.581383519837233,
      "grad_norm": 0.9432336091995239,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8530568480491638,
      "num_tokens": 495393286.0,
      "step": 3109
    },
    {
      "epoch": 1.5818921668362158,
      "grad_norm": 0.9855477809906006,
      "learning_rate": 1e-05,
      "loss": 0.4629,
      "mean_token_accuracy": 0.8511521220207214,
      "num_tokens": 495546408.0,
      "step": 3110
    },
    {
      "epoch": 1.5824008138351984,
      "grad_norm": 0.9574191570281982,
      "learning_rate": 1e-05,
      "loss": 0.4921,
      "mean_token_accuracy": 0.8433710336685181,
      "num_tokens": 495707039.0,
      "step": 3111
    },
    {
      "epoch": 1.582909460834181,
      "grad_norm": 1.2597713470458984,
      "learning_rate": 1e-05,
      "loss": 0.487,
      "mean_token_accuracy": 0.8451650142669678,
      "num_tokens": 495873689.0,
      "step": 3112
    },
    {
      "epoch": 1.5834181078331637,
      "grad_norm": 0.9632464051246643,
      "learning_rate": 1e-05,
      "loss": 0.4677,
      "mean_token_accuracy": 0.8511648178100586,
      "num_tokens": 496043972.0,
      "step": 3113
    },
    {
      "epoch": 1.5839267548321465,
      "grad_norm": 0.8824676871299744,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8504089713096619,
      "num_tokens": 496200993.0,
      "step": 3114
    },
    {
      "epoch": 1.5844354018311293,
      "grad_norm": 0.9932284355163574,
      "learning_rate": 1e-05,
      "loss": 0.4885,
      "mean_token_accuracy": 0.8445684313774109,
      "num_tokens": 496361750.0,
      "step": 3115
    },
    {
      "epoch": 1.5849440488301119,
      "grad_norm": 0.9311143159866333,
      "learning_rate": 1e-05,
      "loss": 0.454,
      "mean_token_accuracy": 0.8534303903579712,
      "num_tokens": 496528159.0,
      "step": 3116
    },
    {
      "epoch": 1.5854526958290946,
      "grad_norm": 0.9174624085426331,
      "learning_rate": 1e-05,
      "loss": 0.4881,
      "mean_token_accuracy": 0.845794141292572,
      "num_tokens": 496683804.0,
      "step": 3117
    },
    {
      "epoch": 1.5859613428280772,
      "grad_norm": 0.9836673736572266,
      "learning_rate": 1e-05,
      "loss": 0.4607,
      "mean_token_accuracy": 0.8526643514633179,
      "num_tokens": 496842977.0,
      "step": 3118
    },
    {
      "epoch": 1.58646998982706,
      "grad_norm": 0.9321286082267761,
      "learning_rate": 1e-05,
      "loss": 0.4901,
      "mean_token_accuracy": 0.8443841338157654,
      "num_tokens": 496996452.0,
      "step": 3119
    },
    {
      "epoch": 1.5869786368260428,
      "grad_norm": 0.9813646674156189,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8516721129417419,
      "num_tokens": 497141795.0,
      "step": 3120
    },
    {
      "epoch": 1.5874872838250256,
      "grad_norm": 0.9760952591896057,
      "learning_rate": 1e-05,
      "loss": 0.4914,
      "mean_token_accuracy": 0.8431864976882935,
      "num_tokens": 497294276.0,
      "step": 3121
    },
    {
      "epoch": 1.5879959308240081,
      "grad_norm": 0.9025095701217651,
      "learning_rate": 1e-05,
      "loss": 0.4497,
      "mean_token_accuracy": 0.8556056022644043,
      "num_tokens": 497453652.0,
      "step": 3122
    },
    {
      "epoch": 1.5885045778229907,
      "grad_norm": 0.8846112489700317,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8493298888206482,
      "num_tokens": 497617124.0,
      "step": 3123
    },
    {
      "epoch": 1.5890132248219735,
      "grad_norm": 0.9269668459892273,
      "learning_rate": 1e-05,
      "loss": 0.4815,
      "mean_token_accuracy": 0.8469797372817993,
      "num_tokens": 497784575.0,
      "step": 3124
    },
    {
      "epoch": 1.5895218718209563,
      "grad_norm": 0.9350006580352783,
      "learning_rate": 1e-05,
      "loss": 0.4618,
      "mean_token_accuracy": 0.8537191152572632,
      "num_tokens": 497936438.0,
      "step": 3125
    },
    {
      "epoch": 1.590030518819939,
      "grad_norm": 0.8660092949867249,
      "learning_rate": 1e-05,
      "loss": 0.4664,
      "mean_token_accuracy": 0.8507699966430664,
      "num_tokens": 498096276.0,
      "step": 3126
    },
    {
      "epoch": 1.5905391658189216,
      "grad_norm": 0.9839761853218079,
      "learning_rate": 1e-05,
      "loss": 0.4714,
      "mean_token_accuracy": 0.8495392799377441,
      "num_tokens": 498259727.0,
      "step": 3127
    },
    {
      "epoch": 1.5910478128179044,
      "grad_norm": 0.9081756472587585,
      "learning_rate": 1e-05,
      "loss": 0.4701,
      "mean_token_accuracy": 0.8499031066894531,
      "num_tokens": 498422502.0,
      "step": 3128
    },
    {
      "epoch": 1.591556459816887,
      "grad_norm": 0.9107279181480408,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8488689064979553,
      "num_tokens": 498583191.0,
      "step": 3129
    },
    {
      "epoch": 1.5920651068158698,
      "grad_norm": 1.0051671266555786,
      "learning_rate": 1e-05,
      "loss": 0.448,
      "mean_token_accuracy": 0.856336236000061,
      "num_tokens": 498743572.0,
      "step": 3130
    },
    {
      "epoch": 1.5925737538148526,
      "grad_norm": 1.0061246156692505,
      "learning_rate": 1e-05,
      "loss": 0.4886,
      "mean_token_accuracy": 0.8462451696395874,
      "num_tokens": 498892569.0,
      "step": 3131
    },
    {
      "epoch": 1.5930824008138353,
      "grad_norm": 0.925137460231781,
      "learning_rate": 1e-05,
      "loss": 0.4897,
      "mean_token_accuracy": 0.8452982902526855,
      "num_tokens": 499051577.0,
      "step": 3132
    },
    {
      "epoch": 1.593591047812818,
      "grad_norm": 1.0042551755905151,
      "learning_rate": 1e-05,
      "loss": 0.4804,
      "mean_token_accuracy": 0.8471189141273499,
      "num_tokens": 499199085.0,
      "step": 3133
    },
    {
      "epoch": 1.5940996948118005,
      "grad_norm": 0.9724690318107605,
      "learning_rate": 1e-05,
      "loss": 0.4791,
      "mean_token_accuracy": 0.8482787013053894,
      "num_tokens": 499360910.0,
      "step": 3134
    },
    {
      "epoch": 1.5946083418107833,
      "grad_norm": 0.8974447846412659,
      "learning_rate": 1e-05,
      "loss": 0.4493,
      "mean_token_accuracy": 0.8557643890380859,
      "num_tokens": 499520756.0,
      "step": 3135
    },
    {
      "epoch": 1.595116988809766,
      "grad_norm": 0.943656861782074,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8480054140090942,
      "num_tokens": 499675969.0,
      "step": 3136
    },
    {
      "epoch": 1.5956256358087488,
      "grad_norm": 0.9828064441680908,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.8466557264328003,
      "num_tokens": 499838062.0,
      "step": 3137
    },
    {
      "epoch": 1.5961342828077314,
      "grad_norm": 0.867965817451477,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8599332571029663,
      "num_tokens": 499988817.0,
      "step": 3138
    },
    {
      "epoch": 1.5966429298067142,
      "grad_norm": 0.9451082944869995,
      "learning_rate": 1e-05,
      "loss": 0.4682,
      "mean_token_accuracy": 0.8521145582199097,
      "num_tokens": 500156820.0,
      "step": 3139
    },
    {
      "epoch": 1.5971515768056967,
      "grad_norm": 0.9329448938369751,
      "learning_rate": 1e-05,
      "loss": 0.5034,
      "mean_token_accuracy": 0.8403822183609009,
      "num_tokens": 500312578.0,
      "step": 3140
    },
    {
      "epoch": 1.5976602238046795,
      "grad_norm": 0.9193881750106812,
      "learning_rate": 1e-05,
      "loss": 0.4543,
      "mean_token_accuracy": 0.8545393943786621,
      "num_tokens": 500465103.0,
      "step": 3141
    },
    {
      "epoch": 1.5981688708036623,
      "grad_norm": 0.9008822441101074,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8494077920913696,
      "num_tokens": 500631973.0,
      "step": 3142
    },
    {
      "epoch": 1.598677517802645,
      "grad_norm": 0.937802791595459,
      "learning_rate": 1e-05,
      "loss": 0.4552,
      "mean_token_accuracy": 0.8539186716079712,
      "num_tokens": 500791918.0,
      "step": 3143
    },
    {
      "epoch": 1.5991861648016277,
      "grad_norm": 2.2799768447875977,
      "learning_rate": 1e-05,
      "loss": 0.4847,
      "mean_token_accuracy": 0.8453787565231323,
      "num_tokens": 500955095.0,
      "step": 3144
    },
    {
      "epoch": 1.5996948118006102,
      "grad_norm": 1.041445016860962,
      "learning_rate": 1e-05,
      "loss": 0.4926,
      "mean_token_accuracy": 0.8438977599143982,
      "num_tokens": 501118364.0,
      "step": 3145
    },
    {
      "epoch": 1.600203458799593,
      "grad_norm": 0.9780532121658325,
      "learning_rate": 1e-05,
      "loss": 0.4739,
      "mean_token_accuracy": 0.849367618560791,
      "num_tokens": 501271088.0,
      "step": 3146
    },
    {
      "epoch": 1.6007121057985758,
      "grad_norm": 0.8955594301223755,
      "learning_rate": 1e-05,
      "loss": 0.4717,
      "mean_token_accuracy": 0.8512964248657227,
      "num_tokens": 501434010.0,
      "step": 3147
    },
    {
      "epoch": 1.6012207527975586,
      "grad_norm": 1.0110737085342407,
      "learning_rate": 1e-05,
      "loss": 0.453,
      "mean_token_accuracy": 0.8546295762062073,
      "num_tokens": 501573475.0,
      "step": 3148
    },
    {
      "epoch": 1.6017293997965412,
      "grad_norm": 1.03841233253479,
      "learning_rate": 1e-05,
      "loss": 0.4524,
      "mean_token_accuracy": 0.8547658920288086,
      "num_tokens": 501727861.0,
      "step": 3149
    },
    {
      "epoch": 1.602238046795524,
      "grad_norm": 0.9409387111663818,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8487837314605713,
      "num_tokens": 501887647.0,
      "step": 3150
    },
    {
      "epoch": 1.6027466937945065,
      "grad_norm": 1.0022470951080322,
      "learning_rate": 1e-05,
      "loss": 0.4441,
      "mean_token_accuracy": 0.8576607704162598,
      "num_tokens": 502035810.0,
      "step": 3151
    },
    {
      "epoch": 1.6032553407934893,
      "grad_norm": 0.9948771595954895,
      "learning_rate": 1e-05,
      "loss": 0.4761,
      "mean_token_accuracy": 0.8488028645515442,
      "num_tokens": 502192904.0,
      "step": 3152
    },
    {
      "epoch": 1.603763987792472,
      "grad_norm": 1.203999400138855,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8586881160736084,
      "num_tokens": 502349122.0,
      "step": 3153
    },
    {
      "epoch": 1.6042726347914549,
      "grad_norm": 0.9819480776786804,
      "learning_rate": 1e-05,
      "loss": 0.486,
      "mean_token_accuracy": 0.8465654850006104,
      "num_tokens": 502505274.0,
      "step": 3154
    },
    {
      "epoch": 1.6047812817904374,
      "grad_norm": 1.0000137090682983,
      "learning_rate": 1e-05,
      "loss": 0.519,
      "mean_token_accuracy": 0.8373558521270752,
      "num_tokens": 502661135.0,
      "step": 3155
    },
    {
      "epoch": 1.60528992878942,
      "grad_norm": 0.9279018044471741,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8506321310997009,
      "num_tokens": 502817024.0,
      "step": 3156
    },
    {
      "epoch": 1.6057985757884028,
      "grad_norm": 1.0391238927841187,
      "learning_rate": 1e-05,
      "loss": 0.4841,
      "mean_token_accuracy": 0.8479592800140381,
      "num_tokens": 502985732.0,
      "step": 3157
    },
    {
      "epoch": 1.6063072227873856,
      "grad_norm": 0.9349561929702759,
      "learning_rate": 1e-05,
      "loss": 0.5151,
      "mean_token_accuracy": 0.8382826447486877,
      "num_tokens": 503156030.0,
      "step": 3158
    },
    {
      "epoch": 1.6068158697863684,
      "grad_norm": 0.9930242896080017,
      "learning_rate": 1e-05,
      "loss": 0.4962,
      "mean_token_accuracy": 0.8423037528991699,
      "num_tokens": 503304262.0,
      "step": 3159
    },
    {
      "epoch": 1.607324516785351,
      "grad_norm": 0.8912495970726013,
      "learning_rate": 1e-05,
      "loss": 0.4212,
      "mean_token_accuracy": 0.8649030327796936,
      "num_tokens": 503459669.0,
      "step": 3160
    },
    {
      "epoch": 1.6078331637843337,
      "grad_norm": 0.9261059165000916,
      "learning_rate": 1e-05,
      "loss": 0.4589,
      "mean_token_accuracy": 0.8540016412734985,
      "num_tokens": 503610334.0,
      "step": 3161
    },
    {
      "epoch": 1.6083418107833163,
      "grad_norm": 0.9560903310775757,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8490592837333679,
      "num_tokens": 503766380.0,
      "step": 3162
    },
    {
      "epoch": 1.608850457782299,
      "grad_norm": 0.9605829119682312,
      "learning_rate": 1e-05,
      "loss": 0.4962,
      "mean_token_accuracy": 0.8422528505325317,
      "num_tokens": 503917967.0,
      "step": 3163
    },
    {
      "epoch": 1.6093591047812819,
      "grad_norm": 0.9508742690086365,
      "learning_rate": 1e-05,
      "loss": 0.4961,
      "mean_token_accuracy": 0.8426268100738525,
      "num_tokens": 504077350.0,
      "step": 3164
    },
    {
      "epoch": 1.6098677517802646,
      "grad_norm": 0.9653372764587402,
      "learning_rate": 1e-05,
      "loss": 0.4481,
      "mean_token_accuracy": 0.8559508323669434,
      "num_tokens": 504239254.0,
      "step": 3165
    },
    {
      "epoch": 1.6103763987792472,
      "grad_norm": 0.9955853223800659,
      "learning_rate": 1e-05,
      "loss": 0.4737,
      "mean_token_accuracy": 0.8484251499176025,
      "num_tokens": 504396848.0,
      "step": 3166
    },
    {
      "epoch": 1.6108850457782298,
      "grad_norm": 0.9986701011657715,
      "learning_rate": 1e-05,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8472124338150024,
      "num_tokens": 504548510.0,
      "step": 3167
    },
    {
      "epoch": 1.6113936927772126,
      "grad_norm": 1.056422233581543,
      "learning_rate": 1e-05,
      "loss": 0.4313,
      "mean_token_accuracy": 0.861392617225647,
      "num_tokens": 504709088.0,
      "step": 3168
    },
    {
      "epoch": 1.6119023397761953,
      "grad_norm": 0.9725289940834045,
      "learning_rate": 1e-05,
      "loss": 0.4898,
      "mean_token_accuracy": 0.844748318195343,
      "num_tokens": 504861511.0,
      "step": 3169
    },
    {
      "epoch": 1.6124109867751781,
      "grad_norm": 0.889563262462616,
      "learning_rate": 1e-05,
      "loss": 0.4743,
      "mean_token_accuracy": 0.8478610515594482,
      "num_tokens": 505040673.0,
      "step": 3170
    },
    {
      "epoch": 1.6129196337741607,
      "grad_norm": 1.020880937576294,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "mean_token_accuracy": 0.8515779972076416,
      "num_tokens": 505192544.0,
      "step": 3171
    },
    {
      "epoch": 1.6134282807731435,
      "grad_norm": 0.9151832461357117,
      "learning_rate": 1e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.8499784469604492,
      "num_tokens": 505369618.0,
      "step": 3172
    },
    {
      "epoch": 1.613936927772126,
      "grad_norm": 1.049196720123291,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.856555163860321,
      "num_tokens": 505522912.0,
      "step": 3173
    },
    {
      "epoch": 1.6144455747711088,
      "grad_norm": 0.9513052701950073,
      "learning_rate": 1e-05,
      "loss": 0.4705,
      "mean_token_accuracy": 0.8521981239318848,
      "num_tokens": 505672795.0,
      "step": 3174
    },
    {
      "epoch": 1.6149542217700916,
      "grad_norm": 1.0559369325637817,
      "learning_rate": 1e-05,
      "loss": 0.4628,
      "mean_token_accuracy": 0.853662371635437,
      "num_tokens": 505848647.0,
      "step": 3175
    },
    {
      "epoch": 1.6154628687690744,
      "grad_norm": 1.0654804706573486,
      "learning_rate": 1e-05,
      "loss": 0.4785,
      "mean_token_accuracy": 0.8477578163146973,
      "num_tokens": 506006904.0,
      "step": 3176
    },
    {
      "epoch": 1.615971515768057,
      "grad_norm": 1.124934196472168,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.8563681840896606,
      "num_tokens": 506170327.0,
      "step": 3177
    },
    {
      "epoch": 1.6164801627670395,
      "grad_norm": 1.0646849870681763,
      "learning_rate": 1e-05,
      "loss": 0.4437,
      "mean_token_accuracy": 0.8576815128326416,
      "num_tokens": 506328403.0,
      "step": 3178
    },
    {
      "epoch": 1.6169888097660223,
      "grad_norm": 1.1809217929840088,
      "learning_rate": 1e-05,
      "loss": 0.4721,
      "mean_token_accuracy": 0.8506449460983276,
      "num_tokens": 506476733.0,
      "step": 3179
    },
    {
      "epoch": 1.6174974567650051,
      "grad_norm": 1.1678529977798462,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8535448312759399,
      "num_tokens": 506646322.0,
      "step": 3180
    },
    {
      "epoch": 1.618006103763988,
      "grad_norm": 0.9654884934425354,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8569132089614868,
      "num_tokens": 506800989.0,
      "step": 3181
    },
    {
      "epoch": 1.6185147507629705,
      "grad_norm": 1.0711699724197388,
      "learning_rate": 1e-05,
      "loss": 0.4402,
      "mean_token_accuracy": 0.8574779033660889,
      "num_tokens": 506963590.0,
      "step": 3182
    },
    {
      "epoch": 1.6190233977619533,
      "grad_norm": 0.9858824610710144,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8534616231918335,
      "num_tokens": 507110698.0,
      "step": 3183
    },
    {
      "epoch": 1.6195320447609358,
      "grad_norm": 1.0069845914840698,
      "learning_rate": 1e-05,
      "loss": 0.4589,
      "mean_token_accuracy": 0.8537305593490601,
      "num_tokens": 507261633.0,
      "step": 3184
    },
    {
      "epoch": 1.6200406917599186,
      "grad_norm": 1.1432651281356812,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "mean_token_accuracy": 0.856499195098877,
      "num_tokens": 507422355.0,
      "step": 3185
    },
    {
      "epoch": 1.6205493387589014,
      "grad_norm": 0.9364808201789856,
      "learning_rate": 1e-05,
      "loss": 0.4864,
      "mean_token_accuracy": 0.8467901945114136,
      "num_tokens": 507586130.0,
      "step": 3186
    },
    {
      "epoch": 1.6210579857578842,
      "grad_norm": 1.058851957321167,
      "learning_rate": 1e-05,
      "loss": 0.4884,
      "mean_token_accuracy": 0.8438219428062439,
      "num_tokens": 507730780.0,
      "step": 3187
    },
    {
      "epoch": 1.6215666327568667,
      "grad_norm": 1.098166823387146,
      "learning_rate": 1e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.8547759056091309,
      "num_tokens": 507895824.0,
      "step": 3188
    },
    {
      "epoch": 1.6220752797558493,
      "grad_norm": 1.0428221225738525,
      "learning_rate": 1e-05,
      "loss": 0.4798,
      "mean_token_accuracy": 0.8461632132530212,
      "num_tokens": 508059047.0,
      "step": 3189
    },
    {
      "epoch": 1.622583926754832,
      "grad_norm": 1.0780528783798218,
      "learning_rate": 1e-05,
      "loss": 0.4984,
      "mean_token_accuracy": 0.8422732353210449,
      "num_tokens": 508213579.0,
      "step": 3190
    },
    {
      "epoch": 1.6230925737538149,
      "grad_norm": 0.9325408935546875,
      "learning_rate": 1e-05,
      "loss": 0.4898,
      "mean_token_accuracy": 0.8447512984275818,
      "num_tokens": 508362321.0,
      "step": 3191
    },
    {
      "epoch": 1.6236012207527977,
      "grad_norm": 1.1149005889892578,
      "learning_rate": 1e-05,
      "loss": 0.4461,
      "mean_token_accuracy": 0.8572701215744019,
      "num_tokens": 508527614.0,
      "step": 3192
    },
    {
      "epoch": 1.6241098677517802,
      "grad_norm": 1.0653656721115112,
      "learning_rate": 1e-05,
      "loss": 0.4562,
      "mean_token_accuracy": 0.8521193265914917,
      "num_tokens": 508675165.0,
      "step": 3193
    },
    {
      "epoch": 1.624618514750763,
      "grad_norm": 1.1204813718795776,
      "learning_rate": 1e-05,
      "loss": 0.491,
      "mean_token_accuracy": 0.8450343012809753,
      "num_tokens": 508844166.0,
      "step": 3194
    },
    {
      "epoch": 1.6251271617497456,
      "grad_norm": 1.0192406177520752,
      "learning_rate": 1e-05,
      "loss": 0.4668,
      "mean_token_accuracy": 0.8499091267585754,
      "num_tokens": 508999496.0,
      "step": 3195
    },
    {
      "epoch": 1.6256358087487284,
      "grad_norm": 1.937743067741394,
      "learning_rate": 1e-05,
      "loss": 0.4516,
      "mean_token_accuracy": 0.854866623878479,
      "num_tokens": 509153388.0,
      "step": 3196
    },
    {
      "epoch": 1.6261444557477112,
      "grad_norm": 1.1351182460784912,
      "learning_rate": 1e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.8531141877174377,
      "num_tokens": 509320466.0,
      "step": 3197
    },
    {
      "epoch": 1.626653102746694,
      "grad_norm": 0.9390095472335815,
      "learning_rate": 1e-05,
      "loss": 0.4827,
      "mean_token_accuracy": 0.8455871343612671,
      "num_tokens": 509478651.0,
      "step": 3198
    },
    {
      "epoch": 1.6271617497456765,
      "grad_norm": 0.9299654960632324,
      "learning_rate": 1e-05,
      "loss": 0.4765,
      "mean_token_accuracy": 0.8481392860412598,
      "num_tokens": 509632225.0,
      "step": 3199
    },
    {
      "epoch": 1.627670396744659,
      "grad_norm": 1.0026236772537231,
      "learning_rate": 1e-05,
      "loss": 0.4672,
      "mean_token_accuracy": 0.8508972525596619,
      "num_tokens": 509789533.0,
      "step": 3200
    },
    {
      "epoch": 1.6281790437436419,
      "grad_norm": 0.8967097401618958,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.849647045135498,
      "num_tokens": 509956113.0,
      "step": 3201
    },
    {
      "epoch": 1.6286876907426246,
      "grad_norm": 1.0988428592681885,
      "learning_rate": 1e-05,
      "loss": 0.458,
      "mean_token_accuracy": 0.853933572769165,
      "num_tokens": 510127406.0,
      "step": 3202
    },
    {
      "epoch": 1.6291963377416074,
      "grad_norm": 0.9219009876251221,
      "learning_rate": 1e-05,
      "loss": 0.4392,
      "mean_token_accuracy": 0.8581788539886475,
      "num_tokens": 510279425.0,
      "step": 3203
    },
    {
      "epoch": 1.62970498474059,
      "grad_norm": 1.0581265687942505,
      "learning_rate": 1e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.852953314781189,
      "num_tokens": 510431387.0,
      "step": 3204
    },
    {
      "epoch": 1.6302136317395728,
      "grad_norm": 0.9527878761291504,
      "learning_rate": 1e-05,
      "loss": 0.4769,
      "mean_token_accuracy": 0.8487293720245361,
      "num_tokens": 510592010.0,
      "step": 3205
    },
    {
      "epoch": 1.6307222787385554,
      "grad_norm": 0.8919398188591003,
      "learning_rate": 1e-05,
      "loss": 0.496,
      "mean_token_accuracy": 0.8447635769844055,
      "num_tokens": 510762803.0,
      "step": 3206
    },
    {
      "epoch": 1.6312309257375381,
      "grad_norm": 0.9669695496559143,
      "learning_rate": 1e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.8502144813537598,
      "num_tokens": 510931494.0,
      "step": 3207
    },
    {
      "epoch": 1.631739572736521,
      "grad_norm": 1.350732684135437,
      "learning_rate": 1e-05,
      "loss": 0.512,
      "mean_token_accuracy": 0.8376517295837402,
      "num_tokens": 511087806.0,
      "step": 3208
    },
    {
      "epoch": 1.6322482197355037,
      "grad_norm": 1.014112949371338,
      "learning_rate": 1e-05,
      "loss": 0.4787,
      "mean_token_accuracy": 0.8486483693122864,
      "num_tokens": 511250013.0,
      "step": 3209
    },
    {
      "epoch": 1.6327568667344863,
      "grad_norm": 0.9049482941627502,
      "learning_rate": 1e-05,
      "loss": 0.4982,
      "mean_token_accuracy": 0.8427013158798218,
      "num_tokens": 511413057.0,
      "step": 3210
    },
    {
      "epoch": 1.6332655137334688,
      "grad_norm": 0.9740550518035889,
      "learning_rate": 1e-05,
      "loss": 0.45,
      "mean_token_accuracy": 0.8555504679679871,
      "num_tokens": 511569137.0,
      "step": 3211
    },
    {
      "epoch": 1.6337741607324516,
      "grad_norm": 0.9177180528640747,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8619349002838135,
      "num_tokens": 511722415.0,
      "step": 3212
    },
    {
      "epoch": 1.6342828077314344,
      "grad_norm": 0.8687677383422852,
      "learning_rate": 1e-05,
      "loss": 0.4424,
      "mean_token_accuracy": 0.8564339876174927,
      "num_tokens": 511879049.0,
      "step": 3213
    },
    {
      "epoch": 1.6347914547304172,
      "grad_norm": 1.2000905275344849,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8552175760269165,
      "num_tokens": 512040496.0,
      "step": 3214
    },
    {
      "epoch": 1.6353001017293998,
      "grad_norm": 0.9453958868980408,
      "learning_rate": 1e-05,
      "loss": 0.4675,
      "mean_token_accuracy": 0.8500424027442932,
      "num_tokens": 512186456.0,
      "step": 3215
    },
    {
      "epoch": 1.6358087487283826,
      "grad_norm": 1.0896542072296143,
      "learning_rate": 1e-05,
      "loss": 0.4889,
      "mean_token_accuracy": 0.8457797765731812,
      "num_tokens": 512348538.0,
      "step": 3216
    },
    {
      "epoch": 1.6363173957273651,
      "grad_norm": 1.0226482152938843,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8481237292289734,
      "num_tokens": 512510611.0,
      "step": 3217
    },
    {
      "epoch": 1.636826042726348,
      "grad_norm": 0.9434159994125366,
      "learning_rate": 1e-05,
      "loss": 0.4704,
      "mean_token_accuracy": 0.8490217924118042,
      "num_tokens": 512672571.0,
      "step": 3218
    },
    {
      "epoch": 1.6373346897253307,
      "grad_norm": 0.9946489334106445,
      "learning_rate": 1e-05,
      "loss": 0.4948,
      "mean_token_accuracy": 0.8430628776550293,
      "num_tokens": 512827563.0,
      "step": 3219
    },
    {
      "epoch": 1.6378433367243135,
      "grad_norm": 0.9760912656784058,
      "learning_rate": 1e-05,
      "loss": 0.4492,
      "mean_token_accuracy": 0.8558773994445801,
      "num_tokens": 512983469.0,
      "step": 3220
    },
    {
      "epoch": 1.638351983723296,
      "grad_norm": 0.9320849180221558,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.8541078567504883,
      "num_tokens": 513142429.0,
      "step": 3221
    },
    {
      "epoch": 1.6388606307222786,
      "grad_norm": 0.915862500667572,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "mean_token_accuracy": 0.8523277044296265,
      "num_tokens": 513304550.0,
      "step": 3222
    },
    {
      "epoch": 1.6393692777212614,
      "grad_norm": 0.8940544128417969,
      "learning_rate": 1e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8485328555107117,
      "num_tokens": 513477949.0,
      "step": 3223
    },
    {
      "epoch": 1.6398779247202442,
      "grad_norm": 0.927986204624176,
      "learning_rate": 1e-05,
      "loss": 0.4848,
      "mean_token_accuracy": 0.8451883792877197,
      "num_tokens": 513639647.0,
      "step": 3224
    },
    {
      "epoch": 1.640386571719227,
      "grad_norm": 0.8704172372817993,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8500398397445679,
      "num_tokens": 513797781.0,
      "step": 3225
    },
    {
      "epoch": 1.6408952187182095,
      "grad_norm": 0.9353987574577332,
      "learning_rate": 1e-05,
      "loss": 0.471,
      "mean_token_accuracy": 0.8504478931427002,
      "num_tokens": 513955189.0,
      "step": 3226
    },
    {
      "epoch": 1.6414038657171923,
      "grad_norm": 0.9797881841659546,
      "learning_rate": 1e-05,
      "loss": 0.4615,
      "mean_token_accuracy": 0.8517624139785767,
      "num_tokens": 514110474.0,
      "step": 3227
    },
    {
      "epoch": 1.6419125127161749,
      "grad_norm": 0.9477497339248657,
      "learning_rate": 1e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.850496768951416,
      "num_tokens": 514262184.0,
      "step": 3228
    },
    {
      "epoch": 1.6424211597151577,
      "grad_norm": 0.8949544429779053,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8600969314575195,
      "num_tokens": 514417324.0,
      "step": 3229
    },
    {
      "epoch": 1.6429298067141405,
      "grad_norm": 0.9569036364555359,
      "learning_rate": 1e-05,
      "loss": 0.4761,
      "mean_token_accuracy": 0.8493359088897705,
      "num_tokens": 514577426.0,
      "step": 3230
    },
    {
      "epoch": 1.6434384537131232,
      "grad_norm": 0.9404508471488953,
      "learning_rate": 1e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.8487488031387329,
      "num_tokens": 514715226.0,
      "step": 3231
    },
    {
      "epoch": 1.6439471007121058,
      "grad_norm": 0.8887859582901001,
      "learning_rate": 1e-05,
      "loss": 0.46,
      "mean_token_accuracy": 0.8533378839492798,
      "num_tokens": 514882566.0,
      "step": 3232
    },
    {
      "epoch": 1.6444557477110884,
      "grad_norm": 0.9008356332778931,
      "learning_rate": 1e-05,
      "loss": 0.4308,
      "mean_token_accuracy": 0.8623565435409546,
      "num_tokens": 515038542.0,
      "step": 3233
    },
    {
      "epoch": 1.6449643947100712,
      "grad_norm": 0.9309217929840088,
      "learning_rate": 1e-05,
      "loss": 0.4989,
      "mean_token_accuracy": 0.8433068990707397,
      "num_tokens": 515212612.0,
      "step": 3234
    },
    {
      "epoch": 1.645473041709054,
      "grad_norm": 0.8963108658790588,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "mean_token_accuracy": 0.854758083820343,
      "num_tokens": 515364485.0,
      "step": 3235
    },
    {
      "epoch": 1.6459816887080367,
      "grad_norm": 0.9246181845664978,
      "learning_rate": 1e-05,
      "loss": 0.498,
      "mean_token_accuracy": 0.8418968915939331,
      "num_tokens": 515525982.0,
      "step": 3236
    },
    {
      "epoch": 1.6464903357070193,
      "grad_norm": 0.8769820332527161,
      "learning_rate": 1e-05,
      "loss": 0.4497,
      "mean_token_accuracy": 0.8552553057670593,
      "num_tokens": 515695826.0,
      "step": 3237
    },
    {
      "epoch": 1.6469989827060019,
      "grad_norm": 0.9627645611763,
      "learning_rate": 1e-05,
      "loss": 0.4615,
      "mean_token_accuracy": 0.8528941869735718,
      "num_tokens": 515853344.0,
      "step": 3238
    },
    {
      "epoch": 1.6475076297049847,
      "grad_norm": 1.0333333015441895,
      "learning_rate": 1e-05,
      "loss": 0.4764,
      "mean_token_accuracy": 0.8480497598648071,
      "num_tokens": 516010142.0,
      "step": 3239
    },
    {
      "epoch": 1.6480162767039674,
      "grad_norm": 0.9454230666160583,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8532935976982117,
      "num_tokens": 516159473.0,
      "step": 3240
    },
    {
      "epoch": 1.6485249237029502,
      "grad_norm": 0.8893861174583435,
      "learning_rate": 1e-05,
      "loss": 0.4733,
      "mean_token_accuracy": 0.8496898412704468,
      "num_tokens": 516332297.0,
      "step": 3241
    },
    {
      "epoch": 1.649033570701933,
      "grad_norm": 0.9125193357467651,
      "learning_rate": 1e-05,
      "loss": 0.4981,
      "mean_token_accuracy": 0.8416143655776978,
      "num_tokens": 516492142.0,
      "step": 3242
    },
    {
      "epoch": 1.6495422177009156,
      "grad_norm": 0.9185030460357666,
      "learning_rate": 1e-05,
      "loss": 0.4638,
      "mean_token_accuracy": 0.852362871170044,
      "num_tokens": 516646489.0,
      "step": 3243
    },
    {
      "epoch": 1.6500508646998981,
      "grad_norm": 0.9418681859970093,
      "learning_rate": 1e-05,
      "loss": 0.466,
      "mean_token_accuracy": 0.8507981300354004,
      "num_tokens": 516798136.0,
      "step": 3244
    },
    {
      "epoch": 1.650559511698881,
      "grad_norm": 0.980635941028595,
      "learning_rate": 1e-05,
      "loss": 0.4736,
      "mean_token_accuracy": 0.8493054509162903,
      "num_tokens": 516964061.0,
      "step": 3245
    },
    {
      "epoch": 1.6510681586978637,
      "grad_norm": 1.006500482559204,
      "learning_rate": 1e-05,
      "loss": 0.4793,
      "mean_token_accuracy": 0.8486887216567993,
      "num_tokens": 517108993.0,
      "step": 3246
    },
    {
      "epoch": 1.6515768056968465,
      "grad_norm": 0.9133172631263733,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8549003005027771,
      "num_tokens": 517276121.0,
      "step": 3247
    },
    {
      "epoch": 1.652085452695829,
      "grad_norm": 0.943731427192688,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8488215208053589,
      "num_tokens": 517439575.0,
      "step": 3248
    },
    {
      "epoch": 1.6525940996948116,
      "grad_norm": 1.0023247003555298,
      "learning_rate": 1e-05,
      "loss": 0.4772,
      "mean_token_accuracy": 0.8474256992340088,
      "num_tokens": 517594951.0,
      "step": 3249
    },
    {
      "epoch": 1.6531027466937944,
      "grad_norm": 0.919856607913971,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8480158448219299,
      "num_tokens": 517744319.0,
      "step": 3250
    },
    {
      "epoch": 1.6536113936927772,
      "grad_norm": 0.993774950504303,
      "learning_rate": 1e-05,
      "loss": 0.4821,
      "mean_token_accuracy": 0.8462665677070618,
      "num_tokens": 517908794.0,
      "step": 3251
    },
    {
      "epoch": 1.65412004069176,
      "grad_norm": 1.0048195123672485,
      "learning_rate": 1e-05,
      "loss": 0.4826,
      "mean_token_accuracy": 0.8465999364852905,
      "num_tokens": 518071406.0,
      "step": 3252
    },
    {
      "epoch": 1.6546286876907428,
      "grad_norm": 1.0032241344451904,
      "learning_rate": 1e-05,
      "loss": 0.488,
      "mean_token_accuracy": 0.8444687724113464,
      "num_tokens": 518229568.0,
      "step": 3253
    },
    {
      "epoch": 1.6551373346897253,
      "grad_norm": 0.9415081739425659,
      "learning_rate": 1e-05,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8483941555023193,
      "num_tokens": 518389903.0,
      "step": 3254
    },
    {
      "epoch": 1.655645981688708,
      "grad_norm": 1.036271333694458,
      "learning_rate": 1e-05,
      "loss": 0.4758,
      "mean_token_accuracy": 0.84796142578125,
      "num_tokens": 518538304.0,
      "step": 3255
    },
    {
      "epoch": 1.6561546286876907,
      "grad_norm": 1.052514910697937,
      "learning_rate": 1e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8535677194595337,
      "num_tokens": 518689792.0,
      "step": 3256
    },
    {
      "epoch": 1.6566632756866735,
      "grad_norm": 1.0288665294647217,
      "learning_rate": 1e-05,
      "loss": 0.4699,
      "mean_token_accuracy": 0.8506037592887878,
      "num_tokens": 518829689.0,
      "step": 3257
    },
    {
      "epoch": 1.6571719226856563,
      "grad_norm": 1.1159483194351196,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8576152324676514,
      "num_tokens": 518982145.0,
      "step": 3258
    },
    {
      "epoch": 1.6576805696846388,
      "grad_norm": 0.9410718083381653,
      "learning_rate": 1e-05,
      "loss": 0.483,
      "mean_token_accuracy": 0.846442461013794,
      "num_tokens": 519145357.0,
      "step": 3259
    },
    {
      "epoch": 1.6581892166836214,
      "grad_norm": 1.0251357555389404,
      "learning_rate": 1e-05,
      "loss": 0.5004,
      "mean_token_accuracy": 0.8411550521850586,
      "num_tokens": 519307621.0,
      "step": 3260
    },
    {
      "epoch": 1.6586978636826042,
      "grad_norm": 0.9591336846351624,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8503050804138184,
      "num_tokens": 519461973.0,
      "step": 3261
    },
    {
      "epoch": 1.659206510681587,
      "grad_norm": 0.9231889247894287,
      "learning_rate": 1e-05,
      "loss": 0.4767,
      "mean_token_accuracy": 0.8487023115158081,
      "num_tokens": 519626545.0,
      "step": 3262
    },
    {
      "epoch": 1.6597151576805698,
      "grad_norm": 1.0337306261062622,
      "learning_rate": 1e-05,
      "loss": 0.4793,
      "mean_token_accuracy": 0.8472877144813538,
      "num_tokens": 519790645.0,
      "step": 3263
    },
    {
      "epoch": 1.6602238046795526,
      "grad_norm": 1.428776741027832,
      "learning_rate": 1e-05,
      "loss": 0.4682,
      "mean_token_accuracy": 0.8494995832443237,
      "num_tokens": 519945118.0,
      "step": 3264
    },
    {
      "epoch": 1.6607324516785351,
      "grad_norm": 0.9849370718002319,
      "learning_rate": 1e-05,
      "loss": 0.493,
      "mean_token_accuracy": 0.8436269760131836,
      "num_tokens": 520102088.0,
      "step": 3265
    },
    {
      "epoch": 1.6612410986775177,
      "grad_norm": 1.0113447904586792,
      "learning_rate": 1e-05,
      "loss": 0.5157,
      "mean_token_accuracy": 0.8369056582450867,
      "num_tokens": 520263275.0,
      "step": 3266
    },
    {
      "epoch": 1.6617497456765005,
      "grad_norm": 0.8960872292518616,
      "learning_rate": 1e-05,
      "loss": 0.4738,
      "mean_token_accuracy": 0.8496039509773254,
      "num_tokens": 520427222.0,
      "step": 3267
    },
    {
      "epoch": 1.6622583926754833,
      "grad_norm": 1.0057883262634277,
      "learning_rate": 1e-05,
      "loss": 0.4931,
      "mean_token_accuracy": 0.8433455228805542,
      "num_tokens": 520587306.0,
      "step": 3268
    },
    {
      "epoch": 1.662767039674466,
      "grad_norm": 0.919667661190033,
      "learning_rate": 1e-05,
      "loss": 0.4534,
      "mean_token_accuracy": 0.8553017377853394,
      "num_tokens": 520752920.0,
      "step": 3269
    },
    {
      "epoch": 1.6632756866734486,
      "grad_norm": 0.9011039733886719,
      "learning_rate": 1e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8484370708465576,
      "num_tokens": 520922639.0,
      "step": 3270
    },
    {
      "epoch": 1.6637843336724312,
      "grad_norm": 0.9278181791305542,
      "learning_rate": 1e-05,
      "loss": 0.4655,
      "mean_token_accuracy": 0.8510851263999939,
      "num_tokens": 521077842.0,
      "step": 3271
    },
    {
      "epoch": 1.664292980671414,
      "grad_norm": 0.9309078454971313,
      "learning_rate": 1e-05,
      "loss": 0.5157,
      "mean_token_accuracy": 0.8380452394485474,
      "num_tokens": 521246669.0,
      "step": 3272
    },
    {
      "epoch": 1.6648016276703967,
      "grad_norm": 0.9402255415916443,
      "learning_rate": 1e-05,
      "loss": 0.4442,
      "mean_token_accuracy": 0.8570981621742249,
      "num_tokens": 521403766.0,
      "step": 3273
    },
    {
      "epoch": 1.6653102746693795,
      "grad_norm": 0.9927259683609009,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8527973890304565,
      "num_tokens": 521552038.0,
      "step": 3274
    },
    {
      "epoch": 1.6658189216683623,
      "grad_norm": 0.9242224097251892,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8570798635482788,
      "num_tokens": 521715941.0,
      "step": 3275
    },
    {
      "epoch": 1.6663275686673449,
      "grad_norm": 1.0413960218429565,
      "learning_rate": 1e-05,
      "loss": 0.5049,
      "mean_token_accuracy": 0.8396327495574951,
      "num_tokens": 521871204.0,
      "step": 3276
    },
    {
      "epoch": 1.6668362156663274,
      "grad_norm": 1.067278265953064,
      "learning_rate": 1e-05,
      "loss": 0.5112,
      "mean_token_accuracy": 0.839201807975769,
      "num_tokens": 522031714.0,
      "step": 3277
    },
    {
      "epoch": 1.6673448626653102,
      "grad_norm": 0.8974424600601196,
      "learning_rate": 1e-05,
      "loss": 0.4461,
      "mean_token_accuracy": 0.8575136661529541,
      "num_tokens": 522180569.0,
      "step": 3278
    },
    {
      "epoch": 1.667853509664293,
      "grad_norm": 1.0562061071395874,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.8524356484413147,
      "num_tokens": 522337674.0,
      "step": 3279
    },
    {
      "epoch": 1.6683621566632758,
      "grad_norm": 1.050594687461853,
      "learning_rate": 1e-05,
      "loss": 0.4799,
      "mean_token_accuracy": 0.8469610214233398,
      "num_tokens": 522491069.0,
      "step": 3280
    },
    {
      "epoch": 1.6688708036622584,
      "grad_norm": 0.9911378026008606,
      "learning_rate": 1e-05,
      "loss": 0.4613,
      "mean_token_accuracy": 0.8525702357292175,
      "num_tokens": 522650079.0,
      "step": 3281
    },
    {
      "epoch": 1.669379450661241,
      "grad_norm": 0.9349995255470276,
      "learning_rate": 1e-05,
      "loss": 0.4731,
      "mean_token_accuracy": 0.85033118724823,
      "num_tokens": 522813823.0,
      "step": 3282
    },
    {
      "epoch": 1.6698880976602237,
      "grad_norm": 0.9397663474082947,
      "learning_rate": 1e-05,
      "loss": 0.5216,
      "mean_token_accuracy": 0.8359174132347107,
      "num_tokens": 522976265.0,
      "step": 3283
    },
    {
      "epoch": 1.6703967446592065,
      "grad_norm": 1.0165876150131226,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8488295078277588,
      "num_tokens": 523142670.0,
      "step": 3284
    },
    {
      "epoch": 1.6709053916581893,
      "grad_norm": 0.9085330367088318,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.851130485534668,
      "num_tokens": 523300688.0,
      "step": 3285
    },
    {
      "epoch": 1.671414038657172,
      "grad_norm": 0.9892270565032959,
      "learning_rate": 1e-05,
      "loss": 0.5006,
      "mean_token_accuracy": 0.8418188095092773,
      "num_tokens": 523461987.0,
      "step": 3286
    },
    {
      "epoch": 1.6719226856561547,
      "grad_norm": 1.0040111541748047,
      "learning_rate": 1e-05,
      "loss": 0.4892,
      "mean_token_accuracy": 0.8444796800613403,
      "num_tokens": 523623517.0,
      "step": 3287
    },
    {
      "epoch": 1.6724313326551372,
      "grad_norm": 0.8981871604919434,
      "learning_rate": 1e-05,
      "loss": 0.4521,
      "mean_token_accuracy": 0.8556160926818848,
      "num_tokens": 523794689.0,
      "step": 3288
    },
    {
      "epoch": 1.67293997965412,
      "grad_norm": 0.9815360307693481,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8480269312858582,
      "num_tokens": 523958641.0,
      "step": 3289
    },
    {
      "epoch": 1.6734486266531028,
      "grad_norm": 0.915729820728302,
      "learning_rate": 1e-05,
      "loss": 0.48,
      "mean_token_accuracy": 0.847346842288971,
      "num_tokens": 524128171.0,
      "step": 3290
    },
    {
      "epoch": 1.6739572736520856,
      "grad_norm": 0.9879436492919922,
      "learning_rate": 1e-05,
      "loss": 0.5075,
      "mean_token_accuracy": 0.8403210639953613,
      "num_tokens": 524291831.0,
      "step": 3291
    },
    {
      "epoch": 1.6744659206510681,
      "grad_norm": 0.9411258697509766,
      "learning_rate": 1e-05,
      "loss": 0.4742,
      "mean_token_accuracy": 0.847053050994873,
      "num_tokens": 524455562.0,
      "step": 3292
    },
    {
      "epoch": 1.6749745676500507,
      "grad_norm": 0.9418855309486389,
      "learning_rate": 1e-05,
      "loss": 0.4988,
      "mean_token_accuracy": 0.8415638208389282,
      "num_tokens": 524615758.0,
      "step": 3293
    },
    {
      "epoch": 1.6754832146490335,
      "grad_norm": 0.9279276728630066,
      "learning_rate": 1e-05,
      "loss": 0.4915,
      "mean_token_accuracy": 0.8443039655685425,
      "num_tokens": 524788549.0,
      "step": 3294
    },
    {
      "epoch": 1.6759918616480163,
      "grad_norm": 0.9746510982513428,
      "learning_rate": 1e-05,
      "loss": 0.4568,
      "mean_token_accuracy": 0.8545578122138977,
      "num_tokens": 524946069.0,
      "step": 3295
    },
    {
      "epoch": 1.676500508646999,
      "grad_norm": 0.955940842628479,
      "learning_rate": 1e-05,
      "loss": 0.5109,
      "mean_token_accuracy": 0.8380922079086304,
      "num_tokens": 525108018.0,
      "step": 3296
    },
    {
      "epoch": 1.6770091556459819,
      "grad_norm": 0.9279923439025879,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.8598830699920654,
      "num_tokens": 525261353.0,
      "step": 3297
    },
    {
      "epoch": 1.6775178026449644,
      "grad_norm": 0.883661687374115,
      "learning_rate": 1e-05,
      "loss": 0.4715,
      "mean_token_accuracy": 0.8499032258987427,
      "num_tokens": 525429783.0,
      "step": 3298
    },
    {
      "epoch": 1.678026449643947,
      "grad_norm": 0.9393394589424133,
      "learning_rate": 1e-05,
      "loss": 0.4914,
      "mean_token_accuracy": 0.8439207673072815,
      "num_tokens": 525596465.0,
      "step": 3299
    },
    {
      "epoch": 1.6785350966429298,
      "grad_norm": 0.9309011101722717,
      "learning_rate": 1e-05,
      "loss": 0.4995,
      "mean_token_accuracy": 0.8434082269668579,
      "num_tokens": 525756468.0,
      "step": 3300
    },
    {
      "epoch": 1.6790437436419126,
      "grad_norm": 0.8695400357246399,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "mean_token_accuracy": 0.8543292284011841,
      "num_tokens": 525916556.0,
      "step": 3301
    },
    {
      "epoch": 1.6795523906408953,
      "grad_norm": 1.1616344451904297,
      "learning_rate": 1e-05,
      "loss": 0.4591,
      "mean_token_accuracy": 0.8515818119049072,
      "num_tokens": 526072332.0,
      "step": 3302
    },
    {
      "epoch": 1.680061037639878,
      "grad_norm": 0.9855133891105652,
      "learning_rate": 1e-05,
      "loss": 0.4694,
      "mean_token_accuracy": 0.8512064218521118,
      "num_tokens": 526232023.0,
      "step": 3303
    },
    {
      "epoch": 1.6805696846388605,
      "grad_norm": 0.922696053981781,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8474718332290649,
      "num_tokens": 526390503.0,
      "step": 3304
    },
    {
      "epoch": 1.6810783316378433,
      "grad_norm": 0.9263838529586792,
      "learning_rate": 1e-05,
      "loss": 0.5006,
      "mean_token_accuracy": 0.8425577282905579,
      "num_tokens": 526553435.0,
      "step": 3305
    },
    {
      "epoch": 1.681586978636826,
      "grad_norm": 0.9973815679550171,
      "learning_rate": 1e-05,
      "loss": 0.483,
      "mean_token_accuracy": 0.8466613292694092,
      "num_tokens": 526702092.0,
      "step": 3306
    },
    {
      "epoch": 1.6820956256358088,
      "grad_norm": 0.9127100110054016,
      "learning_rate": 1e-05,
      "loss": 0.447,
      "mean_token_accuracy": 0.8569740653038025,
      "num_tokens": 526864272.0,
      "step": 3307
    },
    {
      "epoch": 1.6826042726347916,
      "grad_norm": 0.8550009727478027,
      "learning_rate": 1e-05,
      "loss": 0.4777,
      "mean_token_accuracy": 0.8481435775756836,
      "num_tokens": 527035121.0,
      "step": 3308
    },
    {
      "epoch": 1.6831129196337742,
      "grad_norm": 0.8917046189308167,
      "learning_rate": 1e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8557920455932617,
      "num_tokens": 527188367.0,
      "step": 3309
    },
    {
      "epoch": 1.6836215666327567,
      "grad_norm": 0.9677354097366333,
      "learning_rate": 1e-05,
      "loss": 0.4872,
      "mean_token_accuracy": 0.8458775281906128,
      "num_tokens": 527350842.0,
      "step": 3310
    },
    {
      "epoch": 1.6841302136317395,
      "grad_norm": 0.9248644709587097,
      "learning_rate": 1e-05,
      "loss": 0.5035,
      "mean_token_accuracy": 0.84101402759552,
      "num_tokens": 527515994.0,
      "step": 3311
    },
    {
      "epoch": 1.6846388606307223,
      "grad_norm": 0.9055294394493103,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.8526661396026611,
      "num_tokens": 527685407.0,
      "step": 3312
    },
    {
      "epoch": 1.685147507629705,
      "grad_norm": 0.9836543202400208,
      "learning_rate": 1e-05,
      "loss": 0.4398,
      "mean_token_accuracy": 0.8585208654403687,
      "num_tokens": 527850629.0,
      "step": 3313
    },
    {
      "epoch": 1.6856561546286877,
      "grad_norm": 0.9496813416481018,
      "learning_rate": 1e-05,
      "loss": 0.471,
      "mean_token_accuracy": 0.8488203883171082,
      "num_tokens": 528011456.0,
      "step": 3314
    },
    {
      "epoch": 1.6861648016276702,
      "grad_norm": 0.9263770580291748,
      "learning_rate": 1e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.8631601333618164,
      "num_tokens": 528160030.0,
      "step": 3315
    },
    {
      "epoch": 1.686673448626653,
      "grad_norm": 1.0096482038497925,
      "learning_rate": 1e-05,
      "loss": 0.534,
      "mean_token_accuracy": 0.8320564031600952,
      "num_tokens": 528319045.0,
      "step": 3316
    },
    {
      "epoch": 1.6871820956256358,
      "grad_norm": 0.9198050498962402,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8489158153533936,
      "num_tokens": 528482912.0,
      "step": 3317
    },
    {
      "epoch": 1.6876907426246186,
      "grad_norm": 1.0153566598892212,
      "learning_rate": 1e-05,
      "loss": 0.4855,
      "mean_token_accuracy": 0.8455990552902222,
      "num_tokens": 528635401.0,
      "step": 3318
    },
    {
      "epoch": 1.6881993896236012,
      "grad_norm": 0.9170647859573364,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.85591059923172,
      "num_tokens": 528784018.0,
      "step": 3319
    },
    {
      "epoch": 1.688708036622584,
      "grad_norm": 1.0029700994491577,
      "learning_rate": 1e-05,
      "loss": 0.4593,
      "mean_token_accuracy": 0.8527106642723083,
      "num_tokens": 528936587.0,
      "step": 3320
    },
    {
      "epoch": 1.6892166836215665,
      "grad_norm": 0.9631544351577759,
      "learning_rate": 1e-05,
      "loss": 0.4952,
      "mean_token_accuracy": 0.8434855937957764,
      "num_tokens": 529091571.0,
      "step": 3321
    },
    {
      "epoch": 1.6897253306205493,
      "grad_norm": 0.9077914953231812,
      "learning_rate": 1e-05,
      "loss": 0.4763,
      "mean_token_accuracy": 0.8488157987594604,
      "num_tokens": 529263800.0,
      "step": 3322
    },
    {
      "epoch": 1.690233977619532,
      "grad_norm": 1.0570645332336426,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8580660820007324,
      "num_tokens": 529420232.0,
      "step": 3323
    },
    {
      "epoch": 1.6907426246185149,
      "grad_norm": 1.0088614225387573,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8567444682121277,
      "num_tokens": 529579783.0,
      "step": 3324
    },
    {
      "epoch": 1.6912512716174974,
      "grad_norm": 0.9414284825325012,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.8432716727256775,
      "num_tokens": 529751869.0,
      "step": 3325
    },
    {
      "epoch": 1.69175991861648,
      "grad_norm": 0.9591031670570374,
      "learning_rate": 1e-05,
      "loss": 0.4377,
      "mean_token_accuracy": 0.8570700883865356,
      "num_tokens": 529922657.0,
      "step": 3326
    },
    {
      "epoch": 1.6922685656154628,
      "grad_norm": 1.011871099472046,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "mean_token_accuracy": 0.8572059869766235,
      "num_tokens": 530083357.0,
      "step": 3327
    },
    {
      "epoch": 1.6927772126144456,
      "grad_norm": 1.0337574481964111,
      "learning_rate": 1e-05,
      "loss": 0.4375,
      "mean_token_accuracy": 0.858768105506897,
      "num_tokens": 530237269.0,
      "step": 3328
    },
    {
      "epoch": 1.6932858596134284,
      "grad_norm": 1.017116904258728,
      "learning_rate": 1e-05,
      "loss": 0.4935,
      "mean_token_accuracy": 0.8441458940505981,
      "num_tokens": 530400984.0,
      "step": 3329
    },
    {
      "epoch": 1.693794506612411,
      "grad_norm": 0.9320374131202698,
      "learning_rate": 1e-05,
      "loss": 0.479,
      "mean_token_accuracy": 0.8480983972549438,
      "num_tokens": 530563601.0,
      "step": 3330
    },
    {
      "epoch": 1.6943031536113937,
      "grad_norm": 0.9566795825958252,
      "learning_rate": 1e-05,
      "loss": 0.4789,
      "mean_token_accuracy": 0.8461617827415466,
      "num_tokens": 530706352.0,
      "step": 3331
    },
    {
      "epoch": 1.6948118006103763,
      "grad_norm": 0.9467067718505859,
      "learning_rate": 1e-05,
      "loss": 0.4941,
      "mean_token_accuracy": 0.8445472717285156,
      "num_tokens": 530875466.0,
      "step": 3332
    },
    {
      "epoch": 1.695320447609359,
      "grad_norm": 0.9797875285148621,
      "learning_rate": 1e-05,
      "loss": 0.4903,
      "mean_token_accuracy": 0.844483494758606,
      "num_tokens": 531040970.0,
      "step": 3333
    },
    {
      "epoch": 1.6958290946083419,
      "grad_norm": 0.9778969883918762,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8568530082702637,
      "num_tokens": 531199906.0,
      "step": 3334
    },
    {
      "epoch": 1.6963377416073246,
      "grad_norm": 0.9380336403846741,
      "learning_rate": 1e-05,
      "loss": 0.4665,
      "mean_token_accuracy": 0.8512696027755737,
      "num_tokens": 531354083.0,
      "step": 3335
    },
    {
      "epoch": 1.6968463886063072,
      "grad_norm": 0.9415258169174194,
      "learning_rate": 1e-05,
      "loss": 0.4728,
      "mean_token_accuracy": 0.8497647643089294,
      "num_tokens": 531509317.0,
      "step": 3336
    },
    {
      "epoch": 1.6973550356052898,
      "grad_norm": 1.0049771070480347,
      "learning_rate": 1e-05,
      "loss": 0.4709,
      "mean_token_accuracy": 0.8508015871047974,
      "num_tokens": 531661237.0,
      "step": 3337
    },
    {
      "epoch": 1.6978636826042726,
      "grad_norm": 0.9479979276657104,
      "learning_rate": 1e-05,
      "loss": 0.4649,
      "mean_token_accuracy": 0.8530048727989197,
      "num_tokens": 531821184.0,
      "step": 3338
    },
    {
      "epoch": 1.6983723296032553,
      "grad_norm": 0.9929404854774475,
      "learning_rate": 1e-05,
      "loss": 0.4419,
      "mean_token_accuracy": 0.8580644130706787,
      "num_tokens": 531968990.0,
      "step": 3339
    },
    {
      "epoch": 1.6988809766022381,
      "grad_norm": 0.9658786058425903,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8487871289253235,
      "num_tokens": 532137310.0,
      "step": 3340
    },
    {
      "epoch": 1.6993896236012207,
      "grad_norm": 1.012675404548645,
      "learning_rate": 1e-05,
      "loss": 0.5001,
      "mean_token_accuracy": 0.8426909446716309,
      "num_tokens": 532289496.0,
      "step": 3341
    },
    {
      "epoch": 1.6998982706002035,
      "grad_norm": 1.0348429679870605,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8566396832466125,
      "num_tokens": 532440662.0,
      "step": 3342
    },
    {
      "epoch": 1.700406917599186,
      "grad_norm": 0.9613810181617737,
      "learning_rate": 1e-05,
      "loss": 0.4637,
      "mean_token_accuracy": 0.852282702922821,
      "num_tokens": 532597195.0,
      "step": 3343
    },
    {
      "epoch": 1.7009155645981688,
      "grad_norm": 0.9847121238708496,
      "learning_rate": 1e-05,
      "loss": 0.4641,
      "mean_token_accuracy": 0.8513263463973999,
      "num_tokens": 532750623.0,
      "step": 3344
    },
    {
      "epoch": 1.7014242115971516,
      "grad_norm": 0.9310582280158997,
      "learning_rate": 1e-05,
      "loss": 0.4507,
      "mean_token_accuracy": 0.8549737930297852,
      "num_tokens": 532913640.0,
      "step": 3345
    },
    {
      "epoch": 1.7019328585961344,
      "grad_norm": 0.9432613849639893,
      "learning_rate": 1e-05,
      "loss": 0.4538,
      "mean_token_accuracy": 0.8539129495620728,
      "num_tokens": 533079069.0,
      "step": 3346
    },
    {
      "epoch": 1.702441505595117,
      "grad_norm": 0.9648125767707825,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8555746078491211,
      "num_tokens": 533235104.0,
      "step": 3347
    },
    {
      "epoch": 1.7029501525940995,
      "grad_norm": 0.9137231111526489,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8608599901199341,
      "num_tokens": 533404987.0,
      "step": 3348
    },
    {
      "epoch": 1.7034587995930823,
      "grad_norm": 0.9805552959442139,
      "learning_rate": 1e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8472720384597778,
      "num_tokens": 533563719.0,
      "step": 3349
    },
    {
      "epoch": 1.7039674465920651,
      "grad_norm": 1.0190812349319458,
      "learning_rate": 1e-05,
      "loss": 0.4625,
      "mean_token_accuracy": 0.8517487049102783,
      "num_tokens": 533715865.0,
      "step": 3350
    },
    {
      "epoch": 1.704476093591048,
      "grad_norm": 1.0097111463546753,
      "learning_rate": 1e-05,
      "loss": 0.4783,
      "mean_token_accuracy": 0.8472844958305359,
      "num_tokens": 533870888.0,
      "step": 3351
    },
    {
      "epoch": 1.7049847405900305,
      "grad_norm": 1.0112862586975098,
      "learning_rate": 1e-05,
      "loss": 0.4807,
      "mean_token_accuracy": 0.8472076654434204,
      "num_tokens": 534023360.0,
      "step": 3352
    },
    {
      "epoch": 1.7054933875890133,
      "grad_norm": 1.0216953754425049,
      "learning_rate": 1e-05,
      "loss": 0.4389,
      "mean_token_accuracy": 0.8584215641021729,
      "num_tokens": 534175128.0,
      "step": 3353
    },
    {
      "epoch": 1.7060020345879958,
      "grad_norm": 1.0892550945281982,
      "learning_rate": 1e-05,
      "loss": 0.4712,
      "mean_token_accuracy": 0.8495419025421143,
      "num_tokens": 534338342.0,
      "step": 3354
    },
    {
      "epoch": 1.7065106815869786,
      "grad_norm": 0.9826159477233887,
      "learning_rate": 1e-05,
      "loss": 0.472,
      "mean_token_accuracy": 0.8509111404418945,
      "num_tokens": 534485674.0,
      "step": 3355
    },
    {
      "epoch": 1.7070193285859614,
      "grad_norm": 1.0366008281707764,
      "learning_rate": 1e-05,
      "loss": 0.4733,
      "mean_token_accuracy": 0.8497737050056458,
      "num_tokens": 534643540.0,
      "step": 3356
    },
    {
      "epoch": 1.7075279755849442,
      "grad_norm": 0.9178564548492432,
      "learning_rate": 1e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8565972447395325,
      "num_tokens": 534800038.0,
      "step": 3357
    },
    {
      "epoch": 1.7080366225839267,
      "grad_norm": 1.001478672027588,
      "learning_rate": 1e-05,
      "loss": 0.4895,
      "mean_token_accuracy": 0.8442537784576416,
      "num_tokens": 534946089.0,
      "step": 3358
    },
    {
      "epoch": 1.7085452695829093,
      "grad_norm": 1.0083726644515991,
      "learning_rate": 1e-05,
      "loss": 0.4632,
      "mean_token_accuracy": 0.851345419883728,
      "num_tokens": 535099631.0,
      "step": 3359
    },
    {
      "epoch": 1.709053916581892,
      "grad_norm": 0.8685449361801147,
      "learning_rate": 1e-05,
      "loss": 0.5055,
      "mean_token_accuracy": 0.8397232294082642,
      "num_tokens": 535248651.0,
      "step": 3360
    },
    {
      "epoch": 1.7095625635808749,
      "grad_norm": 1.0024147033691406,
      "learning_rate": 1e-05,
      "loss": 0.5107,
      "mean_token_accuracy": 0.8402670621871948,
      "num_tokens": 535415988.0,
      "step": 3361
    },
    {
      "epoch": 1.7100712105798577,
      "grad_norm": 2.0469908714294434,
      "learning_rate": 1e-05,
      "loss": 0.4683,
      "mean_token_accuracy": 0.8498460650444031,
      "num_tokens": 535572330.0,
      "step": 3362
    },
    {
      "epoch": 1.7105798575788402,
      "grad_norm": 0.9829951524734497,
      "learning_rate": 1e-05,
      "loss": 0.4548,
      "mean_token_accuracy": 0.8559287786483765,
      "num_tokens": 535736424.0,
      "step": 3363
    },
    {
      "epoch": 1.711088504577823,
      "grad_norm": 0.948924720287323,
      "learning_rate": 1e-05,
      "loss": 0.4865,
      "mean_token_accuracy": 0.8448041081428528,
      "num_tokens": 535900015.0,
      "step": 3364
    },
    {
      "epoch": 1.7115971515768056,
      "grad_norm": 1.0147477388381958,
      "learning_rate": 1e-05,
      "loss": 0.4883,
      "mean_token_accuracy": 0.8448580503463745,
      "num_tokens": 536054345.0,
      "step": 3365
    },
    {
      "epoch": 1.7121057985757884,
      "grad_norm": 0.9223432540893555,
      "learning_rate": 1e-05,
      "loss": 0.4805,
      "mean_token_accuracy": 0.8478121757507324,
      "num_tokens": 536213601.0,
      "step": 3366
    },
    {
      "epoch": 1.7126144455747712,
      "grad_norm": 0.9273245334625244,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.840658962726593,
      "num_tokens": 536381796.0,
      "step": 3367
    },
    {
      "epoch": 1.713123092573754,
      "grad_norm": 0.9985995888710022,
      "learning_rate": 1e-05,
      "loss": 0.4743,
      "mean_token_accuracy": 0.8496761918067932,
      "num_tokens": 536546179.0,
      "step": 3368
    },
    {
      "epoch": 1.7136317395727365,
      "grad_norm": 0.9761607050895691,
      "learning_rate": 1e-05,
      "loss": 0.4832,
      "mean_token_accuracy": 0.8463513255119324,
      "num_tokens": 536716604.0,
      "step": 3369
    },
    {
      "epoch": 1.714140386571719,
      "grad_norm": 0.9189473986625671,
      "learning_rate": 1e-05,
      "loss": 0.4958,
      "mean_token_accuracy": 0.8437198400497437,
      "num_tokens": 536873341.0,
      "step": 3370
    },
    {
      "epoch": 1.7146490335707019,
      "grad_norm": 1.053405523300171,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.8505102396011353,
      "num_tokens": 537029255.0,
      "step": 3371
    },
    {
      "epoch": 1.7151576805696847,
      "grad_norm": 0.952011227607727,
      "learning_rate": 1e-05,
      "loss": 0.4874,
      "mean_token_accuracy": 0.8456932306289673,
      "num_tokens": 537182989.0,
      "step": 3372
    },
    {
      "epoch": 1.7156663275686674,
      "grad_norm": 0.9287518858909607,
      "learning_rate": 1e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.8501575589179993,
      "num_tokens": 537353885.0,
      "step": 3373
    },
    {
      "epoch": 1.71617497456765,
      "grad_norm": 1.0261600017547607,
      "learning_rate": 1e-05,
      "loss": 0.4848,
      "mean_token_accuracy": 0.845282256603241,
      "num_tokens": 537519346.0,
      "step": 3374
    },
    {
      "epoch": 1.7166836215666328,
      "grad_norm": 1.0151768922805786,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "mean_token_accuracy": 0.8560212850570679,
      "num_tokens": 537690640.0,
      "step": 3375
    },
    {
      "epoch": 1.7171922685656154,
      "grad_norm": 0.9206743240356445,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8524140119552612,
      "num_tokens": 537849156.0,
      "step": 3376
    },
    {
      "epoch": 1.7177009155645981,
      "grad_norm": 1.1978951692581177,
      "learning_rate": 1e-05,
      "loss": 0.477,
      "mean_token_accuracy": 0.8484640121459961,
      "num_tokens": 538012098.0,
      "step": 3377
    },
    {
      "epoch": 1.718209562563581,
      "grad_norm": 0.9505128264427185,
      "learning_rate": 1e-05,
      "loss": 0.5071,
      "mean_token_accuracy": 0.8388642072677612,
      "num_tokens": 538167004.0,
      "step": 3378
    },
    {
      "epoch": 1.7187182095625637,
      "grad_norm": 1.0761715173721313,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.854249894618988,
      "num_tokens": 538306082.0,
      "step": 3379
    },
    {
      "epoch": 1.7192268565615463,
      "grad_norm": 1.0668213367462158,
      "learning_rate": 1e-05,
      "loss": 0.4613,
      "mean_token_accuracy": 0.8530638217926025,
      "num_tokens": 538455354.0,
      "step": 3380
    },
    {
      "epoch": 1.7197355035605288,
      "grad_norm": 0.8889974355697632,
      "learning_rate": 1e-05,
      "loss": 0.4448,
      "mean_token_accuracy": 0.8573060035705566,
      "num_tokens": 538622580.0,
      "step": 3381
    },
    {
      "epoch": 1.7202441505595116,
      "grad_norm": 1.1304421424865723,
      "learning_rate": 1e-05,
      "loss": 0.4882,
      "mean_token_accuracy": 0.84532630443573,
      "num_tokens": 538776413.0,
      "step": 3382
    },
    {
      "epoch": 1.7207527975584944,
      "grad_norm": 0.8981946706771851,
      "learning_rate": 1e-05,
      "loss": 0.4826,
      "mean_token_accuracy": 0.8455299139022827,
      "num_tokens": 538935256.0,
      "step": 3383
    },
    {
      "epoch": 1.7212614445574772,
      "grad_norm": 0.9575291872024536,
      "learning_rate": 1e-05,
      "loss": 0.4575,
      "mean_token_accuracy": 0.8531308174133301,
      "num_tokens": 539089207.0,
      "step": 3384
    },
    {
      "epoch": 1.7217700915564598,
      "grad_norm": 0.9755882620811462,
      "learning_rate": 1e-05,
      "loss": 0.4779,
      "mean_token_accuracy": 0.8488665223121643,
      "num_tokens": 539251537.0,
      "step": 3385
    },
    {
      "epoch": 1.7222787385554426,
      "grad_norm": 0.9240638613700867,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8592078685760498,
      "num_tokens": 539409402.0,
      "step": 3386
    },
    {
      "epoch": 1.7227873855544251,
      "grad_norm": 1.0655521154403687,
      "learning_rate": 1e-05,
      "loss": 0.498,
      "mean_token_accuracy": 0.8432095050811768,
      "num_tokens": 539571156.0,
      "step": 3387
    },
    {
      "epoch": 1.723296032553408,
      "grad_norm": 1.0824614763259888,
      "learning_rate": 1e-05,
      "loss": 0.4777,
      "mean_token_accuracy": 0.8460273742675781,
      "num_tokens": 539724109.0,
      "step": 3388
    },
    {
      "epoch": 1.7238046795523907,
      "grad_norm": 1.0291532278060913,
      "learning_rate": 1e-05,
      "loss": 0.4638,
      "mean_token_accuracy": 0.8528157472610474,
      "num_tokens": 539889499.0,
      "step": 3389
    },
    {
      "epoch": 1.7243133265513735,
      "grad_norm": 1.1965833902359009,
      "learning_rate": 1e-05,
      "loss": 0.4762,
      "mean_token_accuracy": 0.847673773765564,
      "num_tokens": 540055954.0,
      "step": 3390
    },
    {
      "epoch": 1.724821973550356,
      "grad_norm": 0.8850274085998535,
      "learning_rate": 1e-05,
      "loss": 0.4743,
      "mean_token_accuracy": 0.8486838936805725,
      "num_tokens": 540222107.0,
      "step": 3391
    },
    {
      "epoch": 1.7253306205493386,
      "grad_norm": 1.187686800956726,
      "learning_rate": 1e-05,
      "loss": 0.5221,
      "mean_token_accuracy": 0.8357970714569092,
      "num_tokens": 540380282.0,
      "step": 3392
    },
    {
      "epoch": 1.7258392675483214,
      "grad_norm": 1.1276291608810425,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8447475433349609,
      "num_tokens": 540548234.0,
      "step": 3393
    },
    {
      "epoch": 1.7263479145473042,
      "grad_norm": 0.9580877423286438,
      "learning_rate": 1e-05,
      "loss": 0.4854,
      "mean_token_accuracy": 0.8469442129135132,
      "num_tokens": 540703596.0,
      "step": 3394
    },
    {
      "epoch": 1.726856561546287,
      "grad_norm": 1.0890614986419678,
      "learning_rate": 1e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8463518619537354,
      "num_tokens": 540857729.0,
      "step": 3395
    },
    {
      "epoch": 1.7273652085452695,
      "grad_norm": 0.9593095779418945,
      "learning_rate": 1e-05,
      "loss": 0.4937,
      "mean_token_accuracy": 0.8443906307220459,
      "num_tokens": 541024110.0,
      "step": 3396
    },
    {
      "epoch": 1.7278738555442523,
      "grad_norm": 1.0046390295028687,
      "learning_rate": 1e-05,
      "loss": 0.4517,
      "mean_token_accuracy": 0.8553571701049805,
      "num_tokens": 541183931.0,
      "step": 3397
    },
    {
      "epoch": 1.7283825025432349,
      "grad_norm": 1.067029356956482,
      "learning_rate": 1e-05,
      "loss": 0.4856,
      "mean_token_accuracy": 0.8454616069793701,
      "num_tokens": 541343999.0,
      "step": 3398
    },
    {
      "epoch": 1.7288911495422177,
      "grad_norm": 1.0075136423110962,
      "learning_rate": 1e-05,
      "loss": 0.4784,
      "mean_token_accuracy": 0.8482916951179504,
      "num_tokens": 541501447.0,
      "step": 3399
    },
    {
      "epoch": 1.7293997965412005,
      "grad_norm": 0.9967670440673828,
      "learning_rate": 1e-05,
      "loss": 0.4791,
      "mean_token_accuracy": 0.8465551137924194,
      "num_tokens": 541666906.0,
      "step": 3400
    },
    {
      "epoch": 1.7299084435401832,
      "grad_norm": 0.9384946823120117,
      "learning_rate": 1e-05,
      "loss": 0.4472,
      "mean_token_accuracy": 0.8563881516456604,
      "num_tokens": 541823176.0,
      "step": 3401
    },
    {
      "epoch": 1.7304170905391658,
      "grad_norm": 0.9869144558906555,
      "learning_rate": 1e-05,
      "loss": 0.5131,
      "mean_token_accuracy": 0.8384538292884827,
      "num_tokens": 541998847.0,
      "step": 3402
    },
    {
      "epoch": 1.7309257375381484,
      "grad_norm": 0.9360270500183105,
      "learning_rate": 1e-05,
      "loss": 0.4833,
      "mean_token_accuracy": 0.8461273908615112,
      "num_tokens": 542167353.0,
      "step": 3403
    },
    {
      "epoch": 1.7314343845371312,
      "grad_norm": 0.9577409625053406,
      "learning_rate": 1e-05,
      "loss": 0.4472,
      "mean_token_accuracy": 0.8582079410552979,
      "num_tokens": 542330954.0,
      "step": 3404
    },
    {
      "epoch": 1.731943031536114,
      "grad_norm": 0.9778986573219299,
      "learning_rate": 1e-05,
      "loss": 0.4694,
      "mean_token_accuracy": 0.8505184650421143,
      "num_tokens": 542484284.0,
      "step": 3405
    },
    {
      "epoch": 1.7324516785350967,
      "grad_norm": 0.8824731111526489,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.8546308279037476,
      "num_tokens": 542649706.0,
      "step": 3406
    },
    {
      "epoch": 1.7329603255340793,
      "grad_norm": 1.0359128713607788,
      "learning_rate": 1e-05,
      "loss": 0.4899,
      "mean_token_accuracy": 0.8440340757369995,
      "num_tokens": 542812938.0,
      "step": 3407
    },
    {
      "epoch": 1.733468972533062,
      "grad_norm": 1.0210932493209839,
      "learning_rate": 1e-05,
      "loss": 0.5077,
      "mean_token_accuracy": 0.8378579020500183,
      "num_tokens": 542971174.0,
      "step": 3408
    },
    {
      "epoch": 1.7339776195320447,
      "grad_norm": 0.9628419876098633,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8492684960365295,
      "num_tokens": 543130106.0,
      "step": 3409
    },
    {
      "epoch": 1.7344862665310274,
      "grad_norm": 0.9311797022819519,
      "learning_rate": 1e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8529977798461914,
      "num_tokens": 543286863.0,
      "step": 3410
    },
    {
      "epoch": 1.7349949135300102,
      "grad_norm": 0.9846572875976562,
      "learning_rate": 1e-05,
      "loss": 0.4886,
      "mean_token_accuracy": 0.844512939453125,
      "num_tokens": 543444719.0,
      "step": 3411
    },
    {
      "epoch": 1.735503560528993,
      "grad_norm": 0.9851809144020081,
      "learning_rate": 1e-05,
      "loss": 0.5049,
      "mean_token_accuracy": 0.8395944237709045,
      "num_tokens": 543603599.0,
      "step": 3412
    },
    {
      "epoch": 1.7360122075279756,
      "grad_norm": 0.9662925004959106,
      "learning_rate": 1e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8492670655250549,
      "num_tokens": 543764939.0,
      "step": 3413
    },
    {
      "epoch": 1.7365208545269581,
      "grad_norm": 0.9885513186454773,
      "learning_rate": 1e-05,
      "loss": 0.4916,
      "mean_token_accuracy": 0.8415821194648743,
      "num_tokens": 543929074.0,
      "step": 3414
    },
    {
      "epoch": 1.737029501525941,
      "grad_norm": 0.9492428302764893,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8441232442855835,
      "num_tokens": 544087367.0,
      "step": 3415
    },
    {
      "epoch": 1.7375381485249237,
      "grad_norm": 0.9702391028404236,
      "learning_rate": 1e-05,
      "loss": 0.4893,
      "mean_token_accuracy": 0.8451366424560547,
      "num_tokens": 544254565.0,
      "step": 3416
    },
    {
      "epoch": 1.7380467955239065,
      "grad_norm": 0.9493220448493958,
      "learning_rate": 1e-05,
      "loss": 0.4651,
      "mean_token_accuracy": 0.8521139621734619,
      "num_tokens": 544423653.0,
      "step": 3417
    },
    {
      "epoch": 1.738555442522889,
      "grad_norm": 0.92607182264328,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "mean_token_accuracy": 0.8525439500808716,
      "num_tokens": 544594001.0,
      "step": 3418
    },
    {
      "epoch": 1.7390640895218719,
      "grad_norm": 0.9776398539543152,
      "learning_rate": 1e-05,
      "loss": 0.4866,
      "mean_token_accuracy": 0.8456598520278931,
      "num_tokens": 544743578.0,
      "step": 3419
    },
    {
      "epoch": 1.7395727365208544,
      "grad_norm": 0.9213610887527466,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.8459997177124023,
      "num_tokens": 544905906.0,
      "step": 3420
    },
    {
      "epoch": 1.7400813835198372,
      "grad_norm": 1.0641958713531494,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.8510778546333313,
      "num_tokens": 545068059.0,
      "step": 3421
    },
    {
      "epoch": 1.74059003051882,
      "grad_norm": 0.8985027074813843,
      "learning_rate": 1e-05,
      "loss": 0.4428,
      "mean_token_accuracy": 0.8590704202651978,
      "num_tokens": 545218275.0,
      "step": 3422
    },
    {
      "epoch": 1.7410986775178028,
      "grad_norm": 0.9916641712188721,
      "learning_rate": 1e-05,
      "loss": 0.5017,
      "mean_token_accuracy": 0.8406370282173157,
      "num_tokens": 545387771.0,
      "step": 3423
    },
    {
      "epoch": 1.7416073245167853,
      "grad_norm": 1.1028623580932617,
      "learning_rate": 1e-05,
      "loss": 0.4573,
      "mean_token_accuracy": 0.8541989326477051,
      "num_tokens": 545540433.0,
      "step": 3424
    },
    {
      "epoch": 1.742115971515768,
      "grad_norm": 1.007652759552002,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8547618389129639,
      "num_tokens": 545694394.0,
      "step": 3425
    },
    {
      "epoch": 1.7426246185147507,
      "grad_norm": 1.0314555168151855,
      "learning_rate": 1e-05,
      "loss": 0.4774,
      "mean_token_accuracy": 0.8469586372375488,
      "num_tokens": 545853945.0,
      "step": 3426
    },
    {
      "epoch": 1.7431332655137335,
      "grad_norm": 1.0091915130615234,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.8543394804000854,
      "num_tokens": 546017542.0,
      "step": 3427
    },
    {
      "epoch": 1.7436419125127163,
      "grad_norm": 1.021324634552002,
      "learning_rate": 1e-05,
      "loss": 0.4489,
      "mean_token_accuracy": 0.8562377095222473,
      "num_tokens": 546163829.0,
      "step": 3428
    },
    {
      "epoch": 1.7441505595116988,
      "grad_norm": 0.9751710891723633,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.8416702151298523,
      "num_tokens": 546326213.0,
      "step": 3429
    },
    {
      "epoch": 1.7446592065106816,
      "grad_norm": 0.9528321027755737,
      "learning_rate": 1e-05,
      "loss": 0.4336,
      "mean_token_accuracy": 0.8595539331436157,
      "num_tokens": 546476551.0,
      "step": 3430
    },
    {
      "epoch": 1.7451678535096642,
      "grad_norm": 0.8223140239715576,
      "learning_rate": 1e-05,
      "loss": 0.4742,
      "mean_token_accuracy": 0.8496396541595459,
      "num_tokens": 546647073.0,
      "step": 3431
    },
    {
      "epoch": 1.745676500508647,
      "grad_norm": 0.9821622967720032,
      "learning_rate": 1e-05,
      "loss": 0.446,
      "mean_token_accuracy": 0.8567888140678406,
      "num_tokens": 546808037.0,
      "step": 3432
    },
    {
      "epoch": 1.7461851475076298,
      "grad_norm": 0.9108555316925049,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8557436466217041,
      "num_tokens": 546973916.0,
      "step": 3433
    },
    {
      "epoch": 1.7466937945066126,
      "grad_norm": 0.9076216220855713,
      "learning_rate": 1e-05,
      "loss": 0.4517,
      "mean_token_accuracy": 0.8551619052886963,
      "num_tokens": 547138121.0,
      "step": 3434
    },
    {
      "epoch": 1.7472024415055951,
      "grad_norm": 0.9469711780548096,
      "learning_rate": 1e-05,
      "loss": 0.4623,
      "mean_token_accuracy": 0.8510894775390625,
      "num_tokens": 547294960.0,
      "step": 3435
    },
    {
      "epoch": 1.7477110885045777,
      "grad_norm": 0.8839051723480225,
      "learning_rate": 1e-05,
      "loss": 0.4721,
      "mean_token_accuracy": 0.8495970368385315,
      "num_tokens": 547455414.0,
      "step": 3436
    },
    {
      "epoch": 1.7482197355035605,
      "grad_norm": 0.9603871703147888,
      "learning_rate": 1e-05,
      "loss": 0.4371,
      "mean_token_accuracy": 0.8589041233062744,
      "num_tokens": 547604728.0,
      "step": 3437
    },
    {
      "epoch": 1.7487283825025433,
      "grad_norm": 0.9335736036300659,
      "learning_rate": 1e-05,
      "loss": 0.4744,
      "mean_token_accuracy": 0.850881814956665,
      "num_tokens": 547758420.0,
      "step": 3438
    },
    {
      "epoch": 1.749237029501526,
      "grad_norm": 0.9160023331642151,
      "learning_rate": 1e-05,
      "loss": 0.4584,
      "mean_token_accuracy": 0.8534228801727295,
      "num_tokens": 547909778.0,
      "step": 3439
    },
    {
      "epoch": 1.7497456765005086,
      "grad_norm": 0.9696451425552368,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.8512312173843384,
      "num_tokens": 548059321.0,
      "step": 3440
    },
    {
      "epoch": 1.7502543234994914,
      "grad_norm": 1.0125298500061035,
      "learning_rate": 1e-05,
      "loss": 0.4938,
      "mean_token_accuracy": 0.8438988924026489,
      "num_tokens": 548212480.0,
      "step": 3441
    },
    {
      "epoch": 1.750762970498474,
      "grad_norm": 0.9607093930244446,
      "learning_rate": 1e-05,
      "loss": 0.4945,
      "mean_token_accuracy": 0.8422509431838989,
      "num_tokens": 548363768.0,
      "step": 3442
    },
    {
      "epoch": 1.7512716174974567,
      "grad_norm": 0.9785159230232239,
      "learning_rate": 1e-05,
      "loss": 0.4682,
      "mean_token_accuracy": 0.8506906032562256,
      "num_tokens": 548513409.0,
      "step": 3443
    },
    {
      "epoch": 1.7517802644964395,
      "grad_norm": 1.0101968050003052,
      "learning_rate": 1e-05,
      "loss": 0.4716,
      "mean_token_accuracy": 0.8490024209022522,
      "num_tokens": 548670132.0,
      "step": 3444
    },
    {
      "epoch": 1.7522889114954223,
      "grad_norm": 0.9235634207725525,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8533692359924316,
      "num_tokens": 548823795.0,
      "step": 3445
    },
    {
      "epoch": 1.7527975584944049,
      "grad_norm": 1.0365976095199585,
      "learning_rate": 1e-05,
      "loss": 0.4686,
      "mean_token_accuracy": 0.8518203496932983,
      "num_tokens": 548986310.0,
      "step": 3446
    },
    {
      "epoch": 1.7533062054933874,
      "grad_norm": 1.0411449670791626,
      "learning_rate": 1e-05,
      "loss": 0.4776,
      "mean_token_accuracy": 0.8472661375999451,
      "num_tokens": 549140095.0,
      "step": 3447
    },
    {
      "epoch": 1.7538148524923702,
      "grad_norm": 0.8725220561027527,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.8547745943069458,
      "num_tokens": 549294340.0,
      "step": 3448
    },
    {
      "epoch": 1.754323499491353,
      "grad_norm": 0.9153843522071838,
      "learning_rate": 1e-05,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8675376176834106,
      "num_tokens": 549453074.0,
      "step": 3449
    },
    {
      "epoch": 1.7548321464903358,
      "grad_norm": 1.0639971494674683,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8492655754089355,
      "num_tokens": 549610886.0,
      "step": 3450
    },
    {
      "epoch": 1.7553407934893184,
      "grad_norm": 0.9058477878570557,
      "learning_rate": 1e-05,
      "loss": 0.4443,
      "mean_token_accuracy": 0.8552689552307129,
      "num_tokens": 549767763.0,
      "step": 3451
    },
    {
      "epoch": 1.7558494404883012,
      "grad_norm": 0.9886971116065979,
      "learning_rate": 1e-05,
      "loss": 0.4911,
      "mean_token_accuracy": 0.8447088003158569,
      "num_tokens": 549926561.0,
      "step": 3452
    },
    {
      "epoch": 1.7563580874872837,
      "grad_norm": 1.0823899507522583,
      "learning_rate": 1e-05,
      "loss": 0.4934,
      "mean_token_accuracy": 0.8419996500015259,
      "num_tokens": 550090259.0,
      "step": 3453
    },
    {
      "epoch": 1.7568667344862665,
      "grad_norm": 0.9325345754623413,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8606714606285095,
      "num_tokens": 550244848.0,
      "step": 3454
    },
    {
      "epoch": 1.7573753814852493,
      "grad_norm": 1.036022663116455,
      "learning_rate": 1e-05,
      "loss": 0.4409,
      "mean_token_accuracy": 0.8577229976654053,
      "num_tokens": 550406425.0,
      "step": 3455
    },
    {
      "epoch": 1.757884028484232,
      "grad_norm": 0.9253838658332825,
      "learning_rate": 1e-05,
      "loss": 0.453,
      "mean_token_accuracy": 0.8544574975967407,
      "num_tokens": 550565225.0,
      "step": 3456
    },
    {
      "epoch": 1.7583926754832147,
      "grad_norm": 0.9430886507034302,
      "learning_rate": 1e-05,
      "loss": 0.4815,
      "mean_token_accuracy": 0.8463444709777832,
      "num_tokens": 550724486.0,
      "step": 3457
    },
    {
      "epoch": 1.7589013224821972,
      "grad_norm": 1.0232832431793213,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8553723096847534,
      "num_tokens": 550869158.0,
      "step": 3458
    },
    {
      "epoch": 1.75940996948118,
      "grad_norm": 0.8895917534828186,
      "learning_rate": 1e-05,
      "loss": 0.4964,
      "mean_token_accuracy": 0.8435969352722168,
      "num_tokens": 551034748.0,
      "step": 3459
    },
    {
      "epoch": 1.7599186164801628,
      "grad_norm": 1.019453525543213,
      "learning_rate": 1e-05,
      "loss": 0.4516,
      "mean_token_accuracy": 0.8550435900688171,
      "num_tokens": 551189311.0,
      "step": 3460
    },
    {
      "epoch": 1.7604272634791456,
      "grad_norm": 0.9567839503288269,
      "learning_rate": 1e-05,
      "loss": 0.4844,
      "mean_token_accuracy": 0.8456395864486694,
      "num_tokens": 551335943.0,
      "step": 3461
    },
    {
      "epoch": 1.7609359104781281,
      "grad_norm": 1.0265204906463623,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.845983624458313,
      "num_tokens": 551500042.0,
      "step": 3462
    },
    {
      "epoch": 1.761444557477111,
      "grad_norm": 0.9160285592079163,
      "learning_rate": 1e-05,
      "loss": 0.488,
      "mean_token_accuracy": 0.8447883725166321,
      "num_tokens": 551661086.0,
      "step": 3463
    },
    {
      "epoch": 1.7619532044760935,
      "grad_norm": 0.9425185918807983,
      "learning_rate": 1e-05,
      "loss": 0.4344,
      "mean_token_accuracy": 0.8605625033378601,
      "num_tokens": 551811847.0,
      "step": 3464
    },
    {
      "epoch": 1.7624618514750763,
      "grad_norm": 0.9993616938591003,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8575040102005005,
      "num_tokens": 551963564.0,
      "step": 3465
    },
    {
      "epoch": 1.762970498474059,
      "grad_norm": 0.9410571455955505,
      "learning_rate": 1e-05,
      "loss": 0.4712,
      "mean_token_accuracy": 0.8500432372093201,
      "num_tokens": 552131236.0,
      "step": 3466
    },
    {
      "epoch": 1.7634791454730419,
      "grad_norm": 1.0863051414489746,
      "learning_rate": 1e-05,
      "loss": 0.4748,
      "mean_token_accuracy": 0.8502572178840637,
      "num_tokens": 552287004.0,
      "step": 3467
    },
    {
      "epoch": 1.7639877924720244,
      "grad_norm": 0.9163442254066467,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8505086898803711,
      "num_tokens": 552456755.0,
      "step": 3468
    },
    {
      "epoch": 1.764496439471007,
      "grad_norm": 1.024658203125,
      "learning_rate": 1e-05,
      "loss": 0.4303,
      "mean_token_accuracy": 0.8616164326667786,
      "num_tokens": 552619812.0,
      "step": 3469
    },
    {
      "epoch": 1.7650050864699898,
      "grad_norm": 0.965026319026947,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.850260317325592,
      "num_tokens": 552777207.0,
      "step": 3470
    },
    {
      "epoch": 1.7655137334689726,
      "grad_norm": 0.91949063539505,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.8517131805419922,
      "num_tokens": 552931843.0,
      "step": 3471
    },
    {
      "epoch": 1.7660223804679553,
      "grad_norm": 0.9592922925949097,
      "learning_rate": 1e-05,
      "loss": 0.4775,
      "mean_token_accuracy": 0.8479909300804138,
      "num_tokens": 553087376.0,
      "step": 3472
    },
    {
      "epoch": 1.766531027466938,
      "grad_norm": 0.8795325756072998,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8521772623062134,
      "num_tokens": 553258033.0,
      "step": 3473
    },
    {
      "epoch": 1.7670396744659207,
      "grad_norm": 1.0181102752685547,
      "learning_rate": 1e-05,
      "loss": 0.477,
      "mean_token_accuracy": 0.8480532169342041,
      "num_tokens": 553406997.0,
      "step": 3474
    },
    {
      "epoch": 1.7675483214649033,
      "grad_norm": 0.9451496005058289,
      "learning_rate": 1e-05,
      "loss": 0.4944,
      "mean_token_accuracy": 0.8428871631622314,
      "num_tokens": 553571801.0,
      "step": 3475
    },
    {
      "epoch": 1.768056968463886,
      "grad_norm": 1.011832356452942,
      "learning_rate": 1e-05,
      "loss": 0.505,
      "mean_token_accuracy": 0.8400646448135376,
      "num_tokens": 553726380.0,
      "step": 3476
    },
    {
      "epoch": 1.7685656154628688,
      "grad_norm": 0.9240113496780396,
      "learning_rate": 1e-05,
      "loss": 0.4651,
      "mean_token_accuracy": 0.8522868752479553,
      "num_tokens": 553891035.0,
      "step": 3477
    },
    {
      "epoch": 1.7690742624618516,
      "grad_norm": 0.9595242142677307,
      "learning_rate": 1e-05,
      "loss": 0.5011,
      "mean_token_accuracy": 0.8421164751052856,
      "num_tokens": 554060617.0,
      "step": 3478
    },
    {
      "epoch": 1.7695829094608342,
      "grad_norm": 0.8867053389549255,
      "learning_rate": 1e-05,
      "loss": 0.4858,
      "mean_token_accuracy": 0.8462978601455688,
      "num_tokens": 554226977.0,
      "step": 3479
    },
    {
      "epoch": 1.7700915564598168,
      "grad_norm": 0.9090108871459961,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "mean_token_accuracy": 0.8542250394821167,
      "num_tokens": 554385731.0,
      "step": 3480
    },
    {
      "epoch": 1.7706002034587995,
      "grad_norm": 0.9475415349006653,
      "learning_rate": 1e-05,
      "loss": 0.467,
      "mean_token_accuracy": 0.8525370359420776,
      "num_tokens": 554555504.0,
      "step": 3481
    },
    {
      "epoch": 1.7711088504577823,
      "grad_norm": 0.9586099982261658,
      "learning_rate": 1e-05,
      "loss": 0.4636,
      "mean_token_accuracy": 0.8518487215042114,
      "num_tokens": 554714277.0,
      "step": 3482
    },
    {
      "epoch": 1.7716174974567651,
      "grad_norm": 0.8920243382453918,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8517918586730957,
      "num_tokens": 554875296.0,
      "step": 3483
    },
    {
      "epoch": 1.7721261444557477,
      "grad_norm": 0.8927356600761414,
      "learning_rate": 1e-05,
      "loss": 0.484,
      "mean_token_accuracy": 0.8461599946022034,
      "num_tokens": 555033920.0,
      "step": 3484
    },
    {
      "epoch": 1.7726347914547305,
      "grad_norm": 2.070472002029419,
      "learning_rate": 1e-05,
      "loss": 0.4716,
      "mean_token_accuracy": 0.8495326042175293,
      "num_tokens": 555194410.0,
      "step": 3485
    },
    {
      "epoch": 1.773143438453713,
      "grad_norm": 0.9416306614875793,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8557838797569275,
      "num_tokens": 555371365.0,
      "step": 3486
    },
    {
      "epoch": 1.7736520854526958,
      "grad_norm": 0.9121522903442383,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8585523366928101,
      "num_tokens": 555531745.0,
      "step": 3487
    },
    {
      "epoch": 1.7741607324516786,
      "grad_norm": 0.9331509470939636,
      "learning_rate": 1e-05,
      "loss": 0.4758,
      "mean_token_accuracy": 0.8486753702163696,
      "num_tokens": 555695553.0,
      "step": 3488
    },
    {
      "epoch": 1.7746693794506614,
      "grad_norm": 0.9343940615653992,
      "learning_rate": 1e-05,
      "loss": 0.491,
      "mean_token_accuracy": 0.8444896936416626,
      "num_tokens": 555857378.0,
      "step": 3489
    },
    {
      "epoch": 1.775178026449644,
      "grad_norm": 0.9072713255882263,
      "learning_rate": 1e-05,
      "loss": 0.4805,
      "mean_token_accuracy": 0.8467872142791748,
      "num_tokens": 556006070.0,
      "step": 3490
    },
    {
      "epoch": 1.7756866734486265,
      "grad_norm": 0.8870847225189209,
      "learning_rate": 1e-05,
      "loss": 0.4462,
      "mean_token_accuracy": 0.8569936156272888,
      "num_tokens": 556165767.0,
      "step": 3491
    },
    {
      "epoch": 1.7761953204476093,
      "grad_norm": 0.9200275540351868,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8556621074676514,
      "num_tokens": 556324330.0,
      "step": 3492
    },
    {
      "epoch": 1.776703967446592,
      "grad_norm": 0.9575535655021667,
      "learning_rate": 1e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8579325675964355,
      "num_tokens": 556483980.0,
      "step": 3493
    },
    {
      "epoch": 1.7772126144455749,
      "grad_norm": 0.995994508266449,
      "learning_rate": 1e-05,
      "loss": 0.4761,
      "mean_token_accuracy": 0.8466376662254333,
      "num_tokens": 556639463.0,
      "step": 3494
    },
    {
      "epoch": 1.7777212614445574,
      "grad_norm": 0.9333937764167786,
      "learning_rate": 1e-05,
      "loss": 0.5029,
      "mean_token_accuracy": 0.840882420539856,
      "num_tokens": 556796632.0,
      "step": 3495
    },
    {
      "epoch": 1.7782299084435402,
      "grad_norm": 0.9078387022018433,
      "learning_rate": 1e-05,
      "loss": 0.4568,
      "mean_token_accuracy": 0.854224681854248,
      "num_tokens": 556964733.0,
      "step": 3496
    },
    {
      "epoch": 1.7787385554425228,
      "grad_norm": 0.9770086407661438,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.8511612415313721,
      "num_tokens": 557132319.0,
      "step": 3497
    },
    {
      "epoch": 1.7792472024415056,
      "grad_norm": 0.9639984369277954,
      "learning_rate": 1e-05,
      "loss": 0.4995,
      "mean_token_accuracy": 0.8427587747573853,
      "num_tokens": 557293193.0,
      "step": 3498
    },
    {
      "epoch": 1.7797558494404884,
      "grad_norm": 0.9718248844146729,
      "learning_rate": 1e-05,
      "loss": 0.455,
      "mean_token_accuracy": 0.8534946441650391,
      "num_tokens": 557457261.0,
      "step": 3499
    },
    {
      "epoch": 1.7802644964394712,
      "grad_norm": 0.8903071284294128,
      "learning_rate": 1e-05,
      "loss": 0.4606,
      "mean_token_accuracy": 0.8538823127746582,
      "num_tokens": 557609634.0,
      "step": 3500
    },
    {
      "epoch": 1.7807731434384537,
      "grad_norm": 0.9682722091674805,
      "learning_rate": 1e-05,
      "loss": 0.461,
      "mean_token_accuracy": 0.8518296480178833,
      "num_tokens": 557760431.0,
      "step": 3501
    },
    {
      "epoch": 1.7812817904374363,
      "grad_norm": 0.9154918193817139,
      "learning_rate": 1e-05,
      "loss": 0.451,
      "mean_token_accuracy": 0.8548795580863953,
      "num_tokens": 557933267.0,
      "step": 3502
    },
    {
      "epoch": 1.781790437436419,
      "grad_norm": 0.9262132048606873,
      "learning_rate": 1e-05,
      "loss": 0.4929,
      "mean_token_accuracy": 0.8442500233650208,
      "num_tokens": 558081783.0,
      "step": 3503
    },
    {
      "epoch": 1.7822990844354019,
      "grad_norm": 1.055605411529541,
      "learning_rate": 1e-05,
      "loss": 0.5099,
      "mean_token_accuracy": 0.8384379744529724,
      "num_tokens": 558238439.0,
      "step": 3504
    },
    {
      "epoch": 1.7828077314343846,
      "grad_norm": 0.9379321336746216,
      "learning_rate": 1e-05,
      "loss": 0.4678,
      "mean_token_accuracy": 0.8530690670013428,
      "num_tokens": 558406721.0,
      "step": 3505
    },
    {
      "epoch": 1.7833163784333672,
      "grad_norm": 0.8967541456222534,
      "learning_rate": 1e-05,
      "loss": 0.4885,
      "mean_token_accuracy": 0.8444415330886841,
      "num_tokens": 558564521.0,
      "step": 3506
    },
    {
      "epoch": 1.78382502543235,
      "grad_norm": 0.9698563814163208,
      "learning_rate": 1e-05,
      "loss": 0.4949,
      "mean_token_accuracy": 0.8436890840530396,
      "num_tokens": 558715095.0,
      "step": 3507
    },
    {
      "epoch": 1.7843336724313326,
      "grad_norm": 0.9144454598426819,
      "learning_rate": 1e-05,
      "loss": 0.4926,
      "mean_token_accuracy": 0.8440608978271484,
      "num_tokens": 558870711.0,
      "step": 3508
    },
    {
      "epoch": 1.7848423194303153,
      "grad_norm": 1.0247725248336792,
      "learning_rate": 1e-05,
      "loss": 0.4907,
      "mean_token_accuracy": 0.8435952663421631,
      "num_tokens": 559037633.0,
      "step": 3509
    },
    {
      "epoch": 1.7853509664292981,
      "grad_norm": 0.9614184498786926,
      "learning_rate": 1e-05,
      "loss": 0.4447,
      "mean_token_accuracy": 0.8556063175201416,
      "num_tokens": 559184486.0,
      "step": 3510
    },
    {
      "epoch": 1.785859613428281,
      "grad_norm": 0.9238786101341248,
      "learning_rate": 1e-05,
      "loss": 0.4756,
      "mean_token_accuracy": 0.848760724067688,
      "num_tokens": 559340860.0,
      "step": 3511
    },
    {
      "epoch": 1.7863682604272635,
      "grad_norm": 1.0116618871688843,
      "learning_rate": 1e-05,
      "loss": 0.4517,
      "mean_token_accuracy": 0.8541946411132812,
      "num_tokens": 559487831.0,
      "step": 3512
    },
    {
      "epoch": 1.786876907426246,
      "grad_norm": 1.0261667966842651,
      "learning_rate": 1e-05,
      "loss": 0.478,
      "mean_token_accuracy": 0.8468774557113647,
      "num_tokens": 559650109.0,
      "step": 3513
    },
    {
      "epoch": 1.7873855544252288,
      "grad_norm": 0.8953947424888611,
      "learning_rate": 1e-05,
      "loss": 0.4339,
      "mean_token_accuracy": 0.8612636923789978,
      "num_tokens": 559805689.0,
      "step": 3514
    },
    {
      "epoch": 1.7878942014242116,
      "grad_norm": 0.8966084122657776,
      "learning_rate": 1e-05,
      "loss": 0.4536,
      "mean_token_accuracy": 0.8546212911605835,
      "num_tokens": 559966866.0,
      "step": 3515
    },
    {
      "epoch": 1.7884028484231944,
      "grad_norm": 0.8931212425231934,
      "learning_rate": 1e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.8599777221679688,
      "num_tokens": 560117177.0,
      "step": 3516
    },
    {
      "epoch": 1.788911495422177,
      "grad_norm": 0.9382513761520386,
      "learning_rate": 1e-05,
      "loss": 0.4479,
      "mean_token_accuracy": 0.8547048568725586,
      "num_tokens": 560278984.0,
      "step": 3517
    },
    {
      "epoch": 1.7894201424211598,
      "grad_norm": 0.879797101020813,
      "learning_rate": 1e-05,
      "loss": 0.4968,
      "mean_token_accuracy": 0.8421691060066223,
      "num_tokens": 560434600.0,
      "step": 3518
    },
    {
      "epoch": 1.7899287894201423,
      "grad_norm": 0.9176720976829529,
      "learning_rate": 1e-05,
      "loss": 0.4718,
      "mean_token_accuracy": 0.8502687215805054,
      "num_tokens": 560598239.0,
      "step": 3519
    },
    {
      "epoch": 1.7904374364191251,
      "grad_norm": 0.9511398077011108,
      "learning_rate": 1e-05,
      "loss": 0.4632,
      "mean_token_accuracy": 0.8514447212219238,
      "num_tokens": 560752929.0,
      "step": 3520
    },
    {
      "epoch": 1.790946083418108,
      "grad_norm": 0.8993409872055054,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.855494499206543,
      "num_tokens": 560905046.0,
      "step": 3521
    },
    {
      "epoch": 1.7914547304170907,
      "grad_norm": 1.001287579536438,
      "learning_rate": 1e-05,
      "loss": 0.4969,
      "mean_token_accuracy": 0.8412386178970337,
      "num_tokens": 561057758.0,
      "step": 3522
    },
    {
      "epoch": 1.7919633774160733,
      "grad_norm": 0.8689597249031067,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.8530293107032776,
      "num_tokens": 561215568.0,
      "step": 3523
    },
    {
      "epoch": 1.7924720244150558,
      "grad_norm": 1.0701673030853271,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.8540624380111694,
      "num_tokens": 561364341.0,
      "step": 3524
    },
    {
      "epoch": 1.7929806714140386,
      "grad_norm": 0.9774520397186279,
      "learning_rate": 1e-05,
      "loss": 0.4625,
      "mean_token_accuracy": 0.8525039553642273,
      "num_tokens": 561511545.0,
      "step": 3525
    },
    {
      "epoch": 1.7934893184130214,
      "grad_norm": 0.8649917840957642,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8515691757202148,
      "num_tokens": 561682192.0,
      "step": 3526
    },
    {
      "epoch": 1.7939979654120042,
      "grad_norm": 1.0334358215332031,
      "learning_rate": 1e-05,
      "loss": 0.5196,
      "mean_token_accuracy": 0.8361889123916626,
      "num_tokens": 561837097.0,
      "step": 3527
    },
    {
      "epoch": 1.7945066124109867,
      "grad_norm": 0.9365693926811218,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8510040640830994,
      "num_tokens": 561984366.0,
      "step": 3528
    },
    {
      "epoch": 1.7950152594099695,
      "grad_norm": 0.933411180973053,
      "learning_rate": 1e-05,
      "loss": 0.4774,
      "mean_token_accuracy": 0.8487666845321655,
      "num_tokens": 562152993.0,
      "step": 3529
    },
    {
      "epoch": 1.795523906408952,
      "grad_norm": 0.8894389867782593,
      "learning_rate": 1e-05,
      "loss": 0.4619,
      "mean_token_accuracy": 0.8522475957870483,
      "num_tokens": 562319171.0,
      "step": 3530
    },
    {
      "epoch": 1.7960325534079349,
      "grad_norm": 0.9173430800437927,
      "learning_rate": 1e-05,
      "loss": 0.486,
      "mean_token_accuracy": 0.8453357219696045,
      "num_tokens": 562482117.0,
      "step": 3531
    },
    {
      "epoch": 1.7965412004069177,
      "grad_norm": 0.9890989661216736,
      "learning_rate": 1e-05,
      "loss": 0.4921,
      "mean_token_accuracy": 0.8463920950889587,
      "num_tokens": 562649718.0,
      "step": 3532
    },
    {
      "epoch": 1.7970498474059005,
      "grad_norm": 0.8388110995292664,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8532881736755371,
      "num_tokens": 562813763.0,
      "step": 3533
    },
    {
      "epoch": 1.797558494404883,
      "grad_norm": 0.9548352360725403,
      "learning_rate": 1e-05,
      "loss": 0.4609,
      "mean_token_accuracy": 0.8530546426773071,
      "num_tokens": 562962769.0,
      "step": 3534
    },
    {
      "epoch": 1.7980671414038656,
      "grad_norm": 0.9055401682853699,
      "learning_rate": 1e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.8494167327880859,
      "num_tokens": 563125097.0,
      "step": 3535
    },
    {
      "epoch": 1.7985757884028484,
      "grad_norm": 0.8799210786819458,
      "learning_rate": 1e-05,
      "loss": 0.4839,
      "mean_token_accuracy": 0.8458794355392456,
      "num_tokens": 563281657.0,
      "step": 3536
    },
    {
      "epoch": 1.7990844354018312,
      "grad_norm": 1.0961902141571045,
      "learning_rate": 1e-05,
      "loss": 0.4553,
      "mean_token_accuracy": 0.8538365364074707,
      "num_tokens": 563438677.0,
      "step": 3537
    },
    {
      "epoch": 1.799593082400814,
      "grad_norm": 1.0078305006027222,
      "learning_rate": 1e-05,
      "loss": 0.4888,
      "mean_token_accuracy": 0.8463952541351318,
      "num_tokens": 563599442.0,
      "step": 3538
    },
    {
      "epoch": 1.8001017293997965,
      "grad_norm": 0.9052742123603821,
      "learning_rate": 1e-05,
      "loss": 0.4857,
      "mean_token_accuracy": 0.8461458683013916,
      "num_tokens": 563766967.0,
      "step": 3539
    },
    {
      "epoch": 1.8006103763987793,
      "grad_norm": 0.9112515449523926,
      "learning_rate": 1e-05,
      "loss": 0.4644,
      "mean_token_accuracy": 0.8512401580810547,
      "num_tokens": 563938038.0,
      "step": 3540
    },
    {
      "epoch": 1.8011190233977619,
      "grad_norm": 0.9067733883857727,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.856181263923645,
      "num_tokens": 564094217.0,
      "step": 3541
    },
    {
      "epoch": 1.8016276703967447,
      "grad_norm": 0.9368350505828857,
      "learning_rate": 1e-05,
      "loss": 0.4464,
      "mean_token_accuracy": 0.8560757040977478,
      "num_tokens": 564249675.0,
      "step": 3542
    },
    {
      "epoch": 1.8021363173957274,
      "grad_norm": 0.9412768483161926,
      "learning_rate": 1e-05,
      "loss": 0.4787,
      "mean_token_accuracy": 0.8480315208435059,
      "num_tokens": 564412107.0,
      "step": 3543
    },
    {
      "epoch": 1.8026449643947102,
      "grad_norm": 0.9945979118347168,
      "learning_rate": 1e-05,
      "loss": 0.4632,
      "mean_token_accuracy": 0.8529903888702393,
      "num_tokens": 564549442.0,
      "step": 3544
    },
    {
      "epoch": 1.8031536113936928,
      "grad_norm": 1.0442930459976196,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8548238277435303,
      "num_tokens": 564712875.0,
      "step": 3545
    },
    {
      "epoch": 1.8036622583926754,
      "grad_norm": 0.9492664337158203,
      "learning_rate": 1e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.8510094285011292,
      "num_tokens": 564879892.0,
      "step": 3546
    },
    {
      "epoch": 1.8041709053916581,
      "grad_norm": 1.9171106815338135,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8533254861831665,
      "num_tokens": 565048939.0,
      "step": 3547
    },
    {
      "epoch": 1.804679552390641,
      "grad_norm": 0.9038177132606506,
      "learning_rate": 1e-05,
      "loss": 0.48,
      "mean_token_accuracy": 0.8476209044456482,
      "num_tokens": 565210219.0,
      "step": 3548
    },
    {
      "epoch": 1.8051881993896237,
      "grad_norm": 0.9310301542282104,
      "learning_rate": 1e-05,
      "loss": 0.453,
      "mean_token_accuracy": 0.8554023504257202,
      "num_tokens": 565365789.0,
      "step": 3549
    },
    {
      "epoch": 1.8056968463886063,
      "grad_norm": 1.6419590711593628,
      "learning_rate": 1e-05,
      "loss": 0.4635,
      "mean_token_accuracy": 0.853683352470398,
      "num_tokens": 565526228.0,
      "step": 3550
    },
    {
      "epoch": 1.806205493387589,
      "grad_norm": 0.9045491814613342,
      "learning_rate": 1e-05,
      "loss": 0.4818,
      "mean_token_accuracy": 0.8482503294944763,
      "num_tokens": 565690503.0,
      "step": 3551
    },
    {
      "epoch": 1.8067141403865716,
      "grad_norm": 0.8996527194976807,
      "learning_rate": 1e-05,
      "loss": 0.497,
      "mean_token_accuracy": 0.8439865708351135,
      "num_tokens": 565852900.0,
      "step": 3552
    },
    {
      "epoch": 1.8072227873855544,
      "grad_norm": 0.8703089356422424,
      "learning_rate": 1e-05,
      "loss": 0.4649,
      "mean_token_accuracy": 0.8514032959938049,
      "num_tokens": 566016393.0,
      "step": 3553
    },
    {
      "epoch": 1.8077314343845372,
      "grad_norm": 0.9336738586425781,
      "learning_rate": 1e-05,
      "loss": 0.4574,
      "mean_token_accuracy": 0.8547062873840332,
      "num_tokens": 566180405.0,
      "step": 3554
    },
    {
      "epoch": 1.80824008138352,
      "grad_norm": 0.9066407084465027,
      "learning_rate": 1e-05,
      "loss": 0.4934,
      "mean_token_accuracy": 0.8425087928771973,
      "num_tokens": 566345044.0,
      "step": 3555
    },
    {
      "epoch": 1.8087487283825026,
      "grad_norm": 0.8604444265365601,
      "learning_rate": 1e-05,
      "loss": 0.4663,
      "mean_token_accuracy": 0.8514249324798584,
      "num_tokens": 566504302.0,
      "step": 3556
    },
    {
      "epoch": 1.8092573753814851,
      "grad_norm": 1.1371344327926636,
      "learning_rate": 1e-05,
      "loss": 0.4911,
      "mean_token_accuracy": 0.8447737097740173,
      "num_tokens": 566670105.0,
      "step": 3557
    },
    {
      "epoch": 1.809766022380468,
      "grad_norm": 0.9432363510131836,
      "learning_rate": 1e-05,
      "loss": 0.4773,
      "mean_token_accuracy": 0.8473440408706665,
      "num_tokens": 566830753.0,
      "step": 3558
    },
    {
      "epoch": 1.8102746693794507,
      "grad_norm": 0.900536835193634,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.860796332359314,
      "num_tokens": 566970910.0,
      "step": 3559
    },
    {
      "epoch": 1.8107833163784335,
      "grad_norm": 1.056198239326477,
      "learning_rate": 1e-05,
      "loss": 0.4783,
      "mean_token_accuracy": 0.8483229875564575,
      "num_tokens": 567125149.0,
      "step": 3560
    },
    {
      "epoch": 1.811291963377416,
      "grad_norm": 0.8736128807067871,
      "learning_rate": 1e-05,
      "loss": 0.5018,
      "mean_token_accuracy": 0.8416284322738647,
      "num_tokens": 567287899.0,
      "step": 3561
    },
    {
      "epoch": 1.8118006103763988,
      "grad_norm": 0.8953394889831543,
      "learning_rate": 1e-05,
      "loss": 0.448,
      "mean_token_accuracy": 0.8561390042304993,
      "num_tokens": 567439219.0,
      "step": 3562
    },
    {
      "epoch": 1.8123092573753814,
      "grad_norm": 0.9433287978172302,
      "learning_rate": 1e-05,
      "loss": 0.4534,
      "mean_token_accuracy": 0.8540156483650208,
      "num_tokens": 567603590.0,
      "step": 3563
    },
    {
      "epoch": 1.8128179043743642,
      "grad_norm": 0.9658059477806091,
      "learning_rate": 1e-05,
      "loss": 0.4791,
      "mean_token_accuracy": 0.8478438854217529,
      "num_tokens": 567757245.0,
      "step": 3564
    },
    {
      "epoch": 1.813326551373347,
      "grad_norm": 0.958234429359436,
      "learning_rate": 1e-05,
      "loss": 0.4642,
      "mean_token_accuracy": 0.8501091599464417,
      "num_tokens": 567914473.0,
      "step": 3565
    },
    {
      "epoch": 1.8138351983723298,
      "grad_norm": 1.0078924894332886,
      "learning_rate": 1e-05,
      "loss": 0.4494,
      "mean_token_accuracy": 0.8547240495681763,
      "num_tokens": 568063203.0,
      "step": 3566
    },
    {
      "epoch": 1.8143438453713123,
      "grad_norm": 0.9522838592529297,
      "learning_rate": 1e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.8510032296180725,
      "num_tokens": 568219189.0,
      "step": 3567
    },
    {
      "epoch": 1.814852492370295,
      "grad_norm": 0.9724707007408142,
      "learning_rate": 1e-05,
      "loss": 0.477,
      "mean_token_accuracy": 0.8469789028167725,
      "num_tokens": 568375485.0,
      "step": 3568
    },
    {
      "epoch": 1.8153611393692777,
      "grad_norm": 0.9488127827644348,
      "learning_rate": 1e-05,
      "loss": 0.5014,
      "mean_token_accuracy": 0.8410286903381348,
      "num_tokens": 568532599.0,
      "step": 3569
    },
    {
      "epoch": 1.8158697863682605,
      "grad_norm": 0.8995364904403687,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.8555278182029724,
      "num_tokens": 568704776.0,
      "step": 3570
    },
    {
      "epoch": 1.8163784333672433,
      "grad_norm": 0.9317036867141724,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.855113685131073,
      "num_tokens": 568876940.0,
      "step": 3571
    },
    {
      "epoch": 1.8168870803662258,
      "grad_norm": 1.001330018043518,
      "learning_rate": 1e-05,
      "loss": 0.4664,
      "mean_token_accuracy": 0.8509033918380737,
      "num_tokens": 569029468.0,
      "step": 3572
    },
    {
      "epoch": 1.8173957273652084,
      "grad_norm": 0.9407387375831604,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.8527392148971558,
      "num_tokens": 569180360.0,
      "step": 3573
    },
    {
      "epoch": 1.8179043743641912,
      "grad_norm": 0.9812650084495544,
      "learning_rate": 1e-05,
      "loss": 0.4849,
      "mean_token_accuracy": 0.8459858894348145,
      "num_tokens": 569332534.0,
      "step": 3574
    },
    {
      "epoch": 1.818413021363174,
      "grad_norm": 0.9319610595703125,
      "learning_rate": 1e-05,
      "loss": 0.4346,
      "mean_token_accuracy": 0.8603183031082153,
      "num_tokens": 569494821.0,
      "step": 3575
    },
    {
      "epoch": 1.8189216683621567,
      "grad_norm": 0.8834575414657593,
      "learning_rate": 1e-05,
      "loss": 0.4783,
      "mean_token_accuracy": 0.8467259407043457,
      "num_tokens": 569647573.0,
      "step": 3576
    },
    {
      "epoch": 1.8194303153611395,
      "grad_norm": 0.9952992796897888,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8493619561195374,
      "num_tokens": 569797825.0,
      "step": 3577
    },
    {
      "epoch": 1.819938962360122,
      "grad_norm": 0.947803795337677,
      "learning_rate": 1e-05,
      "loss": 0.4603,
      "mean_token_accuracy": 0.8515059351921082,
      "num_tokens": 569960390.0,
      "step": 3578
    },
    {
      "epoch": 1.8204476093591047,
      "grad_norm": 0.9164172410964966,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8544726371765137,
      "num_tokens": 570118035.0,
      "step": 3579
    },
    {
      "epoch": 1.8209562563580874,
      "grad_norm": 0.9546960592269897,
      "learning_rate": 1e-05,
      "loss": 0.4986,
      "mean_token_accuracy": 0.8426315784454346,
      "num_tokens": 570276992.0,
      "step": 3580
    },
    {
      "epoch": 1.8214649033570702,
      "grad_norm": 0.9669478535652161,
      "learning_rate": 1e-05,
      "loss": 0.4513,
      "mean_token_accuracy": 0.8550679087638855,
      "num_tokens": 570437162.0,
      "step": 3581
    },
    {
      "epoch": 1.821973550356053,
      "grad_norm": 0.8663604259490967,
      "learning_rate": 1e-05,
      "loss": 0.4708,
      "mean_token_accuracy": 0.85112065076828,
      "num_tokens": 570601038.0,
      "step": 3582
    },
    {
      "epoch": 1.8224821973550356,
      "grad_norm": 1.0352712869644165,
      "learning_rate": 1e-05,
      "loss": 0.4778,
      "mean_token_accuracy": 0.8478525876998901,
      "num_tokens": 570764320.0,
      "step": 3583
    },
    {
      "epoch": 1.8229908443540181,
      "grad_norm": 0.8922290205955505,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8514959812164307,
      "num_tokens": 570927501.0,
      "step": 3584
    },
    {
      "epoch": 1.823499491353001,
      "grad_norm": 0.937830924987793,
      "learning_rate": 1e-05,
      "loss": 0.4735,
      "mean_token_accuracy": 0.8487762808799744,
      "num_tokens": 571072594.0,
      "step": 3585
    },
    {
      "epoch": 1.8240081383519837,
      "grad_norm": 0.9236015677452087,
      "learning_rate": 1e-05,
      "loss": 0.47,
      "mean_token_accuracy": 0.849860429763794,
      "num_tokens": 571225760.0,
      "step": 3586
    },
    {
      "epoch": 1.8245167853509665,
      "grad_norm": 0.9017430543899536,
      "learning_rate": 1e-05,
      "loss": 0.4913,
      "mean_token_accuracy": 0.8431181311607361,
      "num_tokens": 571383451.0,
      "step": 3587
    },
    {
      "epoch": 1.8250254323499493,
      "grad_norm": 0.9215783476829529,
      "learning_rate": 1e-05,
      "loss": 0.4537,
      "mean_token_accuracy": 0.8537969589233398,
      "num_tokens": 571536763.0,
      "step": 3588
    },
    {
      "epoch": 1.8255340793489319,
      "grad_norm": 0.9324192404747009,
      "learning_rate": 1e-05,
      "loss": 0.4644,
      "mean_token_accuracy": 0.8512316942214966,
      "num_tokens": 571709893.0,
      "step": 3589
    },
    {
      "epoch": 1.8260427263479144,
      "grad_norm": 0.9075586199760437,
      "learning_rate": 1e-05,
      "loss": 0.4368,
      "mean_token_accuracy": 0.8588082790374756,
      "num_tokens": 571865425.0,
      "step": 3590
    },
    {
      "epoch": 1.8265513733468972,
      "grad_norm": 0.926460325717926,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8537696599960327,
      "num_tokens": 572024420.0,
      "step": 3591
    },
    {
      "epoch": 1.82706002034588,
      "grad_norm": 0.9003691673278809,
      "learning_rate": 1e-05,
      "loss": 0.478,
      "mean_token_accuracy": 0.8481897711753845,
      "num_tokens": 572184872.0,
      "step": 3592
    },
    {
      "epoch": 1.8275686673448628,
      "grad_norm": 0.9034071564674377,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.8484635353088379,
      "num_tokens": 572353776.0,
      "step": 3593
    },
    {
      "epoch": 1.8280773143438453,
      "grad_norm": 0.8958525061607361,
      "learning_rate": 1e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.8530550003051758,
      "num_tokens": 572506248.0,
      "step": 3594
    },
    {
      "epoch": 1.828585961342828,
      "grad_norm": 0.9878848195075989,
      "learning_rate": 1e-05,
      "loss": 0.5038,
      "mean_token_accuracy": 0.8405113220214844,
      "num_tokens": 572672814.0,
      "step": 3595
    },
    {
      "epoch": 1.8290946083418107,
      "grad_norm": 1.085959553718567,
      "learning_rate": 1e-05,
      "loss": 0.4639,
      "mean_token_accuracy": 0.8505579233169556,
      "num_tokens": 572820670.0,
      "step": 3596
    },
    {
      "epoch": 1.8296032553407935,
      "grad_norm": 0.9331508278846741,
      "learning_rate": 1e-05,
      "loss": 0.4324,
      "mean_token_accuracy": 0.8600931167602539,
      "num_tokens": 572974866.0,
      "step": 3597
    },
    {
      "epoch": 1.8301119023397763,
      "grad_norm": 0.917675793170929,
      "learning_rate": 1e-05,
      "loss": 0.4578,
      "mean_token_accuracy": 0.8545408248901367,
      "num_tokens": 573129521.0,
      "step": 3598
    },
    {
      "epoch": 1.830620549338759,
      "grad_norm": 3.0825843811035156,
      "learning_rate": 1e-05,
      "loss": 0.4918,
      "mean_token_accuracy": 0.8449922204017639,
      "num_tokens": 573296454.0,
      "step": 3599
    },
    {
      "epoch": 1.8311291963377416,
      "grad_norm": 0.9181584715843201,
      "learning_rate": 1e-05,
      "loss": 0.4642,
      "mean_token_accuracy": 0.8511255979537964,
      "num_tokens": 573458787.0,
      "step": 3600
    },
    {
      "epoch": 1.8316378433367242,
      "grad_norm": 0.8957834839820862,
      "learning_rate": 1e-05,
      "loss": 0.495,
      "mean_token_accuracy": 0.8457474708557129,
      "num_tokens": 573619962.0,
      "step": 3601
    },
    {
      "epoch": 1.832146490335707,
      "grad_norm": 0.9297384023666382,
      "learning_rate": 1e-05,
      "loss": 0.4773,
      "mean_token_accuracy": 0.8491976261138916,
      "num_tokens": 573794196.0,
      "step": 3602
    },
    {
      "epoch": 1.8326551373346898,
      "grad_norm": 0.906842052936554,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.8586156368255615,
      "num_tokens": 573952177.0,
      "step": 3603
    },
    {
      "epoch": 1.8331637843336726,
      "grad_norm": 0.9280809164047241,
      "learning_rate": 1e-05,
      "loss": 0.4908,
      "mean_token_accuracy": 0.8437187671661377,
      "num_tokens": 574118594.0,
      "step": 3604
    },
    {
      "epoch": 1.8336724313326551,
      "grad_norm": 0.9042379856109619,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.850511372089386,
      "num_tokens": 574274813.0,
      "step": 3605
    },
    {
      "epoch": 1.8341810783316377,
      "grad_norm": 1.0138126611709595,
      "learning_rate": 1e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8595215082168579,
      "num_tokens": 574429910.0,
      "step": 3606
    },
    {
      "epoch": 1.8346897253306205,
      "grad_norm": 0.9278191924095154,
      "learning_rate": 1e-05,
      "loss": 0.4736,
      "mean_token_accuracy": 0.8506954908370972,
      "num_tokens": 574598948.0,
      "step": 3607
    },
    {
      "epoch": 1.8351983723296033,
      "grad_norm": 1.0183500051498413,
      "learning_rate": 1e-05,
      "loss": 0.4376,
      "mean_token_accuracy": 0.8583874702453613,
      "num_tokens": 574750525.0,
      "step": 3608
    },
    {
      "epoch": 1.835707019328586,
      "grad_norm": 0.9347324967384338,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8530687689781189,
      "num_tokens": 574920331.0,
      "step": 3609
    },
    {
      "epoch": 1.8362156663275688,
      "grad_norm": 1.1582096815109253,
      "learning_rate": 1e-05,
      "loss": 0.4834,
      "mean_token_accuracy": 0.8469882011413574,
      "num_tokens": 575086614.0,
      "step": 3610
    },
    {
      "epoch": 1.8367243133265514,
      "grad_norm": 1.0273844003677368,
      "learning_rate": 1e-05,
      "loss": 0.4602,
      "mean_token_accuracy": 0.8511192202568054,
      "num_tokens": 575243072.0,
      "step": 3611
    },
    {
      "epoch": 1.837232960325534,
      "grad_norm": 0.8761646151542664,
      "learning_rate": 1e-05,
      "loss": 0.445,
      "mean_token_accuracy": 0.8564286828041077,
      "num_tokens": 575407175.0,
      "step": 3612
    },
    {
      "epoch": 1.8377416073245167,
      "grad_norm": 0.9320701956748962,
      "learning_rate": 1e-05,
      "loss": 0.454,
      "mean_token_accuracy": 0.8560305237770081,
      "num_tokens": 575560729.0,
      "step": 3613
    },
    {
      "epoch": 1.8382502543234995,
      "grad_norm": 0.9284968376159668,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8568035364151001,
      "num_tokens": 575734331.0,
      "step": 3614
    },
    {
      "epoch": 1.8387589013224823,
      "grad_norm": 0.9640108346939087,
      "learning_rate": 1e-05,
      "loss": 0.4764,
      "mean_token_accuracy": 0.8468776345252991,
      "num_tokens": 575903732.0,
      "step": 3615
    },
    {
      "epoch": 1.8392675483214649,
      "grad_norm": 0.9263799786567688,
      "learning_rate": 1e-05,
      "loss": 0.4829,
      "mean_token_accuracy": 0.8463630676269531,
      "num_tokens": 576048212.0,
      "step": 3616
    },
    {
      "epoch": 1.8397761953204474,
      "grad_norm": 0.9382118582725525,
      "learning_rate": 1e-05,
      "loss": 0.4596,
      "mean_token_accuracy": 0.8531448841094971,
      "num_tokens": 576207321.0,
      "step": 3617
    },
    {
      "epoch": 1.8402848423194302,
      "grad_norm": 1.1217350959777832,
      "learning_rate": 1e-05,
      "loss": 0.5083,
      "mean_token_accuracy": 0.840814471244812,
      "num_tokens": 576379541.0,
      "step": 3618
    },
    {
      "epoch": 1.840793489318413,
      "grad_norm": 0.875610888004303,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8584346771240234,
      "num_tokens": 576542203.0,
      "step": 3619
    },
    {
      "epoch": 1.8413021363173958,
      "grad_norm": 0.9747274518013,
      "learning_rate": 1e-05,
      "loss": 0.4954,
      "mean_token_accuracy": 0.8422902822494507,
      "num_tokens": 576693938.0,
      "step": 3620
    },
    {
      "epoch": 1.8418107833163786,
      "grad_norm": 1.030494213104248,
      "learning_rate": 1e-05,
      "loss": 0.4637,
      "mean_token_accuracy": 0.8514543771743774,
      "num_tokens": 576848972.0,
      "step": 3621
    },
    {
      "epoch": 1.8423194303153612,
      "grad_norm": 0.9475911259651184,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.851582407951355,
      "num_tokens": 576996432.0,
      "step": 3622
    },
    {
      "epoch": 1.8428280773143437,
      "grad_norm": 0.9580101370811462,
      "learning_rate": 1e-05,
      "loss": 0.5064,
      "mean_token_accuracy": 0.8397974967956543,
      "num_tokens": 577165777.0,
      "step": 3623
    },
    {
      "epoch": 1.8433367243133265,
      "grad_norm": 0.9337804913520813,
      "learning_rate": 1e-05,
      "loss": 0.4639,
      "mean_token_accuracy": 0.8514394760131836,
      "num_tokens": 577320221.0,
      "step": 3624
    },
    {
      "epoch": 1.8438453713123093,
      "grad_norm": 0.8653109669685364,
      "learning_rate": 1e-05,
      "loss": 0.4743,
      "mean_token_accuracy": 0.8486281633377075,
      "num_tokens": 577482320.0,
      "step": 3625
    },
    {
      "epoch": 1.844354018311292,
      "grad_norm": 2.005347967147827,
      "learning_rate": 1e-05,
      "loss": 0.4726,
      "mean_token_accuracy": 0.8509484529495239,
      "num_tokens": 577643317.0,
      "step": 3626
    },
    {
      "epoch": 1.8448626653102747,
      "grad_norm": 1.2213129997253418,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.8512935638427734,
      "num_tokens": 577797333.0,
      "step": 3627
    },
    {
      "epoch": 1.8453713123092572,
      "grad_norm": 0.9152913689613342,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.852689802646637,
      "num_tokens": 577962353.0,
      "step": 3628
    },
    {
      "epoch": 1.84587995930824,
      "grad_norm": 0.9832363128662109,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.849490225315094,
      "num_tokens": 578124756.0,
      "step": 3629
    },
    {
      "epoch": 1.8463886063072228,
      "grad_norm": 0.9140488505363464,
      "learning_rate": 1e-05,
      "loss": 0.4898,
      "mean_token_accuracy": 0.8460453748703003,
      "num_tokens": 578280116.0,
      "step": 3630
    },
    {
      "epoch": 1.8468972533062056,
      "grad_norm": 1.0018895864486694,
      "learning_rate": 1e-05,
      "loss": 0.4624,
      "mean_token_accuracy": 0.8521567583084106,
      "num_tokens": 578428148.0,
      "step": 3631
    },
    {
      "epoch": 1.8474059003051884,
      "grad_norm": 0.8730851411819458,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.8521086573600769,
      "num_tokens": 578587724.0,
      "step": 3632
    },
    {
      "epoch": 1.847914547304171,
      "grad_norm": 0.9751337766647339,
      "learning_rate": 1e-05,
      "loss": 0.4795,
      "mean_token_accuracy": 0.8483328223228455,
      "num_tokens": 578744811.0,
      "step": 3633
    },
    {
      "epoch": 1.8484231943031535,
      "grad_norm": 0.952669084072113,
      "learning_rate": 1e-05,
      "loss": 0.4775,
      "mean_token_accuracy": 0.8489563465118408,
      "num_tokens": 578907271.0,
      "step": 3634
    },
    {
      "epoch": 1.8489318413021363,
      "grad_norm": 0.9171006083488464,
      "learning_rate": 1e-05,
      "loss": 0.4773,
      "mean_token_accuracy": 0.8486084342002869,
      "num_tokens": 579067155.0,
      "step": 3635
    },
    {
      "epoch": 1.849440488301119,
      "grad_norm": 1.0136206150054932,
      "learning_rate": 1e-05,
      "loss": 0.4782,
      "mean_token_accuracy": 0.8471975326538086,
      "num_tokens": 579229124.0,
      "step": 3636
    },
    {
      "epoch": 1.8499491353001019,
      "grad_norm": 0.9842079877853394,
      "learning_rate": 1e-05,
      "loss": 0.4558,
      "mean_token_accuracy": 0.8547145128250122,
      "num_tokens": 579380265.0,
      "step": 3637
    },
    {
      "epoch": 1.8504577822990844,
      "grad_norm": 0.8879310488700867,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8601967096328735,
      "num_tokens": 579541087.0,
      "step": 3638
    },
    {
      "epoch": 1.850966429298067,
      "grad_norm": 0.9165470004081726,
      "learning_rate": 1e-05,
      "loss": 0.4499,
      "mean_token_accuracy": 0.8543002009391785,
      "num_tokens": 579691325.0,
      "step": 3639
    },
    {
      "epoch": 1.8514750762970498,
      "grad_norm": 1.1003382205963135,
      "learning_rate": 1e-05,
      "loss": 0.5004,
      "mean_token_accuracy": 0.8400880098342896,
      "num_tokens": 579853781.0,
      "step": 3640
    },
    {
      "epoch": 1.8519837232960326,
      "grad_norm": 0.9798530340194702,
      "learning_rate": 1e-05,
      "loss": 0.4806,
      "mean_token_accuracy": 0.8484410047531128,
      "num_tokens": 580011285.0,
      "step": 3641
    },
    {
      "epoch": 1.8524923702950153,
      "grad_norm": 0.8885270357131958,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8579368591308594,
      "num_tokens": 580175507.0,
      "step": 3642
    },
    {
      "epoch": 1.8530010172939981,
      "grad_norm": 0.8953379988670349,
      "learning_rate": 1e-05,
      "loss": 0.4748,
      "mean_token_accuracy": 0.8487176895141602,
      "num_tokens": 580334727.0,
      "step": 3643
    },
    {
      "epoch": 1.8535096642929807,
      "grad_norm": 1.237749457359314,
      "learning_rate": 1e-05,
      "loss": 0.4699,
      "mean_token_accuracy": 0.8505109548568726,
      "num_tokens": 580499293.0,
      "step": 3644
    },
    {
      "epoch": 1.8540183112919633,
      "grad_norm": 1.0147404670715332,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "mean_token_accuracy": 0.8560020327568054,
      "num_tokens": 580643374.0,
      "step": 3645
    },
    {
      "epoch": 1.854526958290946,
      "grad_norm": 0.908033549785614,
      "learning_rate": 1e-05,
      "loss": 0.501,
      "mean_token_accuracy": 0.8429970741271973,
      "num_tokens": 580805832.0,
      "step": 3646
    },
    {
      "epoch": 1.8550356052899288,
      "grad_norm": 0.9476537704467773,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "mean_token_accuracy": 0.8594100475311279,
      "num_tokens": 580956923.0,
      "step": 3647
    },
    {
      "epoch": 1.8555442522889116,
      "grad_norm": 0.925979495048523,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8583248853683472,
      "num_tokens": 581107989.0,
      "step": 3648
    },
    {
      "epoch": 1.8560528992878942,
      "grad_norm": 0.9476032257080078,
      "learning_rate": 1e-05,
      "loss": 0.4507,
      "mean_token_accuracy": 0.8557018041610718,
      "num_tokens": 581257593.0,
      "step": 3649
    },
    {
      "epoch": 1.8565615462868768,
      "grad_norm": 0.9103021621704102,
      "learning_rate": 1e-05,
      "loss": 0.4687,
      "mean_token_accuracy": 0.8508168458938599,
      "num_tokens": 581423977.0,
      "step": 3650
    },
    {
      "epoch": 1.8570701932858595,
      "grad_norm": 0.8881806135177612,
      "learning_rate": 1e-05,
      "loss": 0.465,
      "mean_token_accuracy": 0.8523017764091492,
      "num_tokens": 581587347.0,
      "step": 3651
    },
    {
      "epoch": 1.8575788402848423,
      "grad_norm": 1.0127854347229004,
      "learning_rate": 1e-05,
      "loss": 0.5131,
      "mean_token_accuracy": 0.8396639823913574,
      "num_tokens": 581742263.0,
      "step": 3652
    },
    {
      "epoch": 1.8580874872838251,
      "grad_norm": 0.9147594571113586,
      "learning_rate": 1e-05,
      "loss": 0.4786,
      "mean_token_accuracy": 0.8464944958686829,
      "num_tokens": 581907547.0,
      "step": 3653
    },
    {
      "epoch": 1.8585961342828077,
      "grad_norm": 0.919609785079956,
      "learning_rate": 1e-05,
      "loss": 0.4675,
      "mean_token_accuracy": 0.850050151348114,
      "num_tokens": 582060741.0,
      "step": 3654
    },
    {
      "epoch": 1.8591047812817905,
      "grad_norm": 0.9043321013450623,
      "learning_rate": 1e-05,
      "loss": 0.4559,
      "mean_token_accuracy": 0.8531760573387146,
      "num_tokens": 582214993.0,
      "step": 3655
    },
    {
      "epoch": 1.859613428280773,
      "grad_norm": 0.8717383146286011,
      "learning_rate": 1e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.8501156568527222,
      "num_tokens": 582383233.0,
      "step": 3656
    },
    {
      "epoch": 1.8601220752797558,
      "grad_norm": 0.8819180130958557,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.8476206064224243,
      "num_tokens": 582545163.0,
      "step": 3657
    },
    {
      "epoch": 1.8606307222787386,
      "grad_norm": 0.989859938621521,
      "learning_rate": 1e-05,
      "loss": 0.5048,
      "mean_token_accuracy": 0.8376556634902954,
      "num_tokens": 582700928.0,
      "step": 3658
    },
    {
      "epoch": 1.8611393692777214,
      "grad_norm": 0.8996127247810364,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8479574918746948,
      "num_tokens": 582871470.0,
      "step": 3659
    },
    {
      "epoch": 1.861648016276704,
      "grad_norm": 0.8768028616905212,
      "learning_rate": 1e-05,
      "loss": 0.4492,
      "mean_token_accuracy": 0.8561135530471802,
      "num_tokens": 583019968.0,
      "step": 3660
    },
    {
      "epoch": 1.8621566632756865,
      "grad_norm": 0.9248018264770508,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8482380509376526,
      "num_tokens": 583166808.0,
      "step": 3661
    },
    {
      "epoch": 1.8626653102746693,
      "grad_norm": 0.9277891516685486,
      "learning_rate": 1e-05,
      "loss": 0.4679,
      "mean_token_accuracy": 0.8508737087249756,
      "num_tokens": 583318374.0,
      "step": 3662
    },
    {
      "epoch": 1.863173957273652,
      "grad_norm": 0.9190016388893127,
      "learning_rate": 1e-05,
      "loss": 0.4715,
      "mean_token_accuracy": 0.8488146662712097,
      "num_tokens": 583484043.0,
      "step": 3663
    },
    {
      "epoch": 1.8636826042726349,
      "grad_norm": 1.0272608995437622,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.857718825340271,
      "num_tokens": 583646646.0,
      "step": 3664
    },
    {
      "epoch": 1.8641912512716174,
      "grad_norm": 0.8706140518188477,
      "learning_rate": 1e-05,
      "loss": 0.4779,
      "mean_token_accuracy": 0.847746729850769,
      "num_tokens": 583808674.0,
      "step": 3665
    },
    {
      "epoch": 1.8646998982706002,
      "grad_norm": 1.0054000616073608,
      "learning_rate": 1e-05,
      "loss": 0.4471,
      "mean_token_accuracy": 0.8554672002792358,
      "num_tokens": 583961908.0,
      "step": 3666
    },
    {
      "epoch": 1.8652085452695828,
      "grad_norm": 0.9957459568977356,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8503569960594177,
      "num_tokens": 584121007.0,
      "step": 3667
    },
    {
      "epoch": 1.8657171922685656,
      "grad_norm": 0.9274004101753235,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.8552700877189636,
      "num_tokens": 584274112.0,
      "step": 3668
    },
    {
      "epoch": 1.8662258392675484,
      "grad_norm": 1.0314558744430542,
      "learning_rate": 1e-05,
      "loss": 0.4322,
      "mean_token_accuracy": 0.8610309362411499,
      "num_tokens": 584437482.0,
      "step": 3669
    },
    {
      "epoch": 1.8667344862665312,
      "grad_norm": 0.9328179955482483,
      "learning_rate": 1e-05,
      "loss": 0.4557,
      "mean_token_accuracy": 0.8532506227493286,
      "num_tokens": 584604753.0,
      "step": 3670
    },
    {
      "epoch": 1.8672431332655137,
      "grad_norm": 0.9097633361816406,
      "learning_rate": 1e-05,
      "loss": 0.4708,
      "mean_token_accuracy": 0.8506937026977539,
      "num_tokens": 584768863.0,
      "step": 3671
    },
    {
      "epoch": 1.8677517802644963,
      "grad_norm": 0.9894227385520935,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8492318391799927,
      "num_tokens": 584922779.0,
      "step": 3672
    },
    {
      "epoch": 1.868260427263479,
      "grad_norm": 1.0254460573196411,
      "learning_rate": 1e-05,
      "loss": 0.475,
      "mean_token_accuracy": 0.848928689956665,
      "num_tokens": 585082703.0,
      "step": 3673
    },
    {
      "epoch": 1.8687690742624619,
      "grad_norm": 0.9415755271911621,
      "learning_rate": 1e-05,
      "loss": 0.5065,
      "mean_token_accuracy": 0.8420661687850952,
      "num_tokens": 585258146.0,
      "step": 3674
    },
    {
      "epoch": 1.8692777212614446,
      "grad_norm": 0.9524638652801514,
      "learning_rate": 1e-05,
      "loss": 0.4575,
      "mean_token_accuracy": 0.8552057147026062,
      "num_tokens": 585416065.0,
      "step": 3675
    },
    {
      "epoch": 1.8697863682604272,
      "grad_norm": 0.9377021789550781,
      "learning_rate": 1e-05,
      "loss": 0.4464,
      "mean_token_accuracy": 0.8568128347396851,
      "num_tokens": 585575451.0,
      "step": 3676
    },
    {
      "epoch": 1.87029501525941,
      "grad_norm": 0.9163416028022766,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8492084741592407,
      "num_tokens": 585739539.0,
      "step": 3677
    },
    {
      "epoch": 1.8708036622583926,
      "grad_norm": 0.959039568901062,
      "learning_rate": 1e-05,
      "loss": 0.4717,
      "mean_token_accuracy": 0.8480402231216431,
      "num_tokens": 585907402.0,
      "step": 3678
    },
    {
      "epoch": 1.8713123092573754,
      "grad_norm": 0.8940199613571167,
      "learning_rate": 1e-05,
      "loss": 0.4905,
      "mean_token_accuracy": 0.8439282178878784,
      "num_tokens": 586083334.0,
      "step": 3679
    },
    {
      "epoch": 1.8718209562563581,
      "grad_norm": 0.9631999135017395,
      "learning_rate": 1e-05,
      "loss": 0.4664,
      "mean_token_accuracy": 0.8514367341995239,
      "num_tokens": 586248102.0,
      "step": 3680
    },
    {
      "epoch": 1.872329603255341,
      "grad_norm": 1.0510187149047852,
      "learning_rate": 1e-05,
      "loss": 0.498,
      "mean_token_accuracy": 0.8436463475227356,
      "num_tokens": 586402076.0,
      "step": 3681
    },
    {
      "epoch": 1.8728382502543235,
      "grad_norm": 0.9832147359848022,
      "learning_rate": 1e-05,
      "loss": 0.4784,
      "mean_token_accuracy": 0.847326934337616,
      "num_tokens": 586569377.0,
      "step": 3682
    },
    {
      "epoch": 1.873346897253306,
      "grad_norm": 0.9283510446548462,
      "learning_rate": 1e-05,
      "loss": 0.4705,
      "mean_token_accuracy": 0.8506478071212769,
      "num_tokens": 586729634.0,
      "step": 3683
    },
    {
      "epoch": 1.8738555442522888,
      "grad_norm": 0.9281595349311829,
      "learning_rate": 1e-05,
      "loss": 0.4558,
      "mean_token_accuracy": 0.8545751571655273,
      "num_tokens": 586891543.0,
      "step": 3684
    },
    {
      "epoch": 1.8743641912512716,
      "grad_norm": 1.007712960243225,
      "learning_rate": 1e-05,
      "loss": 0.4496,
      "mean_token_accuracy": 0.8549709320068359,
      "num_tokens": 587053469.0,
      "step": 3685
    },
    {
      "epoch": 1.8748728382502544,
      "grad_norm": 0.9378703236579895,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8443782329559326,
      "num_tokens": 587214972.0,
      "step": 3686
    },
    {
      "epoch": 1.875381485249237,
      "grad_norm": 0.9779534339904785,
      "learning_rate": 1e-05,
      "loss": 0.5014,
      "mean_token_accuracy": 0.8421754240989685,
      "num_tokens": 587380598.0,
      "step": 3687
    },
    {
      "epoch": 1.8758901322482198,
      "grad_norm": 0.9017239809036255,
      "learning_rate": 1e-05,
      "loss": 0.4774,
      "mean_token_accuracy": 0.8475435972213745,
      "num_tokens": 587541589.0,
      "step": 3688
    },
    {
      "epoch": 1.8763987792472023,
      "grad_norm": 0.9856732487678528,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8449099659919739,
      "num_tokens": 587704662.0,
      "step": 3689
    },
    {
      "epoch": 1.8769074262461851,
      "grad_norm": 0.9052358865737915,
      "learning_rate": 1e-05,
      "loss": 0.4933,
      "mean_token_accuracy": 0.8428623080253601,
      "num_tokens": 587860976.0,
      "step": 3690
    },
    {
      "epoch": 1.877416073245168,
      "grad_norm": 0.9494747519493103,
      "learning_rate": 1e-05,
      "loss": 0.5027,
      "mean_token_accuracy": 0.8417651653289795,
      "num_tokens": 588018853.0,
      "step": 3691
    },
    {
      "epoch": 1.8779247202441507,
      "grad_norm": 0.9226229786872864,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "mean_token_accuracy": 0.8583682775497437,
      "num_tokens": 588168649.0,
      "step": 3692
    },
    {
      "epoch": 1.8784333672431333,
      "grad_norm": 0.8945475816726685,
      "learning_rate": 1e-05,
      "loss": 0.4372,
      "mean_token_accuracy": 0.8595737218856812,
      "num_tokens": 588328963.0,
      "step": 3693
    },
    {
      "epoch": 1.8789420142421158,
      "grad_norm": 0.9041112065315247,
      "learning_rate": 1e-05,
      "loss": 0.4589,
      "mean_token_accuracy": 0.8514023423194885,
      "num_tokens": 588491275.0,
      "step": 3694
    },
    {
      "epoch": 1.8794506612410986,
      "grad_norm": 0.8546394109725952,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8534486889839172,
      "num_tokens": 588654078.0,
      "step": 3695
    },
    {
      "epoch": 1.8799593082400814,
      "grad_norm": 0.8748382925987244,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8489651679992676,
      "num_tokens": 588817168.0,
      "step": 3696
    },
    {
      "epoch": 1.8804679552390642,
      "grad_norm": 0.9193595051765442,
      "learning_rate": 1e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.8599328994750977,
      "num_tokens": 588976724.0,
      "step": 3697
    },
    {
      "epoch": 1.8809766022380467,
      "grad_norm": 0.8867758512496948,
      "learning_rate": 1e-05,
      "loss": 0.4854,
      "mean_token_accuracy": 0.8471266627311707,
      "num_tokens": 589135465.0,
      "step": 3698
    },
    {
      "epoch": 1.8814852492370295,
      "grad_norm": 0.9284715056419373,
      "learning_rate": 1e-05,
      "loss": 0.4877,
      "mean_token_accuracy": 0.8443389534950256,
      "num_tokens": 589304160.0,
      "step": 3699
    },
    {
      "epoch": 1.881993896236012,
      "grad_norm": 0.9017884731292725,
      "learning_rate": 1e-05,
      "loss": 0.4997,
      "mean_token_accuracy": 0.8431593179702759,
      "num_tokens": 589459066.0,
      "step": 3700
    },
    {
      "epoch": 1.8825025432349949,
      "grad_norm": 0.8958366513252258,
      "learning_rate": 1e-05,
      "loss": 0.4609,
      "mean_token_accuracy": 0.8517111539840698,
      "num_tokens": 589634237.0,
      "step": 3701
    },
    {
      "epoch": 1.8830111902339777,
      "grad_norm": 0.9634583592414856,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8483635187149048,
      "num_tokens": 589787765.0,
      "step": 3702
    },
    {
      "epoch": 1.8835198372329605,
      "grad_norm": 0.8931185007095337,
      "learning_rate": 1e-05,
      "loss": 0.4692,
      "mean_token_accuracy": 0.8505688905715942,
      "num_tokens": 589955052.0,
      "step": 3703
    },
    {
      "epoch": 1.884028484231943,
      "grad_norm": 0.9743666648864746,
      "learning_rate": 1e-05,
      "loss": 0.466,
      "mean_token_accuracy": 0.8499509692192078,
      "num_tokens": 590119500.0,
      "step": 3704
    },
    {
      "epoch": 1.8845371312309256,
      "grad_norm": 0.8730661273002625,
      "learning_rate": 1e-05,
      "loss": 0.4745,
      "mean_token_accuracy": 0.849245011806488,
      "num_tokens": 590274290.0,
      "step": 3705
    },
    {
      "epoch": 1.8850457782299084,
      "grad_norm": 1.013344168663025,
      "learning_rate": 1e-05,
      "loss": 0.4978,
      "mean_token_accuracy": 0.8414533138275146,
      "num_tokens": 590427085.0,
      "step": 3706
    },
    {
      "epoch": 1.8855544252288912,
      "grad_norm": 0.902656614780426,
      "learning_rate": 1e-05,
      "loss": 0.4469,
      "mean_token_accuracy": 0.8566779494285583,
      "num_tokens": 590586848.0,
      "step": 3707
    },
    {
      "epoch": 1.886063072227874,
      "grad_norm": 0.9930974841117859,
      "learning_rate": 1e-05,
      "loss": 0.4867,
      "mean_token_accuracy": 0.8459283113479614,
      "num_tokens": 590750486.0,
      "step": 3708
    },
    {
      "epoch": 1.8865717192268565,
      "grad_norm": 0.9296849966049194,
      "learning_rate": 1e-05,
      "loss": 0.469,
      "mean_token_accuracy": 0.8507835268974304,
      "num_tokens": 590909813.0,
      "step": 3709
    },
    {
      "epoch": 1.8870803662258393,
      "grad_norm": 1.0177860260009766,
      "learning_rate": 1e-05,
      "loss": 0.4679,
      "mean_token_accuracy": 0.8514139652252197,
      "num_tokens": 591063252.0,
      "step": 3710
    },
    {
      "epoch": 1.8875890132248219,
      "grad_norm": 0.9661931991577148,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.853483259677887,
      "num_tokens": 591212843.0,
      "step": 3711
    },
    {
      "epoch": 1.8880976602238047,
      "grad_norm": 0.9466366171836853,
      "learning_rate": 1e-05,
      "loss": 0.4627,
      "mean_token_accuracy": 0.8509161472320557,
      "num_tokens": 591367281.0,
      "step": 3712
    },
    {
      "epoch": 1.8886063072227874,
      "grad_norm": 0.9832890033721924,
      "learning_rate": 1e-05,
      "loss": 0.4941,
      "mean_token_accuracy": 0.8426257371902466,
      "num_tokens": 591533903.0,
      "step": 3713
    },
    {
      "epoch": 1.8891149542217702,
      "grad_norm": 0.9140143990516663,
      "learning_rate": 1e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8483765721321106,
      "num_tokens": 591689473.0,
      "step": 3714
    },
    {
      "epoch": 1.8896236012207528,
      "grad_norm": 0.9094839096069336,
      "learning_rate": 1e-05,
      "loss": 0.4704,
      "mean_token_accuracy": 0.848656177520752,
      "num_tokens": 591853239.0,
      "step": 3715
    },
    {
      "epoch": 1.8901322482197354,
      "grad_norm": 0.9416212439537048,
      "learning_rate": 1e-05,
      "loss": 0.4584,
      "mean_token_accuracy": 0.8517597913742065,
      "num_tokens": 592022697.0,
      "step": 3716
    },
    {
      "epoch": 1.8906408952187181,
      "grad_norm": 0.9125696420669556,
      "learning_rate": 1e-05,
      "loss": 0.5124,
      "mean_token_accuracy": 0.8381770849227905,
      "num_tokens": 592183835.0,
      "step": 3717
    },
    {
      "epoch": 1.891149542217701,
      "grad_norm": 0.8517329096794128,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8600273728370667,
      "num_tokens": 592341817.0,
      "step": 3718
    },
    {
      "epoch": 1.8916581892166837,
      "grad_norm": 0.911518931388855,
      "learning_rate": 1e-05,
      "loss": 0.4754,
      "mean_token_accuracy": 0.8477194905281067,
      "num_tokens": 592503410.0,
      "step": 3719
    },
    {
      "epoch": 1.8921668362156663,
      "grad_norm": 0.9388558864593506,
      "learning_rate": 1e-05,
      "loss": 0.4723,
      "mean_token_accuracy": 0.8482939004898071,
      "num_tokens": 592667852.0,
      "step": 3720
    },
    {
      "epoch": 1.892675483214649,
      "grad_norm": 0.9579901695251465,
      "learning_rate": 1e-05,
      "loss": 0.4793,
      "mean_token_accuracy": 0.8475110530853271,
      "num_tokens": 592827432.0,
      "step": 3721
    },
    {
      "epoch": 1.8931841302136316,
      "grad_norm": 0.8853926062583923,
      "learning_rate": 1e-05,
      "loss": 0.4786,
      "mean_token_accuracy": 0.8468433022499084,
      "num_tokens": 592990364.0,
      "step": 3722
    },
    {
      "epoch": 1.8936927772126144,
      "grad_norm": 0.9219601154327393,
      "learning_rate": 1e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8503814935684204,
      "num_tokens": 593145072.0,
      "step": 3723
    },
    {
      "epoch": 1.8942014242115972,
      "grad_norm": 0.8941967487335205,
      "learning_rate": 1e-05,
      "loss": 0.4716,
      "mean_token_accuracy": 0.8500727415084839,
      "num_tokens": 593313033.0,
      "step": 3724
    },
    {
      "epoch": 1.89471007121058,
      "grad_norm": 0.9067884087562561,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.8485181331634521,
      "num_tokens": 593473823.0,
      "step": 3725
    },
    {
      "epoch": 1.8952187182095626,
      "grad_norm": 0.9031772017478943,
      "learning_rate": 1e-05,
      "loss": 0.4473,
      "mean_token_accuracy": 0.8565289974212646,
      "num_tokens": 593629565.0,
      "step": 3726
    },
    {
      "epoch": 1.8957273652085451,
      "grad_norm": 0.9642429351806641,
      "learning_rate": 1e-05,
      "loss": 0.4716,
      "mean_token_accuracy": 0.8493976593017578,
      "num_tokens": 593780513.0,
      "step": 3727
    },
    {
      "epoch": 1.896236012207528,
      "grad_norm": 0.9130746126174927,
      "learning_rate": 1e-05,
      "loss": 0.4846,
      "mean_token_accuracy": 0.8470790386199951,
      "num_tokens": 593932811.0,
      "step": 3728
    },
    {
      "epoch": 1.8967446592065107,
      "grad_norm": 0.9594608545303345,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.851692795753479,
      "num_tokens": 594084116.0,
      "step": 3729
    },
    {
      "epoch": 1.8972533062054935,
      "grad_norm": 0.9444839954376221,
      "learning_rate": 1e-05,
      "loss": 0.4491,
      "mean_token_accuracy": 0.856663703918457,
      "num_tokens": 594251800.0,
      "step": 3730
    },
    {
      "epoch": 1.897761953204476,
      "grad_norm": 0.9218647480010986,
      "learning_rate": 1e-05,
      "loss": 0.4788,
      "mean_token_accuracy": 0.8491775393486023,
      "num_tokens": 594412917.0,
      "step": 3731
    },
    {
      "epoch": 1.8982706002034588,
      "grad_norm": 0.9814186096191406,
      "learning_rate": 1e-05,
      "loss": 0.4792,
      "mean_token_accuracy": 0.8472409844398499,
      "num_tokens": 594566525.0,
      "step": 3732
    },
    {
      "epoch": 1.8987792472024414,
      "grad_norm": 0.9457241892814636,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.8417283296585083,
      "num_tokens": 594735192.0,
      "step": 3733
    },
    {
      "epoch": 1.8992878942014242,
      "grad_norm": 0.8630687594413757,
      "learning_rate": 1e-05,
      "loss": 0.4424,
      "mean_token_accuracy": 0.8581480383872986,
      "num_tokens": 594903927.0,
      "step": 3734
    },
    {
      "epoch": 1.899796541200407,
      "grad_norm": 0.9283967018127441,
      "learning_rate": 1e-05,
      "loss": 0.4681,
      "mean_token_accuracy": 0.8492915630340576,
      "num_tokens": 595059127.0,
      "step": 3735
    },
    {
      "epoch": 1.9003051881993898,
      "grad_norm": 0.9585233330726624,
      "learning_rate": 1e-05,
      "loss": 0.4506,
      "mean_token_accuracy": 0.8559272289276123,
      "num_tokens": 595203411.0,
      "step": 3736
    },
    {
      "epoch": 1.9008138351983723,
      "grad_norm": 0.9517952799797058,
      "learning_rate": 1e-05,
      "loss": 0.5178,
      "mean_token_accuracy": 0.8370660543441772,
      "num_tokens": 595365030.0,
      "step": 3737
    },
    {
      "epoch": 1.901322482197355,
      "grad_norm": 0.9465552568435669,
      "learning_rate": 1e-05,
      "loss": 0.4722,
      "mean_token_accuracy": 0.8481013774871826,
      "num_tokens": 595523298.0,
      "step": 3738
    },
    {
      "epoch": 1.9018311291963377,
      "grad_norm": 0.9181548953056335,
      "learning_rate": 1e-05,
      "loss": 0.465,
      "mean_token_accuracy": 0.8516472578048706,
      "num_tokens": 595685557.0,
      "step": 3739
    },
    {
      "epoch": 1.9023397761953205,
      "grad_norm": 0.9235841035842896,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8593134880065918,
      "num_tokens": 595851295.0,
      "step": 3740
    },
    {
      "epoch": 1.9028484231943033,
      "grad_norm": 0.9412876963615417,
      "learning_rate": 1e-05,
      "loss": 0.4595,
      "mean_token_accuracy": 0.8510797619819641,
      "num_tokens": 596006533.0,
      "step": 3741
    },
    {
      "epoch": 1.9033570701932858,
      "grad_norm": 0.9959353804588318,
      "learning_rate": 1e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8455770015716553,
      "num_tokens": 596165520.0,
      "step": 3742
    },
    {
      "epoch": 1.9038657171922686,
      "grad_norm": 0.968315064907074,
      "learning_rate": 1e-05,
      "loss": 0.483,
      "mean_token_accuracy": 0.8484834432601929,
      "num_tokens": 596331455.0,
      "step": 3743
    },
    {
      "epoch": 1.9043743641912512,
      "grad_norm": 0.9146832227706909,
      "learning_rate": 1e-05,
      "loss": 0.4644,
      "mean_token_accuracy": 0.8519558310508728,
      "num_tokens": 596492724.0,
      "step": 3744
    },
    {
      "epoch": 1.904883011190234,
      "grad_norm": 0.8960966467857361,
      "learning_rate": 1e-05,
      "loss": 0.4594,
      "mean_token_accuracy": 0.8516942858695984,
      "num_tokens": 596655813.0,
      "step": 3745
    },
    {
      "epoch": 1.9053916581892167,
      "grad_norm": 0.8998123407363892,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8601040244102478,
      "num_tokens": 596815498.0,
      "step": 3746
    },
    {
      "epoch": 1.9059003051881995,
      "grad_norm": 0.9438427686691284,
      "learning_rate": 1e-05,
      "loss": 0.5112,
      "mean_token_accuracy": 0.8384228348731995,
      "num_tokens": 596976097.0,
      "step": 3747
    },
    {
      "epoch": 1.906408952187182,
      "grad_norm": 0.8788058161735535,
      "learning_rate": 1e-05,
      "loss": 0.4693,
      "mean_token_accuracy": 0.8497171401977539,
      "num_tokens": 597137505.0,
      "step": 3748
    },
    {
      "epoch": 1.9069175991861647,
      "grad_norm": 0.9773399829864502,
      "learning_rate": 1e-05,
      "loss": 0.4739,
      "mean_token_accuracy": 0.8490920066833496,
      "num_tokens": 597290245.0,
      "step": 3749
    },
    {
      "epoch": 1.9074262461851474,
      "grad_norm": 0.9383112788200378,
      "learning_rate": 1e-05,
      "loss": 0.4779,
      "mean_token_accuracy": 0.8457980751991272,
      "num_tokens": 597439964.0,
      "step": 3750
    },
    {
      "epoch": 1.9079348931841302,
      "grad_norm": 0.894231379032135,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.8520057797431946,
      "num_tokens": 597601954.0,
      "step": 3751
    },
    {
      "epoch": 1.908443540183113,
      "grad_norm": 0.9662824273109436,
      "learning_rate": 1e-05,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8424139022827148,
      "num_tokens": 597762524.0,
      "step": 3752
    },
    {
      "epoch": 1.9089521871820956,
      "grad_norm": 0.9119341373443604,
      "learning_rate": 1e-05,
      "loss": 0.4681,
      "mean_token_accuracy": 0.8486347198486328,
      "num_tokens": 597915383.0,
      "step": 3753
    },
    {
      "epoch": 1.9094608341810784,
      "grad_norm": 0.9073940515518188,
      "learning_rate": 1e-05,
      "loss": 0.5003,
      "mean_token_accuracy": 0.8417505025863647,
      "num_tokens": 598088815.0,
      "step": 3754
    },
    {
      "epoch": 1.909969481180061,
      "grad_norm": 0.9285572171211243,
      "learning_rate": 1e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8495465517044067,
      "num_tokens": 598243445.0,
      "step": 3755
    },
    {
      "epoch": 1.9104781281790437,
      "grad_norm": 0.920777440071106,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8522055149078369,
      "num_tokens": 598408435.0,
      "step": 3756
    },
    {
      "epoch": 1.9109867751780265,
      "grad_norm": 0.9562985301017761,
      "learning_rate": 1e-05,
      "loss": 0.4431,
      "mean_token_accuracy": 0.8574119806289673,
      "num_tokens": 598545916.0,
      "step": 3757
    },
    {
      "epoch": 1.9114954221770093,
      "grad_norm": 0.9341357350349426,
      "learning_rate": 1e-05,
      "loss": 0.4815,
      "mean_token_accuracy": 0.8483132123947144,
      "num_tokens": 598703467.0,
      "step": 3758
    },
    {
      "epoch": 1.9120040691759919,
      "grad_norm": 0.9093040823936462,
      "learning_rate": 1e-05,
      "loss": 0.4778,
      "mean_token_accuracy": 0.8470190167427063,
      "num_tokens": 598878943.0,
      "step": 3759
    },
    {
      "epoch": 1.9125127161749744,
      "grad_norm": 0.8956009745597839,
      "learning_rate": 1e-05,
      "loss": 0.4641,
      "mean_token_accuracy": 0.8517540693283081,
      "num_tokens": 599047533.0,
      "step": 3760
    },
    {
      "epoch": 1.9130213631739572,
      "grad_norm": 0.9305961728096008,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8554340600967407,
      "num_tokens": 599208433.0,
      "step": 3761
    },
    {
      "epoch": 1.91353001017294,
      "grad_norm": 0.9422276020050049,
      "learning_rate": 1e-05,
      "loss": 0.4778,
      "mean_token_accuracy": 0.8485419750213623,
      "num_tokens": 599369815.0,
      "step": 3762
    },
    {
      "epoch": 1.9140386571719228,
      "grad_norm": 0.9445098638534546,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "mean_token_accuracy": 0.8574182391166687,
      "num_tokens": 599529177.0,
      "step": 3763
    },
    {
      "epoch": 1.9145473041709054,
      "grad_norm": 1.0218015909194946,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "mean_token_accuracy": 0.858683705329895,
      "num_tokens": 599687509.0,
      "step": 3764
    },
    {
      "epoch": 1.9150559511698881,
      "grad_norm": 0.9703574180603027,
      "learning_rate": 1e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8607935905456543,
      "num_tokens": 599837337.0,
      "step": 3765
    },
    {
      "epoch": 1.9155645981688707,
      "grad_norm": 0.9127360582351685,
      "learning_rate": 1e-05,
      "loss": 0.4582,
      "mean_token_accuracy": 0.8523753881454468,
      "num_tokens": 599989016.0,
      "step": 3766
    },
    {
      "epoch": 1.9160732451678535,
      "grad_norm": 0.8837084770202637,
      "learning_rate": 1e-05,
      "loss": 0.4835,
      "mean_token_accuracy": 0.845979630947113,
      "num_tokens": 600150348.0,
      "step": 3767
    },
    {
      "epoch": 1.9165818921668363,
      "grad_norm": 0.993800163269043,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8495070934295654,
      "num_tokens": 600296480.0,
      "step": 3768
    },
    {
      "epoch": 1.917090539165819,
      "grad_norm": 0.9457738995552063,
      "learning_rate": 1e-05,
      "loss": 0.453,
      "mean_token_accuracy": 0.8544372320175171,
      "num_tokens": 600457101.0,
      "step": 3769
    },
    {
      "epoch": 1.9175991861648016,
      "grad_norm": 0.8690885305404663,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.853378176689148,
      "num_tokens": 600617011.0,
      "step": 3770
    },
    {
      "epoch": 1.9181078331637842,
      "grad_norm": 1.2345824241638184,
      "learning_rate": 1e-05,
      "loss": 0.5044,
      "mean_token_accuracy": 0.840251088142395,
      "num_tokens": 600776697.0,
      "step": 3771
    },
    {
      "epoch": 1.918616480162767,
      "grad_norm": 0.9083021283149719,
      "learning_rate": 1e-05,
      "loss": 0.4732,
      "mean_token_accuracy": 0.849480152130127,
      "num_tokens": 600942695.0,
      "step": 3772
    },
    {
      "epoch": 1.9191251271617498,
      "grad_norm": 0.9992514252662659,
      "learning_rate": 1e-05,
      "loss": 0.4724,
      "mean_token_accuracy": 0.848492443561554,
      "num_tokens": 601098526.0,
      "step": 3773
    },
    {
      "epoch": 1.9196337741607326,
      "grad_norm": 0.9319624304771423,
      "learning_rate": 1e-05,
      "loss": 0.4845,
      "mean_token_accuracy": 0.8464127779006958,
      "num_tokens": 601254189.0,
      "step": 3774
    },
    {
      "epoch": 1.9201424211597151,
      "grad_norm": 0.8973638415336609,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8527525663375854,
      "num_tokens": 601408256.0,
      "step": 3775
    },
    {
      "epoch": 1.920651068158698,
      "grad_norm": 0.9371774196624756,
      "learning_rate": 1e-05,
      "loss": 0.4829,
      "mean_token_accuracy": 0.8458545207977295,
      "num_tokens": 601571439.0,
      "step": 3776
    },
    {
      "epoch": 1.9211597151576805,
      "grad_norm": 0.9487343430519104,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.8576250076293945,
      "num_tokens": 601734945.0,
      "step": 3777
    },
    {
      "epoch": 1.9216683621566633,
      "grad_norm": 0.937295138835907,
      "learning_rate": 1e-05,
      "loss": 0.4883,
      "mean_token_accuracy": 0.8438887000083923,
      "num_tokens": 601896520.0,
      "step": 3778
    },
    {
      "epoch": 1.922177009155646,
      "grad_norm": 0.9660419821739197,
      "learning_rate": 1e-05,
      "loss": 0.4683,
      "mean_token_accuracy": 0.8490821123123169,
      "num_tokens": 602048648.0,
      "step": 3779
    },
    {
      "epoch": 1.9226856561546288,
      "grad_norm": 0.9533944725990295,
      "learning_rate": 1e-05,
      "loss": 0.4638,
      "mean_token_accuracy": 0.8520326614379883,
      "num_tokens": 602203498.0,
      "step": 3780
    },
    {
      "epoch": 1.9231943031536114,
      "grad_norm": 0.9050354361534119,
      "learning_rate": 1e-05,
      "loss": 0.4609,
      "mean_token_accuracy": 0.8517625331878662,
      "num_tokens": 602353361.0,
      "step": 3781
    },
    {
      "epoch": 1.923702950152594,
      "grad_norm": 0.9708511233329773,
      "learning_rate": 1e-05,
      "loss": 0.4756,
      "mean_token_accuracy": 0.8472548723220825,
      "num_tokens": 602510131.0,
      "step": 3782
    },
    {
      "epoch": 1.9242115971515767,
      "grad_norm": 0.8962816596031189,
      "learning_rate": 1e-05,
      "loss": 0.4759,
      "mean_token_accuracy": 0.8474619388580322,
      "num_tokens": 602679344.0,
      "step": 3783
    },
    {
      "epoch": 1.9247202441505595,
      "grad_norm": 0.9285560250282288,
      "learning_rate": 1e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8549748063087463,
      "num_tokens": 602835451.0,
      "step": 3784
    },
    {
      "epoch": 1.9252288911495423,
      "grad_norm": 0.9716651439666748,
      "learning_rate": 1e-05,
      "loss": 0.4734,
      "mean_token_accuracy": 0.8499223589897156,
      "num_tokens": 602995790.0,
      "step": 3785
    },
    {
      "epoch": 1.9257375381485249,
      "grad_norm": 0.9989169239997864,
      "learning_rate": 1e-05,
      "loss": 0.4802,
      "mean_token_accuracy": 0.846304714679718,
      "num_tokens": 603141191.0,
      "step": 3786
    },
    {
      "epoch": 1.9262461851475077,
      "grad_norm": 0.8882007002830505,
      "learning_rate": 1e-05,
      "loss": 0.4587,
      "mean_token_accuracy": 0.8536261320114136,
      "num_tokens": 603312610.0,
      "step": 3787
    },
    {
      "epoch": 1.9267548321464902,
      "grad_norm": 0.9280068278312683,
      "learning_rate": 1e-05,
      "loss": 0.4574,
      "mean_token_accuracy": 0.8533625602722168,
      "num_tokens": 603470478.0,
      "step": 3788
    },
    {
      "epoch": 1.927263479145473,
      "grad_norm": 1.2315386533737183,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8540915846824646,
      "num_tokens": 603640127.0,
      "step": 3789
    },
    {
      "epoch": 1.9277721261444558,
      "grad_norm": 0.9625018835067749,
      "learning_rate": 1e-05,
      "loss": 0.4568,
      "mean_token_accuracy": 0.8532472252845764,
      "num_tokens": 603791455.0,
      "step": 3790
    },
    {
      "epoch": 1.9282807731434386,
      "grad_norm": 0.9542778134346008,
      "learning_rate": 1e-05,
      "loss": 0.4507,
      "mean_token_accuracy": 0.8542107343673706,
      "num_tokens": 603950683.0,
      "step": 3791
    },
    {
      "epoch": 1.9287894201424212,
      "grad_norm": 1.1490851640701294,
      "learning_rate": 1e-05,
      "loss": 0.5005,
      "mean_token_accuracy": 0.8407775163650513,
      "num_tokens": 604109105.0,
      "step": 3792
    },
    {
      "epoch": 1.9292980671414037,
      "grad_norm": 0.9285960793495178,
      "learning_rate": 1e-05,
      "loss": 0.493,
      "mean_token_accuracy": 0.8430469036102295,
      "num_tokens": 604268766.0,
      "step": 3793
    },
    {
      "epoch": 1.9298067141403865,
      "grad_norm": 0.954233705997467,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8519355058670044,
      "num_tokens": 604425281.0,
      "step": 3794
    },
    {
      "epoch": 1.9303153611393693,
      "grad_norm": 0.9654402136802673,
      "learning_rate": 1e-05,
      "loss": 0.5106,
      "mean_token_accuracy": 0.8387899398803711,
      "num_tokens": 604591434.0,
      "step": 3795
    },
    {
      "epoch": 1.930824008138352,
      "grad_norm": 0.9682581424713135,
      "learning_rate": 1e-05,
      "loss": 0.4742,
      "mean_token_accuracy": 0.8516649007797241,
      "num_tokens": 604760916.0,
      "step": 3796
    },
    {
      "epoch": 1.9313326551373347,
      "grad_norm": 0.9230402112007141,
      "learning_rate": 1e-05,
      "loss": 0.497,
      "mean_token_accuracy": 0.8434281945228577,
      "num_tokens": 604925201.0,
      "step": 3797
    },
    {
      "epoch": 1.9318413021363174,
      "grad_norm": 0.9125484824180603,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8544818162918091,
      "num_tokens": 605069512.0,
      "step": 3798
    },
    {
      "epoch": 1.9323499491353,
      "grad_norm": 0.9310237169265747,
      "learning_rate": 1e-05,
      "loss": 0.4991,
      "mean_token_accuracy": 0.8439196348190308,
      "num_tokens": 605236735.0,
      "step": 3799
    },
    {
      "epoch": 1.9328585961342828,
      "grad_norm": 0.9001007080078125,
      "learning_rate": 1e-05,
      "loss": 0.4428,
      "mean_token_accuracy": 0.8576923608779907,
      "num_tokens": 605388317.0,
      "step": 3800
    },
    {
      "epoch": 1.9333672431332656,
      "grad_norm": 0.9525231122970581,
      "learning_rate": 1e-05,
      "loss": 0.5133,
      "mean_token_accuracy": 0.8374272584915161,
      "num_tokens": 605553305.0,
      "step": 3801
    },
    {
      "epoch": 1.9338758901322484,
      "grad_norm": 0.9368336796760559,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8605653643608093,
      "num_tokens": 605720362.0,
      "step": 3802
    },
    {
      "epoch": 1.934384537131231,
      "grad_norm": 0.9109113216400146,
      "learning_rate": 1e-05,
      "loss": 0.4945,
      "mean_token_accuracy": 0.8430343866348267,
      "num_tokens": 605884975.0,
      "step": 3803
    },
    {
      "epoch": 1.9348931841302135,
      "grad_norm": 0.9070175886154175,
      "learning_rate": 1e-05,
      "loss": 0.485,
      "mean_token_accuracy": 0.8453513979911804,
      "num_tokens": 606043288.0,
      "step": 3804
    },
    {
      "epoch": 1.9354018311291963,
      "grad_norm": 0.9492138624191284,
      "learning_rate": 1e-05,
      "loss": 0.4822,
      "mean_token_accuracy": 0.8475384712219238,
      "num_tokens": 606192599.0,
      "step": 3805
    },
    {
      "epoch": 1.935910478128179,
      "grad_norm": 0.8870913982391357,
      "learning_rate": 1e-05,
      "loss": 0.4756,
      "mean_token_accuracy": 0.8476678729057312,
      "num_tokens": 606348632.0,
      "step": 3806
    },
    {
      "epoch": 1.9364191251271619,
      "grad_norm": 0.9374875426292419,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8640589118003845,
      "num_tokens": 606512463.0,
      "step": 3807
    },
    {
      "epoch": 1.9369277721261444,
      "grad_norm": 0.927408754825592,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8542242050170898,
      "num_tokens": 606676157.0,
      "step": 3808
    },
    {
      "epoch": 1.9374364191251272,
      "grad_norm": 0.8937217593193054,
      "learning_rate": 1e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8511909246444702,
      "num_tokens": 606843986.0,
      "step": 3809
    },
    {
      "epoch": 1.9379450661241098,
      "grad_norm": 0.9094258546829224,
      "learning_rate": 1e-05,
      "loss": 0.4873,
      "mean_token_accuracy": 0.8464955687522888,
      "num_tokens": 607011481.0,
      "step": 3810
    },
    {
      "epoch": 1.9384537131230926,
      "grad_norm": 0.9036715030670166,
      "learning_rate": 1e-05,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8398281335830688,
      "num_tokens": 607178849.0,
      "step": 3811
    },
    {
      "epoch": 1.9389623601220753,
      "grad_norm": 0.9271400570869446,
      "learning_rate": 1e-05,
      "loss": 0.4615,
      "mean_token_accuracy": 0.851945161819458,
      "num_tokens": 607334595.0,
      "step": 3812
    },
    {
      "epoch": 1.9394710071210581,
      "grad_norm": 0.907317042350769,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8564980030059814,
      "num_tokens": 607488276.0,
      "step": 3813
    },
    {
      "epoch": 1.9399796541200407,
      "grad_norm": 1.031917929649353,
      "learning_rate": 1e-05,
      "loss": 0.452,
      "mean_token_accuracy": 0.8556896448135376,
      "num_tokens": 607647463.0,
      "step": 3814
    },
    {
      "epoch": 1.9404883011190233,
      "grad_norm": 0.9541162252426147,
      "learning_rate": 1e-05,
      "loss": 0.4722,
      "mean_token_accuracy": 0.8490946888923645,
      "num_tokens": 607808299.0,
      "step": 3815
    },
    {
      "epoch": 1.940996948118006,
      "grad_norm": 1.3289088010787964,
      "learning_rate": 1e-05,
      "loss": 0.4582,
      "mean_token_accuracy": 0.8528410196304321,
      "num_tokens": 607969176.0,
      "step": 3816
    },
    {
      "epoch": 1.9415055951169888,
      "grad_norm": 0.9064154624938965,
      "learning_rate": 1e-05,
      "loss": 0.5181,
      "mean_token_accuracy": 0.835903525352478,
      "num_tokens": 608134248.0,
      "step": 3817
    },
    {
      "epoch": 1.9420142421159716,
      "grad_norm": 0.936055064201355,
      "learning_rate": 1e-05,
      "loss": 0.4516,
      "mean_token_accuracy": 0.8539500832557678,
      "num_tokens": 608294081.0,
      "step": 3818
    },
    {
      "epoch": 1.9425228891149542,
      "grad_norm": 0.9543054103851318,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8487963676452637,
      "num_tokens": 608454699.0,
      "step": 3819
    },
    {
      "epoch": 1.943031536113937,
      "grad_norm": 0.8393166065216064,
      "learning_rate": 1e-05,
      "loss": 0.4812,
      "mean_token_accuracy": 0.8456804752349854,
      "num_tokens": 608626842.0,
      "step": 3820
    },
    {
      "epoch": 1.9435401831129195,
      "grad_norm": 0.8989098072052002,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.8546445369720459,
      "num_tokens": 608803353.0,
      "step": 3821
    },
    {
      "epoch": 1.9440488301119023,
      "grad_norm": 0.9190250039100647,
      "learning_rate": 1e-05,
      "loss": 0.467,
      "mean_token_accuracy": 0.8513407707214355,
      "num_tokens": 608969123.0,
      "step": 3822
    },
    {
      "epoch": 1.9445574771108851,
      "grad_norm": 0.9733286499977112,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8549008369445801,
      "num_tokens": 609112871.0,
      "step": 3823
    },
    {
      "epoch": 1.945066124109868,
      "grad_norm": 1.037373423576355,
      "learning_rate": 1e-05,
      "loss": 0.4717,
      "mean_token_accuracy": 0.8502563238143921,
      "num_tokens": 609267308.0,
      "step": 3824
    },
    {
      "epoch": 1.9455747711088505,
      "grad_norm": 0.9469419121742249,
      "learning_rate": 1e-05,
      "loss": 0.4755,
      "mean_token_accuracy": 0.846880316734314,
      "num_tokens": 609429279.0,
      "step": 3825
    },
    {
      "epoch": 1.946083418107833,
      "grad_norm": 0.9067391157150269,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8481675386428833,
      "num_tokens": 609605979.0,
      "step": 3826
    },
    {
      "epoch": 1.9465920651068158,
      "grad_norm": 1.0206915140151978,
      "learning_rate": 1e-05,
      "loss": 0.4456,
      "mean_token_accuracy": 0.8566611409187317,
      "num_tokens": 609754605.0,
      "step": 3827
    },
    {
      "epoch": 1.9471007121057986,
      "grad_norm": 0.9272356033325195,
      "learning_rate": 1e-05,
      "loss": 0.4788,
      "mean_token_accuracy": 0.8478904962539673,
      "num_tokens": 609905194.0,
      "step": 3828
    },
    {
      "epoch": 1.9476093591047814,
      "grad_norm": 0.889542818069458,
      "learning_rate": 1e-05,
      "loss": 0.4646,
      "mean_token_accuracy": 0.8507856130599976,
      "num_tokens": 610068975.0,
      "step": 3829
    },
    {
      "epoch": 1.948118006103764,
      "grad_norm": 0.9088366031646729,
      "learning_rate": 1e-05,
      "loss": 0.4919,
      "mean_token_accuracy": 0.8457269668579102,
      "num_tokens": 610232914.0,
      "step": 3830
    },
    {
      "epoch": 1.9486266531027467,
      "grad_norm": 0.9174922108650208,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8533996343612671,
      "num_tokens": 610377414.0,
      "step": 3831
    },
    {
      "epoch": 1.9491353001017293,
      "grad_norm": 0.952247679233551,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8532795906066895,
      "num_tokens": 610524356.0,
      "step": 3832
    },
    {
      "epoch": 1.949643947100712,
      "grad_norm": 0.8449368476867676,
      "learning_rate": 1e-05,
      "loss": 0.4706,
      "mean_token_accuracy": 0.8495937585830688,
      "num_tokens": 610687425.0,
      "step": 3833
    },
    {
      "epoch": 1.9501525940996949,
      "grad_norm": 0.9186002612113953,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "mean_token_accuracy": 0.8576358556747437,
      "num_tokens": 610848911.0,
      "step": 3834
    },
    {
      "epoch": 1.9506612410986777,
      "grad_norm": 0.9892232418060303,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "mean_token_accuracy": 0.8579406142234802,
      "num_tokens": 610992150.0,
      "step": 3835
    },
    {
      "epoch": 1.9511698880976602,
      "grad_norm": 0.9537031650543213,
      "learning_rate": 1e-05,
      "loss": 0.4697,
      "mean_token_accuracy": 0.8509016036987305,
      "num_tokens": 611151896.0,
      "step": 3836
    },
    {
      "epoch": 1.9516785350966428,
      "grad_norm": 0.8665832877159119,
      "learning_rate": 1e-05,
      "loss": 0.4631,
      "mean_token_accuracy": 0.852397084236145,
      "num_tokens": 611322548.0,
      "step": 3837
    },
    {
      "epoch": 1.9521871820956256,
      "grad_norm": 0.9472004771232605,
      "learning_rate": 1e-05,
      "loss": 0.4391,
      "mean_token_accuracy": 0.8582301735877991,
      "num_tokens": 611477707.0,
      "step": 3838
    },
    {
      "epoch": 1.9526958290946084,
      "grad_norm": 1.0446181297302246,
      "learning_rate": 1e-05,
      "loss": 0.4804,
      "mean_token_accuracy": 0.8453882932662964,
      "num_tokens": 611633623.0,
      "step": 3839
    },
    {
      "epoch": 1.9532044760935912,
      "grad_norm": 0.9511471390724182,
      "learning_rate": 1e-05,
      "loss": 0.4801,
      "mean_token_accuracy": 0.8465399742126465,
      "num_tokens": 611786310.0,
      "step": 3840
    },
    {
      "epoch": 1.9537131230925737,
      "grad_norm": 1.009140968322754,
      "learning_rate": 1e-05,
      "loss": 0.4833,
      "mean_token_accuracy": 0.8464534282684326,
      "num_tokens": 611939348.0,
      "step": 3841
    },
    {
      "epoch": 1.9542217700915565,
      "grad_norm": 0.9803367257118225,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8545321822166443,
      "num_tokens": 612106242.0,
      "step": 3842
    },
    {
      "epoch": 1.954730417090539,
      "grad_norm": 0.9608579277992249,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.8532856702804565,
      "num_tokens": 612259738.0,
      "step": 3843
    },
    {
      "epoch": 1.9552390640895219,
      "grad_norm": 1.0368393659591675,
      "learning_rate": 1e-05,
      "loss": 0.4404,
      "mean_token_accuracy": 0.8589491844177246,
      "num_tokens": 612408694.0,
      "step": 3844
    },
    {
      "epoch": 1.9557477110885046,
      "grad_norm": 1.0139336585998535,
      "learning_rate": 1e-05,
      "loss": 0.4909,
      "mean_token_accuracy": 0.8452524542808533,
      "num_tokens": 612570005.0,
      "step": 3845
    },
    {
      "epoch": 1.9562563580874874,
      "grad_norm": 0.9647461175918579,
      "learning_rate": 1e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8602635860443115,
      "num_tokens": 612723003.0,
      "step": 3846
    },
    {
      "epoch": 1.95676500508647,
      "grad_norm": 0.9352390170097351,
      "learning_rate": 1e-05,
      "loss": 0.4532,
      "mean_token_accuracy": 0.8548156023025513,
      "num_tokens": 612893200.0,
      "step": 3847
    },
    {
      "epoch": 1.9572736520854526,
      "grad_norm": 1.0115501880645752,
      "learning_rate": 1e-05,
      "loss": 0.4864,
      "mean_token_accuracy": 0.8474304676055908,
      "num_tokens": 613044117.0,
      "step": 3848
    },
    {
      "epoch": 1.9577822990844354,
      "grad_norm": 0.969480037689209,
      "learning_rate": 1e-05,
      "loss": 0.4243,
      "mean_token_accuracy": 0.8618882894515991,
      "num_tokens": 613176151.0,
      "step": 3849
    },
    {
      "epoch": 1.9582909460834181,
      "grad_norm": 0.8841404914855957,
      "learning_rate": 1e-05,
      "loss": 0.4612,
      "mean_token_accuracy": 0.8527694344520569,
      "num_tokens": 613322625.0,
      "step": 3850
    },
    {
      "epoch": 1.958799593082401,
      "grad_norm": 1.028300404548645,
      "learning_rate": 1e-05,
      "loss": 0.4302,
      "mean_token_accuracy": 0.8599109649658203,
      "num_tokens": 613484461.0,
      "step": 3851
    },
    {
      "epoch": 1.9593082400813835,
      "grad_norm": 1.005175232887268,
      "learning_rate": 1e-05,
      "loss": 0.4941,
      "mean_token_accuracy": 0.8423586487770081,
      "num_tokens": 613652844.0,
      "step": 3852
    },
    {
      "epoch": 1.9598168870803663,
      "grad_norm": 1.6460040807724,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8538551330566406,
      "num_tokens": 613810470.0,
      "step": 3853
    },
    {
      "epoch": 1.9603255340793488,
      "grad_norm": 1.1169441938400269,
      "learning_rate": 1e-05,
      "loss": 0.459,
      "mean_token_accuracy": 0.8533007502555847,
      "num_tokens": 613969421.0,
      "step": 3854
    },
    {
      "epoch": 1.9608341810783316,
      "grad_norm": 0.9980722665786743,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8556852340698242,
      "num_tokens": 614130745.0,
      "step": 3855
    },
    {
      "epoch": 1.9613428280773144,
      "grad_norm": 0.9866142868995667,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8559843301773071,
      "num_tokens": 614295637.0,
      "step": 3856
    },
    {
      "epoch": 1.9618514750762972,
      "grad_norm": 1.0588595867156982,
      "learning_rate": 1e-05,
      "loss": 0.4838,
      "mean_token_accuracy": 0.8471004962921143,
      "num_tokens": 614455524.0,
      "step": 3857
    },
    {
      "epoch": 1.9623601220752798,
      "grad_norm": 0.9205344915390015,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8560166358947754,
      "num_tokens": 614612191.0,
      "step": 3858
    },
    {
      "epoch": 1.9628687690742623,
      "grad_norm": 0.9048281908035278,
      "learning_rate": 1e-05,
      "loss": 0.4677,
      "mean_token_accuracy": 0.8504944443702698,
      "num_tokens": 614782233.0,
      "step": 3859
    },
    {
      "epoch": 1.9633774160732451,
      "grad_norm": 1.050432562828064,
      "learning_rate": 1e-05,
      "loss": 0.4825,
      "mean_token_accuracy": 0.8455482125282288,
      "num_tokens": 614941005.0,
      "step": 3860
    },
    {
      "epoch": 1.963886063072228,
      "grad_norm": 0.865611732006073,
      "learning_rate": 1e-05,
      "loss": 0.4472,
      "mean_token_accuracy": 0.8554700613021851,
      "num_tokens": 615100608.0,
      "step": 3861
    },
    {
      "epoch": 1.9643947100712107,
      "grad_norm": 0.9181163311004639,
      "learning_rate": 1e-05,
      "loss": 0.4519,
      "mean_token_accuracy": 0.8545412421226501,
      "num_tokens": 615260901.0,
      "step": 3862
    },
    {
      "epoch": 1.9649033570701933,
      "grad_norm": 1.0207750797271729,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.853108286857605,
      "num_tokens": 615416052.0,
      "step": 3863
    },
    {
      "epoch": 1.965412004069176,
      "grad_norm": 0.8985828161239624,
      "learning_rate": 1e-05,
      "loss": 0.4758,
      "mean_token_accuracy": 0.8490435481071472,
      "num_tokens": 615576002.0,
      "step": 3864
    },
    {
      "epoch": 1.9659206510681586,
      "grad_norm": 1.0118473768234253,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8502198457717896,
      "num_tokens": 615742455.0,
      "step": 3865
    },
    {
      "epoch": 1.9664292980671414,
      "grad_norm": 0.9573753476142883,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8564034700393677,
      "num_tokens": 615906641.0,
      "step": 3866
    },
    {
      "epoch": 1.9669379450661242,
      "grad_norm": 0.9813663363456726,
      "learning_rate": 1e-05,
      "loss": 0.4824,
      "mean_token_accuracy": 0.8461043834686279,
      "num_tokens": 616075924.0,
      "step": 3867
    },
    {
      "epoch": 1.967446592065107,
      "grad_norm": 1.0511075258255005,
      "learning_rate": 1e-05,
      "loss": 0.5225,
      "mean_token_accuracy": 0.8352948427200317,
      "num_tokens": 616248968.0,
      "step": 3868
    },
    {
      "epoch": 1.9679552390640895,
      "grad_norm": 0.9328294396400452,
      "learning_rate": 1e-05,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8543003797531128,
      "num_tokens": 616416342.0,
      "step": 3869
    },
    {
      "epoch": 1.968463886063072,
      "grad_norm": 1.0230273008346558,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.84980309009552,
      "num_tokens": 616578475.0,
      "step": 3870
    },
    {
      "epoch": 1.9689725330620549,
      "grad_norm": 0.910184919834137,
      "learning_rate": 1e-05,
      "loss": 0.5046,
      "mean_token_accuracy": 0.8411458730697632,
      "num_tokens": 616746225.0,
      "step": 3871
    },
    {
      "epoch": 1.9694811800610377,
      "grad_norm": 0.8828282356262207,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8588099479675293,
      "num_tokens": 616899976.0,
      "step": 3872
    },
    {
      "epoch": 1.9699898270600205,
      "grad_norm": 1.0311261415481567,
      "learning_rate": 1e-05,
      "loss": 0.4711,
      "mean_token_accuracy": 0.8495997190475464,
      "num_tokens": 617064650.0,
      "step": 3873
    },
    {
      "epoch": 1.970498474059003,
      "grad_norm": 0.947348952293396,
      "learning_rate": 1e-05,
      "loss": 0.4353,
      "mean_token_accuracy": 0.8598195910453796,
      "num_tokens": 617214839.0,
      "step": 3874
    },
    {
      "epoch": 1.9710071210579858,
      "grad_norm": 1.03840172290802,
      "learning_rate": 1e-05,
      "loss": 0.4986,
      "mean_token_accuracy": 0.8405134677886963,
      "num_tokens": 617367210.0,
      "step": 3875
    },
    {
      "epoch": 1.9715157680569684,
      "grad_norm": 0.9516762495040894,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.8516678214073181,
      "num_tokens": 617512120.0,
      "step": 3876
    },
    {
      "epoch": 1.9720244150559512,
      "grad_norm": 0.9232913255691528,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "mean_token_accuracy": 0.8564640879631042,
      "num_tokens": 617668697.0,
      "step": 3877
    },
    {
      "epoch": 1.972533062054934,
      "grad_norm": 1.0209755897521973,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8471225500106812,
      "num_tokens": 617823554.0,
      "step": 3878
    },
    {
      "epoch": 1.9730417090539167,
      "grad_norm": 1.0261807441711426,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8503115773200989,
      "num_tokens": 617983196.0,
      "step": 3879
    },
    {
      "epoch": 1.9735503560528993,
      "grad_norm": 0.923484206199646,
      "learning_rate": 1e-05,
      "loss": 0.4574,
      "mean_token_accuracy": 0.8527460098266602,
      "num_tokens": 618137243.0,
      "step": 3880
    },
    {
      "epoch": 1.9740590030518819,
      "grad_norm": 0.9335493445396423,
      "learning_rate": 1e-05,
      "loss": 0.454,
      "mean_token_accuracy": 0.8534466624259949,
      "num_tokens": 618302801.0,
      "step": 3881
    },
    {
      "epoch": 1.9745676500508647,
      "grad_norm": 0.9859886169433594,
      "learning_rate": 1e-05,
      "loss": 0.5018,
      "mean_token_accuracy": 0.8423020839691162,
      "num_tokens": 618473609.0,
      "step": 3882
    },
    {
      "epoch": 1.9750762970498474,
      "grad_norm": 0.962785005569458,
      "learning_rate": 1e-05,
      "loss": 0.4451,
      "mean_token_accuracy": 0.8573927879333496,
      "num_tokens": 618635491.0,
      "step": 3883
    },
    {
      "epoch": 1.9755849440488302,
      "grad_norm": 0.9044583439826965,
      "learning_rate": 1e-05,
      "loss": 0.4552,
      "mean_token_accuracy": 0.8537480235099792,
      "num_tokens": 618796427.0,
      "step": 3884
    },
    {
      "epoch": 1.9760935910478128,
      "grad_norm": 0.9657168984413147,
      "learning_rate": 1e-05,
      "loss": 0.48,
      "mean_token_accuracy": 0.8483161330223083,
      "num_tokens": 618958212.0,
      "step": 3885
    },
    {
      "epoch": 1.9766022380467956,
      "grad_norm": 0.9363645911216736,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.8569280505180359,
      "num_tokens": 619108709.0,
      "step": 3886
    },
    {
      "epoch": 1.9771108850457781,
      "grad_norm": 0.8866698145866394,
      "learning_rate": 1e-05,
      "loss": 0.4345,
      "mean_token_accuracy": 0.8591384291648865,
      "num_tokens": 619264010.0,
      "step": 3887
    },
    {
      "epoch": 1.977619532044761,
      "grad_norm": 0.9166999459266663,
      "learning_rate": 1e-05,
      "loss": 0.5002,
      "mean_token_accuracy": 0.8409247398376465,
      "num_tokens": 619435220.0,
      "step": 3888
    },
    {
      "epoch": 1.9781281790437437,
      "grad_norm": 0.9197640419006348,
      "learning_rate": 1e-05,
      "loss": 0.4705,
      "mean_token_accuracy": 0.848235011100769,
      "num_tokens": 619595878.0,
      "step": 3889
    },
    {
      "epoch": 1.9786368260427265,
      "grad_norm": 0.918525755405426,
      "learning_rate": 1e-05,
      "loss": 0.5253,
      "mean_token_accuracy": 0.8348990678787231,
      "num_tokens": 619759819.0,
      "step": 3890
    },
    {
      "epoch": 1.979145473041709,
      "grad_norm": 1.0172572135925293,
      "learning_rate": 1e-05,
      "loss": 0.4545,
      "mean_token_accuracy": 0.8554086685180664,
      "num_tokens": 619915648.0,
      "step": 3891
    },
    {
      "epoch": 1.9796541200406916,
      "grad_norm": 0.9677065014839172,
      "learning_rate": 1e-05,
      "loss": 0.4713,
      "mean_token_accuracy": 0.8500718474388123,
      "num_tokens": 620070498.0,
      "step": 3892
    },
    {
      "epoch": 1.9801627670396744,
      "grad_norm": 0.9305118322372437,
      "learning_rate": 1e-05,
      "loss": 0.4747,
      "mean_token_accuracy": 0.8483951687812805,
      "num_tokens": 620230320.0,
      "step": 3893
    },
    {
      "epoch": 1.9806714140386572,
      "grad_norm": 0.8849009275436401,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8593703508377075,
      "num_tokens": 620393836.0,
      "step": 3894
    },
    {
      "epoch": 1.98118006103764,
      "grad_norm": 0.9445408582687378,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8560349941253662,
      "num_tokens": 620550585.0,
      "step": 3895
    },
    {
      "epoch": 1.9816887080366226,
      "grad_norm": 0.8695283532142639,
      "learning_rate": 1e-05,
      "loss": 0.4788,
      "mean_token_accuracy": 0.8496056795120239,
      "num_tokens": 620719684.0,
      "step": 3896
    },
    {
      "epoch": 1.9821973550356051,
      "grad_norm": 0.9322693943977356,
      "learning_rate": 1e-05,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8461048603057861,
      "num_tokens": 620887069.0,
      "step": 3897
    },
    {
      "epoch": 1.982706002034588,
      "grad_norm": 0.9163092970848083,
      "learning_rate": 1e-05,
      "loss": 0.4684,
      "mean_token_accuracy": 0.8510671854019165,
      "num_tokens": 621050520.0,
      "step": 3898
    },
    {
      "epoch": 1.9832146490335707,
      "grad_norm": 1.0227288007736206,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8476278185844421,
      "num_tokens": 621205316.0,
      "step": 3899
    },
    {
      "epoch": 1.9837232960325535,
      "grad_norm": 0.9090538620948792,
      "learning_rate": 1e-05,
      "loss": 0.4524,
      "mean_token_accuracy": 0.8544093370437622,
      "num_tokens": 621363543.0,
      "step": 3900
    },
    {
      "epoch": 1.9842319430315363,
      "grad_norm": 0.8535559773445129,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.8503838181495667,
      "num_tokens": 621525240.0,
      "step": 3901
    },
    {
      "epoch": 1.9847405900305188,
      "grad_norm": 0.928386390209198,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.856638491153717,
      "num_tokens": 621674098.0,
      "step": 3902
    },
    {
      "epoch": 1.9852492370295014,
      "grad_norm": 0.8908268213272095,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.8504698276519775,
      "num_tokens": 621831203.0,
      "step": 3903
    },
    {
      "epoch": 1.9857578840284842,
      "grad_norm": 0.8927626013755798,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.8538251519203186,
      "num_tokens": 621989155.0,
      "step": 3904
    },
    {
      "epoch": 1.986266531027467,
      "grad_norm": 0.9428632855415344,
      "learning_rate": 1e-05,
      "loss": 0.4319,
      "mean_token_accuracy": 0.8602399826049805,
      "num_tokens": 622136977.0,
      "step": 3905
    },
    {
      "epoch": 1.9867751780264498,
      "grad_norm": 0.909707248210907,
      "learning_rate": 1e-05,
      "loss": 0.5064,
      "mean_token_accuracy": 0.8414626121520996,
      "num_tokens": 622317261.0,
      "step": 3906
    },
    {
      "epoch": 1.9872838250254323,
      "grad_norm": 0.8628239035606384,
      "learning_rate": 1e-05,
      "loss": 0.4943,
      "mean_token_accuracy": 0.8427656292915344,
      "num_tokens": 622482892.0,
      "step": 3907
    },
    {
      "epoch": 1.987792472024415,
      "grad_norm": 0.8984335064888,
      "learning_rate": 1e-05,
      "loss": 0.4598,
      "mean_token_accuracy": 0.8521856665611267,
      "num_tokens": 622647732.0,
      "step": 3908
    },
    {
      "epoch": 1.9883011190233977,
      "grad_norm": 0.8870931267738342,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8595765233039856,
      "num_tokens": 622821700.0,
      "step": 3909
    },
    {
      "epoch": 1.9888097660223805,
      "grad_norm": 0.9406011700630188,
      "learning_rate": 1e-05,
      "loss": 0.4693,
      "mean_token_accuracy": 0.8493680953979492,
      "num_tokens": 622984287.0,
      "step": 3910
    },
    {
      "epoch": 1.9893184130213633,
      "grad_norm": 1.0311652421951294,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.8638802766799927,
      "num_tokens": 623129707.0,
      "step": 3911
    },
    {
      "epoch": 1.989827060020346,
      "grad_norm": 0.9765109419822693,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8548345565795898,
      "num_tokens": 623279377.0,
      "step": 3912
    },
    {
      "epoch": 1.9903357070193286,
      "grad_norm": 0.9124788045883179,
      "learning_rate": 1e-05,
      "loss": 0.471,
      "mean_token_accuracy": 0.8486224412918091,
      "num_tokens": 623451034.0,
      "step": 3913
    },
    {
      "epoch": 1.9908443540183112,
      "grad_norm": 0.9086233973503113,
      "learning_rate": 1e-05,
      "loss": 0.4584,
      "mean_token_accuracy": 0.8524532318115234,
      "num_tokens": 623617093.0,
      "step": 3914
    },
    {
      "epoch": 1.991353001017294,
      "grad_norm": 0.8954272270202637,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8520581126213074,
      "num_tokens": 623769321.0,
      "step": 3915
    },
    {
      "epoch": 1.9918616480162767,
      "grad_norm": 0.916243314743042,
      "learning_rate": 1e-05,
      "loss": 0.4947,
      "mean_token_accuracy": 0.8445735573768616,
      "num_tokens": 623935367.0,
      "step": 3916
    },
    {
      "epoch": 1.9923702950152595,
      "grad_norm": 0.8950624465942383,
      "learning_rate": 1e-05,
      "loss": 0.4808,
      "mean_token_accuracy": 0.8477447032928467,
      "num_tokens": 624097806.0,
      "step": 3917
    },
    {
      "epoch": 1.992878942014242,
      "grad_norm": 0.8239250183105469,
      "learning_rate": 1e-05,
      "loss": 0.4581,
      "mean_token_accuracy": 0.8526318073272705,
      "num_tokens": 624268903.0,
      "step": 3918
    },
    {
      "epoch": 1.9933875890132247,
      "grad_norm": 0.9191027283668518,
      "learning_rate": 1e-05,
      "loss": 0.4719,
      "mean_token_accuracy": 0.8489470481872559,
      "num_tokens": 624421868.0,
      "step": 3919
    },
    {
      "epoch": 1.9938962360122074,
      "grad_norm": 0.8871040940284729,
      "learning_rate": 1e-05,
      "loss": 0.476,
      "mean_token_accuracy": 0.8485807180404663,
      "num_tokens": 624585621.0,
      "step": 3920
    },
    {
      "epoch": 1.9944048830111902,
      "grad_norm": 0.9006506204605103,
      "learning_rate": 1e-05,
      "loss": 0.5007,
      "mean_token_accuracy": 0.8404300808906555,
      "num_tokens": 624745618.0,
      "step": 3921
    },
    {
      "epoch": 1.994913530010173,
      "grad_norm": 0.8424310684204102,
      "learning_rate": 1e-05,
      "loss": 0.4745,
      "mean_token_accuracy": 0.8490932583808899,
      "num_tokens": 624919983.0,
      "step": 3922
    },
    {
      "epoch": 1.9954221770091558,
      "grad_norm": 0.8778648376464844,
      "learning_rate": 1e-05,
      "loss": 0.4957,
      "mean_token_accuracy": 0.8425777554512024,
      "num_tokens": 625088682.0,
      "step": 3923
    },
    {
      "epoch": 1.9959308240081384,
      "grad_norm": 0.9202529191970825,
      "learning_rate": 1e-05,
      "loss": 0.5111,
      "mean_token_accuracy": 0.8390426635742188,
      "num_tokens": 625244861.0,
      "step": 3924
    },
    {
      "epoch": 1.996439471007121,
      "grad_norm": 0.8500998616218567,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.853619396686554,
      "num_tokens": 625402853.0,
      "step": 3925
    },
    {
      "epoch": 1.9969481180061037,
      "grad_norm": 0.9115105271339417,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.853884756565094,
      "num_tokens": 625553633.0,
      "step": 3926
    },
    {
      "epoch": 1.9974567650050865,
      "grad_norm": 0.8809705972671509,
      "learning_rate": 1e-05,
      "loss": 0.4812,
      "mean_token_accuracy": 0.8468331098556519,
      "num_tokens": 625717102.0,
      "step": 3927
    },
    {
      "epoch": 1.9979654120040693,
      "grad_norm": 0.841272234916687,
      "learning_rate": 1e-05,
      "loss": 0.4817,
      "mean_token_accuracy": 0.8469640016555786,
      "num_tokens": 625879310.0,
      "step": 3928
    },
    {
      "epoch": 1.9984740590030519,
      "grad_norm": 0.9201867580413818,
      "learning_rate": 1e-05,
      "loss": 0.4555,
      "mean_token_accuracy": 0.8538429737091064,
      "num_tokens": 626035166.0,
      "step": 3929
    },
    {
      "epoch": 1.9989827060020344,
      "grad_norm": 0.8562085628509521,
      "learning_rate": 1e-05,
      "loss": 0.457,
      "mean_token_accuracy": 0.8542649149894714,
      "num_tokens": 626194056.0,
      "step": 3930
    },
    {
      "epoch": 1.9994913530010172,
      "grad_norm": 0.8933717012405396,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.846855103969574,
      "num_tokens": 626349258.0,
      "step": 3931
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.915454626083374,
      "learning_rate": 1e-05,
      "loss": 0.4431,
      "mean_token_accuracy": 0.8570587038993835,
      "num_tokens": 626514393.0,
      "step": 3932
    },
    {
      "epoch": 2.000508646998983,
      "grad_norm": 0.9072067737579346,
      "learning_rate": 1e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8628668785095215,
      "num_tokens": 626678112.0,
      "step": 3933
    },
    {
      "epoch": 2.0010172939979656,
      "grad_norm": 0.9203000068664551,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.860080897808075,
      "num_tokens": 626843728.0,
      "step": 3934
    },
    {
      "epoch": 2.001525940996948,
      "grad_norm": 0.9907317161560059,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8583724498748779,
      "num_tokens": 627012779.0,
      "step": 3935
    },
    {
      "epoch": 2.0020345879959307,
      "grad_norm": 0.9471038579940796,
      "learning_rate": 1e-05,
      "loss": 0.4623,
      "mean_token_accuracy": 0.8516335487365723,
      "num_tokens": 627179506.0,
      "step": 3936
    },
    {
      "epoch": 2.0025432349949135,
      "grad_norm": 1.0517023801803589,
      "learning_rate": 1e-05,
      "loss": 0.4371,
      "mean_token_accuracy": 0.8596482872962952,
      "num_tokens": 627321312.0,
      "step": 3937
    },
    {
      "epoch": 2.0030518819938963,
      "grad_norm": 1.1122736930847168,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8612704277038574,
      "num_tokens": 627483549.0,
      "step": 3938
    },
    {
      "epoch": 2.003560528992879,
      "grad_norm": 0.9688953757286072,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8595166802406311,
      "num_tokens": 627652408.0,
      "step": 3939
    },
    {
      "epoch": 2.004069175991862,
      "grad_norm": 1.0007293224334717,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8554959893226624,
      "num_tokens": 627813505.0,
      "step": 3940
    },
    {
      "epoch": 2.004577822990844,
      "grad_norm": 0.981654703617096,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8577959537506104,
      "num_tokens": 627970583.0,
      "step": 3941
    },
    {
      "epoch": 2.005086469989827,
      "grad_norm": 0.9396583437919617,
      "learning_rate": 1e-05,
      "loss": 0.4526,
      "mean_token_accuracy": 0.8536440134048462,
      "num_tokens": 628136119.0,
      "step": 3942
    },
    {
      "epoch": 2.0055951169888098,
      "grad_norm": 0.9022424817085266,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8626178503036499,
      "num_tokens": 628294673.0,
      "step": 3943
    },
    {
      "epoch": 2.0061037639877926,
      "grad_norm": 0.9768242239952087,
      "learning_rate": 1e-05,
      "loss": 0.4266,
      "mean_token_accuracy": 0.8636049032211304,
      "num_tokens": 628450141.0,
      "step": 3944
    },
    {
      "epoch": 2.0066124109867753,
      "grad_norm": 0.9007858037948608,
      "learning_rate": 1e-05,
      "loss": 0.4216,
      "mean_token_accuracy": 0.8628674745559692,
      "num_tokens": 628608570.0,
      "step": 3945
    },
    {
      "epoch": 2.0071210579857577,
      "grad_norm": 0.9086043238639832,
      "learning_rate": 1e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.8627064228057861,
      "num_tokens": 628770893.0,
      "step": 3946
    },
    {
      "epoch": 2.0076297049847405,
      "grad_norm": 0.9539920687675476,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.8553727865219116,
      "num_tokens": 628932043.0,
      "step": 3947
    },
    {
      "epoch": 2.0081383519837233,
      "grad_norm": 0.9863036274909973,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8575353026390076,
      "num_tokens": 629091061.0,
      "step": 3948
    },
    {
      "epoch": 2.008646998982706,
      "grad_norm": 0.9033030271530151,
      "learning_rate": 1e-05,
      "loss": 0.4115,
      "mean_token_accuracy": 0.8660039901733398,
      "num_tokens": 629252044.0,
      "step": 3949
    },
    {
      "epoch": 2.009155645981689,
      "grad_norm": 0.9724506735801697,
      "learning_rate": 1e-05,
      "loss": 0.4186,
      "mean_token_accuracy": 0.8633401393890381,
      "num_tokens": 629412815.0,
      "step": 3950
    },
    {
      "epoch": 2.0096642929806716,
      "grad_norm": 0.8986982107162476,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8590998649597168,
      "num_tokens": 629565057.0,
      "step": 3951
    },
    {
      "epoch": 2.010172939979654,
      "grad_norm": 0.9415915012359619,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "mean_token_accuracy": 0.8536491990089417,
      "num_tokens": 629722640.0,
      "step": 3952
    },
    {
      "epoch": 2.0106815869786367,
      "grad_norm": 0.8877673149108887,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.8628571033477783,
      "num_tokens": 629883654.0,
      "step": 3953
    },
    {
      "epoch": 2.0111902339776195,
      "grad_norm": 0.8674871325492859,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8509914875030518,
      "num_tokens": 630044682.0,
      "step": 3954
    },
    {
      "epoch": 2.0116988809766023,
      "grad_norm": 0.9668797254562378,
      "learning_rate": 1e-05,
      "loss": 0.5095,
      "mean_token_accuracy": 0.8393916487693787,
      "num_tokens": 630189841.0,
      "step": 3955
    },
    {
      "epoch": 2.012207527975585,
      "grad_norm": 0.9578163623809814,
      "learning_rate": 1e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.8615664839744568,
      "num_tokens": 630333856.0,
      "step": 3956
    },
    {
      "epoch": 2.0127161749745675,
      "grad_norm": 0.9142828583717346,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.8505555391311646,
      "num_tokens": 630485294.0,
      "step": 3957
    },
    {
      "epoch": 2.0132248219735502,
      "grad_norm": 1.0230820178985596,
      "learning_rate": 1e-05,
      "loss": 0.4553,
      "mean_token_accuracy": 0.8525019884109497,
      "num_tokens": 630642252.0,
      "step": 3958
    },
    {
      "epoch": 2.013733468972533,
      "grad_norm": 0.9404370784759521,
      "learning_rate": 1e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8559948205947876,
      "num_tokens": 630810155.0,
      "step": 3959
    },
    {
      "epoch": 2.014242115971516,
      "grad_norm": 0.9590718150138855,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.854925274848938,
      "num_tokens": 630967182.0,
      "step": 3960
    },
    {
      "epoch": 2.0147507629704986,
      "grad_norm": 0.8920328617095947,
      "learning_rate": 1e-05,
      "loss": 0.4026,
      "mean_token_accuracy": 0.8685610294342041,
      "num_tokens": 631119247.0,
      "step": 3961
    },
    {
      "epoch": 2.0152594099694814,
      "grad_norm": 0.9542795419692993,
      "learning_rate": 1e-05,
      "loss": 0.4006,
      "mean_token_accuracy": 0.8695034980773926,
      "num_tokens": 631277861.0,
      "step": 3962
    },
    {
      "epoch": 2.0157680569684637,
      "grad_norm": 0.8820130825042725,
      "learning_rate": 1e-05,
      "loss": 0.3956,
      "mean_token_accuracy": 0.8703972101211548,
      "num_tokens": 631431425.0,
      "step": 3963
    },
    {
      "epoch": 2.0162767039674465,
      "grad_norm": 1.0530736446380615,
      "learning_rate": 1e-05,
      "loss": 0.4092,
      "mean_token_accuracy": 0.8659566640853882,
      "num_tokens": 631581074.0,
      "step": 3964
    },
    {
      "epoch": 2.0167853509664293,
      "grad_norm": 0.9052873253822327,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8610852956771851,
      "num_tokens": 631737047.0,
      "step": 3965
    },
    {
      "epoch": 2.017293997965412,
      "grad_norm": 0.8882894515991211,
      "learning_rate": 1e-05,
      "loss": 0.4342,
      "mean_token_accuracy": 0.8599810600280762,
      "num_tokens": 631908085.0,
      "step": 3966
    },
    {
      "epoch": 2.017802644964395,
      "grad_norm": 1.0958209037780762,
      "learning_rate": 1e-05,
      "loss": 0.4618,
      "mean_token_accuracy": 0.8512526750564575,
      "num_tokens": 632082603.0,
      "step": 3967
    },
    {
      "epoch": 2.018311291963377,
      "grad_norm": 1.1128910779953003,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8599388599395752,
      "num_tokens": 632228951.0,
      "step": 3968
    },
    {
      "epoch": 2.01881993896236,
      "grad_norm": 0.9493024945259094,
      "learning_rate": 1e-05,
      "loss": 0.4291,
      "mean_token_accuracy": 0.8602290153503418,
      "num_tokens": 632395510.0,
      "step": 3969
    },
    {
      "epoch": 2.019328585961343,
      "grad_norm": 0.9741524457931519,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.858225405216217,
      "num_tokens": 632557630.0,
      "step": 3970
    },
    {
      "epoch": 2.0198372329603256,
      "grad_norm": 0.9519668221473694,
      "learning_rate": 1e-05,
      "loss": 0.431,
      "mean_token_accuracy": 0.8605380058288574,
      "num_tokens": 632721024.0,
      "step": 3971
    },
    {
      "epoch": 2.0203458799593084,
      "grad_norm": 0.9784402847290039,
      "learning_rate": 1e-05,
      "loss": 0.4625,
      "mean_token_accuracy": 0.8513401746749878,
      "num_tokens": 632888435.0,
      "step": 3972
    },
    {
      "epoch": 2.020854526958291,
      "grad_norm": 0.8853612542152405,
      "learning_rate": 1e-05,
      "loss": 0.4005,
      "mean_token_accuracy": 0.8683898448944092,
      "num_tokens": 633047095.0,
      "step": 3973
    },
    {
      "epoch": 2.0213631739572735,
      "grad_norm": 1.0621803998947144,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8565826416015625,
      "num_tokens": 633213083.0,
      "step": 3974
    },
    {
      "epoch": 2.0218718209562563,
      "grad_norm": 0.9953148365020752,
      "learning_rate": 1e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8580330610275269,
      "num_tokens": 633369616.0,
      "step": 3975
    },
    {
      "epoch": 2.022380467955239,
      "grad_norm": 0.9554272890090942,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8606266975402832,
      "num_tokens": 633521507.0,
      "step": 3976
    },
    {
      "epoch": 2.022889114954222,
      "grad_norm": 1.0419212579727173,
      "learning_rate": 1e-05,
      "loss": 0.4395,
      "mean_token_accuracy": 0.8574559688568115,
      "num_tokens": 633678242.0,
      "step": 3977
    },
    {
      "epoch": 2.0233977619532046,
      "grad_norm": 1.0358978509902954,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8566417694091797,
      "num_tokens": 633851293.0,
      "step": 3978
    },
    {
      "epoch": 2.023906408952187,
      "grad_norm": 1.0834808349609375,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8513150811195374,
      "num_tokens": 634019931.0,
      "step": 3979
    },
    {
      "epoch": 2.0244150559511698,
      "grad_norm": 0.8701474666595459,
      "learning_rate": 1e-05,
      "loss": 0.4333,
      "mean_token_accuracy": 0.859501838684082,
      "num_tokens": 634182314.0,
      "step": 3980
    },
    {
      "epoch": 2.0249237029501526,
      "grad_norm": 0.9976544976234436,
      "learning_rate": 1e-05,
      "loss": 0.4501,
      "mean_token_accuracy": 0.855332612991333,
      "num_tokens": 634351784.0,
      "step": 3981
    },
    {
      "epoch": 2.0254323499491353,
      "grad_norm": 0.9032160043716431,
      "learning_rate": 1e-05,
      "loss": 0.4203,
      "mean_token_accuracy": 0.8629529476165771,
      "num_tokens": 634503716.0,
      "step": 3982
    },
    {
      "epoch": 2.025940996948118,
      "grad_norm": 0.9581498503684998,
      "learning_rate": 1e-05,
      "loss": 0.4547,
      "mean_token_accuracy": 0.8519119024276733,
      "num_tokens": 634660553.0,
      "step": 3983
    },
    {
      "epoch": 2.026449643947101,
      "grad_norm": 0.9951674938201904,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8657628893852234,
      "num_tokens": 634812750.0,
      "step": 3984
    },
    {
      "epoch": 2.0269582909460833,
      "grad_norm": 1.1226887702941895,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8587206602096558,
      "num_tokens": 634964107.0,
      "step": 3985
    },
    {
      "epoch": 2.027466937945066,
      "grad_norm": 1.053794503211975,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8575358986854553,
      "num_tokens": 635124979.0,
      "step": 3986
    },
    {
      "epoch": 2.027975584944049,
      "grad_norm": 0.8636070489883423,
      "learning_rate": 1e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8556486964225769,
      "num_tokens": 635294168.0,
      "step": 3987
    },
    {
      "epoch": 2.0284842319430316,
      "grad_norm": 0.9470566511154175,
      "learning_rate": 1e-05,
      "loss": 0.4219,
      "mean_token_accuracy": 0.8628208041191101,
      "num_tokens": 635455215.0,
      "step": 3988
    },
    {
      "epoch": 2.0289928789420144,
      "grad_norm": 0.9437407851219177,
      "learning_rate": 1e-05,
      "loss": 0.451,
      "mean_token_accuracy": 0.8550064563751221,
      "num_tokens": 635615392.0,
      "step": 3989
    },
    {
      "epoch": 2.0295015259409968,
      "grad_norm": 0.887579083442688,
      "learning_rate": 1e-05,
      "loss": 0.4342,
      "mean_token_accuracy": 0.8611311912536621,
      "num_tokens": 635778771.0,
      "step": 3990
    },
    {
      "epoch": 2.0300101729399795,
      "grad_norm": 0.9505531191825867,
      "learning_rate": 1e-05,
      "loss": 0.4582,
      "mean_token_accuracy": 0.8537662029266357,
      "num_tokens": 635940023.0,
      "step": 3991
    },
    {
      "epoch": 2.0305188199389623,
      "grad_norm": 0.9255881309509277,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.8540250062942505,
      "num_tokens": 636091977.0,
      "step": 3992
    },
    {
      "epoch": 2.031027466937945,
      "grad_norm": 1.0332192182540894,
      "learning_rate": 1e-05,
      "loss": 0.4233,
      "mean_token_accuracy": 0.8634705543518066,
      "num_tokens": 636260118.0,
      "step": 3993
    },
    {
      "epoch": 2.031536113936928,
      "grad_norm": 0.9251944422721863,
      "learning_rate": 1e-05,
      "loss": 0.4222,
      "mean_token_accuracy": 0.862652063369751,
      "num_tokens": 636426068.0,
      "step": 3994
    },
    {
      "epoch": 2.0320447609359107,
      "grad_norm": 0.9779582619667053,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8561336994171143,
      "num_tokens": 636574295.0,
      "step": 3995
    },
    {
      "epoch": 2.032553407934893,
      "grad_norm": 0.9237884879112244,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8642343878746033,
      "num_tokens": 636728987.0,
      "step": 3996
    },
    {
      "epoch": 2.033062054933876,
      "grad_norm": 0.8933495283126831,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8535547256469727,
      "num_tokens": 636889748.0,
      "step": 3997
    },
    {
      "epoch": 2.0335707019328586,
      "grad_norm": 1.0079468488693237,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8520773649215698,
      "num_tokens": 637048170.0,
      "step": 3998
    },
    {
      "epoch": 2.0340793489318414,
      "grad_norm": 0.9072739481925964,
      "learning_rate": 1e-05,
      "loss": 0.4405,
      "mean_token_accuracy": 0.8564949631690979,
      "num_tokens": 637205011.0,
      "step": 3999
    },
    {
      "epoch": 2.034587995930824,
      "grad_norm": 1.0024240016937256,
      "learning_rate": 1e-05,
      "loss": 0.447,
      "mean_token_accuracy": 0.8577027320861816,
      "num_tokens": 637376842.0,
      "step": 4000
    },
    {
      "epoch": 2.0350966429298065,
      "grad_norm": 0.8983991146087646,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.8515585660934448,
      "num_tokens": 637544257.0,
      "step": 4001
    },
    {
      "epoch": 2.0356052899287893,
      "grad_norm": 0.9078333973884583,
      "learning_rate": 1e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8519089818000793,
      "num_tokens": 637708776.0,
      "step": 4002
    },
    {
      "epoch": 2.036113936927772,
      "grad_norm": 1.111275315284729,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8573497533798218,
      "num_tokens": 637866051.0,
      "step": 4003
    },
    {
      "epoch": 2.036622583926755,
      "grad_norm": 0.9840702414512634,
      "learning_rate": 1e-05,
      "loss": 0.4424,
      "mean_token_accuracy": 0.857101559638977,
      "num_tokens": 638026761.0,
      "step": 4004
    },
    {
      "epoch": 2.0371312309257377,
      "grad_norm": 0.9235756397247314,
      "learning_rate": 1e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.8585684299468994,
      "num_tokens": 638183663.0,
      "step": 4005
    },
    {
      "epoch": 2.0376398779247205,
      "grad_norm": 1.0132715702056885,
      "learning_rate": 1e-05,
      "loss": 0.4557,
      "mean_token_accuracy": 0.8540298938751221,
      "num_tokens": 638337687.0,
      "step": 4006
    },
    {
      "epoch": 2.038148524923703,
      "grad_norm": 0.9375067353248596,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "mean_token_accuracy": 0.8654524087905884,
      "num_tokens": 638500144.0,
      "step": 4007
    },
    {
      "epoch": 2.0386571719226856,
      "grad_norm": 0.9727280139923096,
      "learning_rate": 1e-05,
      "loss": 0.4133,
      "mean_token_accuracy": 0.8654862642288208,
      "num_tokens": 638663203.0,
      "step": 4008
    },
    {
      "epoch": 2.0391658189216684,
      "grad_norm": 1.001802682876587,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.8518264293670654,
      "num_tokens": 638816564.0,
      "step": 4009
    },
    {
      "epoch": 2.039674465920651,
      "grad_norm": 0.9148018956184387,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.860122561454773,
      "num_tokens": 638985653.0,
      "step": 4010
    },
    {
      "epoch": 2.040183112919634,
      "grad_norm": 1.015397548675537,
      "learning_rate": 1e-05,
      "loss": 0.4381,
      "mean_token_accuracy": 0.857643187046051,
      "num_tokens": 639144412.0,
      "step": 4011
    },
    {
      "epoch": 2.0406917599186163,
      "grad_norm": 0.9982504844665527,
      "learning_rate": 1e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8505040407180786,
      "num_tokens": 639296652.0,
      "step": 4012
    },
    {
      "epoch": 2.041200406917599,
      "grad_norm": 1.0248266458511353,
      "learning_rate": 1e-05,
      "loss": 0.4815,
      "mean_token_accuracy": 0.8467176556587219,
      "num_tokens": 639467379.0,
      "step": 4013
    },
    {
      "epoch": 2.041709053916582,
      "grad_norm": 0.9136324524879456,
      "learning_rate": 1e-05,
      "loss": 0.4248,
      "mean_token_accuracy": 0.8617210388183594,
      "num_tokens": 639629812.0,
      "step": 4014
    },
    {
      "epoch": 2.0422177009155646,
      "grad_norm": 0.9851207137107849,
      "learning_rate": 1e-05,
      "loss": 0.3942,
      "mean_token_accuracy": 0.871574878692627,
      "num_tokens": 639790695.0,
      "step": 4015
    },
    {
      "epoch": 2.0427263479145474,
      "grad_norm": 0.9477247595787048,
      "learning_rate": 1e-05,
      "loss": 0.4488,
      "mean_token_accuracy": 0.855469822883606,
      "num_tokens": 639953979.0,
      "step": 4016
    },
    {
      "epoch": 2.0432349949135302,
      "grad_norm": 1.1952741146087646,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8568698763847351,
      "num_tokens": 640110677.0,
      "step": 4017
    },
    {
      "epoch": 2.0437436419125126,
      "grad_norm": 0.9191387295722961,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "mean_token_accuracy": 0.8653135299682617,
      "num_tokens": 640271057.0,
      "step": 4018
    },
    {
      "epoch": 2.0442522889114954,
      "grad_norm": 0.886969268321991,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "mean_token_accuracy": 0.862865686416626,
      "num_tokens": 640421336.0,
      "step": 4019
    },
    {
      "epoch": 2.044760935910478,
      "grad_norm": 0.9596548080444336,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8615579605102539,
      "num_tokens": 640574916.0,
      "step": 4020
    },
    {
      "epoch": 2.045269582909461,
      "grad_norm": 0.9139665365219116,
      "learning_rate": 1e-05,
      "loss": 0.4235,
      "mean_token_accuracy": 0.8608518838882446,
      "num_tokens": 640731484.0,
      "step": 4021
    },
    {
      "epoch": 2.0457782299084437,
      "grad_norm": 0.9046313166618347,
      "learning_rate": 1e-05,
      "loss": 0.4313,
      "mean_token_accuracy": 0.8604253530502319,
      "num_tokens": 640899252.0,
      "step": 4022
    },
    {
      "epoch": 2.046286876907426,
      "grad_norm": 0.9831649661064148,
      "learning_rate": 1e-05,
      "loss": 0.4448,
      "mean_token_accuracy": 0.8558986186981201,
      "num_tokens": 641049998.0,
      "step": 4023
    },
    {
      "epoch": 2.046795523906409,
      "grad_norm": 0.9268972873687744,
      "learning_rate": 1e-05,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8665068745613098,
      "num_tokens": 641225765.0,
      "step": 4024
    },
    {
      "epoch": 2.0473041709053916,
      "grad_norm": 1.0772287845611572,
      "learning_rate": 1e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8591381311416626,
      "num_tokens": 641395261.0,
      "step": 4025
    },
    {
      "epoch": 2.0478128179043744,
      "grad_norm": 0.9058048129081726,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8644649982452393,
      "num_tokens": 641553952.0,
      "step": 4026
    },
    {
      "epoch": 2.048321464903357,
      "grad_norm": 0.9972060322761536,
      "learning_rate": 1e-05,
      "loss": 0.4822,
      "mean_token_accuracy": 0.8465654253959656,
      "num_tokens": 641713306.0,
      "step": 4027
    },
    {
      "epoch": 2.04883011190234,
      "grad_norm": 0.9957790970802307,
      "learning_rate": 1e-05,
      "loss": 0.4189,
      "mean_token_accuracy": 0.862214207649231,
      "num_tokens": 641871688.0,
      "step": 4028
    },
    {
      "epoch": 2.0493387589013223,
      "grad_norm": 0.9785782098770142,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8533655405044556,
      "num_tokens": 642028112.0,
      "step": 4029
    },
    {
      "epoch": 2.049847405900305,
      "grad_norm": 1.0586796998977661,
      "learning_rate": 1e-05,
      "loss": 0.4646,
      "mean_token_accuracy": 0.8501312136650085,
      "num_tokens": 642182839.0,
      "step": 4030
    },
    {
      "epoch": 2.050356052899288,
      "grad_norm": 0.9724810719490051,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8565281629562378,
      "num_tokens": 642333528.0,
      "step": 4031
    },
    {
      "epoch": 2.0508646998982707,
      "grad_norm": 1.0518903732299805,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.8530091047286987,
      "num_tokens": 642490309.0,
      "step": 4032
    },
    {
      "epoch": 2.0513733468972535,
      "grad_norm": 0.9552509784698486,
      "learning_rate": 1e-05,
      "loss": 0.4149,
      "mean_token_accuracy": 0.8634740114212036,
      "num_tokens": 642640172.0,
      "step": 4033
    },
    {
      "epoch": 2.051881993896236,
      "grad_norm": 1.0244795083999634,
      "learning_rate": 1e-05,
      "loss": 0.4371,
      "mean_token_accuracy": 0.8561656475067139,
      "num_tokens": 642789777.0,
      "step": 4034
    },
    {
      "epoch": 2.0523906408952186,
      "grad_norm": 1.0393922328948975,
      "learning_rate": 1e-05,
      "loss": 0.4637,
      "mean_token_accuracy": 0.8515689969062805,
      "num_tokens": 642951366.0,
      "step": 4035
    },
    {
      "epoch": 2.0528992878942014,
      "grad_norm": 0.9119642972946167,
      "learning_rate": 1e-05,
      "loss": 0.4285,
      "mean_token_accuracy": 0.8608314990997314,
      "num_tokens": 643121797.0,
      "step": 4036
    },
    {
      "epoch": 2.053407934893184,
      "grad_norm": 1.0008642673492432,
      "learning_rate": 1e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8494364023208618,
      "num_tokens": 643279667.0,
      "step": 4037
    },
    {
      "epoch": 2.053916581892167,
      "grad_norm": 0.9313419461250305,
      "learning_rate": 1e-05,
      "loss": 0.4313,
      "mean_token_accuracy": 0.8611078262329102,
      "num_tokens": 643442233.0,
      "step": 4038
    },
    {
      "epoch": 2.0544252288911498,
      "grad_norm": 0.9488298892974854,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8599388003349304,
      "num_tokens": 643614205.0,
      "step": 4039
    },
    {
      "epoch": 2.054933875890132,
      "grad_norm": 0.9752569198608398,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8609519004821777,
      "num_tokens": 643770297.0,
      "step": 4040
    },
    {
      "epoch": 2.055442522889115,
      "grad_norm": 0.9722061157226562,
      "learning_rate": 1e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8619430065155029,
      "num_tokens": 643916346.0,
      "step": 4041
    },
    {
      "epoch": 2.0559511698880977,
      "grad_norm": 1.0140228271484375,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8532426357269287,
      "num_tokens": 644077898.0,
      "step": 4042
    },
    {
      "epoch": 2.0564598168870805,
      "grad_norm": 0.9496407508850098,
      "learning_rate": 1e-05,
      "loss": 0.4408,
      "mean_token_accuracy": 0.8577236533164978,
      "num_tokens": 644239431.0,
      "step": 4043
    },
    {
      "epoch": 2.0569684638860632,
      "grad_norm": 1.0077439546585083,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.8524376153945923,
      "num_tokens": 644396823.0,
      "step": 4044
    },
    {
      "epoch": 2.0574771108850456,
      "grad_norm": 0.8989365100860596,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.8576108813285828,
      "num_tokens": 644551897.0,
      "step": 4045
    },
    {
      "epoch": 2.0579857578840284,
      "grad_norm": 0.9695183634757996,
      "learning_rate": 1e-05,
      "loss": 0.4428,
      "mean_token_accuracy": 0.8565598130226135,
      "num_tokens": 644715765.0,
      "step": 4046
    },
    {
      "epoch": 2.058494404883011,
      "grad_norm": 0.9331927299499512,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "mean_token_accuracy": 0.8558671474456787,
      "num_tokens": 644885356.0,
      "step": 4047
    },
    {
      "epoch": 2.059003051881994,
      "grad_norm": 0.9889021515846252,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8565429449081421,
      "num_tokens": 645039148.0,
      "step": 4048
    },
    {
      "epoch": 2.0595116988809767,
      "grad_norm": 0.9650540351867676,
      "learning_rate": 1e-05,
      "loss": 0.4473,
      "mean_token_accuracy": 0.8539718389511108,
      "num_tokens": 645202089.0,
      "step": 4049
    },
    {
      "epoch": 2.0600203458799595,
      "grad_norm": 0.9073891043663025,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.859544038772583,
      "num_tokens": 645359940.0,
      "step": 4050
    },
    {
      "epoch": 2.060528992878942,
      "grad_norm": 0.9147620797157288,
      "learning_rate": 1e-05,
      "loss": 0.4402,
      "mean_token_accuracy": 0.8565871715545654,
      "num_tokens": 645530000.0,
      "step": 4051
    },
    {
      "epoch": 2.0610376398779247,
      "grad_norm": 0.9782045483589172,
      "learning_rate": 1e-05,
      "loss": 0.4254,
      "mean_token_accuracy": 0.8615814447402954,
      "num_tokens": 645695713.0,
      "step": 4052
    },
    {
      "epoch": 2.0615462868769074,
      "grad_norm": 1.0421298742294312,
      "learning_rate": 1e-05,
      "loss": 0.4073,
      "mean_token_accuracy": 0.8664171695709229,
      "num_tokens": 645845526.0,
      "step": 4053
    },
    {
      "epoch": 2.0620549338758902,
      "grad_norm": 0.9888449311256409,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8537807464599609,
      "num_tokens": 646007060.0,
      "step": 4054
    },
    {
      "epoch": 2.062563580874873,
      "grad_norm": 1.046412467956543,
      "learning_rate": 1e-05,
      "loss": 0.4361,
      "mean_token_accuracy": 0.8585008382797241,
      "num_tokens": 646158897.0,
      "step": 4055
    },
    {
      "epoch": 2.0630722278738554,
      "grad_norm": 1.0488885641098022,
      "learning_rate": 1e-05,
      "loss": 0.4464,
      "mean_token_accuracy": 0.8543227910995483,
      "num_tokens": 646310789.0,
      "step": 4056
    },
    {
      "epoch": 2.063580874872838,
      "grad_norm": 0.9440844655036926,
      "learning_rate": 1e-05,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8615769147872925,
      "num_tokens": 646489056.0,
      "step": 4057
    },
    {
      "epoch": 2.064089521871821,
      "grad_norm": 1.0296015739440918,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8658794164657593,
      "num_tokens": 646651450.0,
      "step": 4058
    },
    {
      "epoch": 2.0645981688708037,
      "grad_norm": 1.0239273309707642,
      "learning_rate": 1e-05,
      "loss": 0.4552,
      "mean_token_accuracy": 0.8538615703582764,
      "num_tokens": 646808789.0,
      "step": 4059
    },
    {
      "epoch": 2.0651068158697865,
      "grad_norm": 0.9998471736907959,
      "learning_rate": 1e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8558937907218933,
      "num_tokens": 646974846.0,
      "step": 4060
    },
    {
      "epoch": 2.0656154628687693,
      "grad_norm": 0.9561287760734558,
      "learning_rate": 1e-05,
      "loss": 0.4152,
      "mean_token_accuracy": 0.8642705082893372,
      "num_tokens": 647131712.0,
      "step": 4061
    },
    {
      "epoch": 2.0661241098677516,
      "grad_norm": 1.3276989459991455,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8598711490631104,
      "num_tokens": 647294337.0,
      "step": 4062
    },
    {
      "epoch": 2.0666327568667344,
      "grad_norm": 1.076204776763916,
      "learning_rate": 1e-05,
      "loss": 0.4626,
      "mean_token_accuracy": 0.8524924516677856,
      "num_tokens": 647453242.0,
      "step": 4063
    },
    {
      "epoch": 2.067141403865717,
      "grad_norm": 0.8968902230262756,
      "learning_rate": 1e-05,
      "loss": 0.4183,
      "mean_token_accuracy": 0.863444447517395,
      "num_tokens": 647614724.0,
      "step": 4064
    },
    {
      "epoch": 2.0676500508647,
      "grad_norm": 1.007246732711792,
      "learning_rate": 1e-05,
      "loss": 0.4166,
      "mean_token_accuracy": 0.8639742732048035,
      "num_tokens": 647774121.0,
      "step": 4065
    },
    {
      "epoch": 2.068158697863683,
      "grad_norm": 0.9755057692527771,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.855438232421875,
      "num_tokens": 647933285.0,
      "step": 4066
    },
    {
      "epoch": 2.068667344862665,
      "grad_norm": 0.9740897417068481,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8596476316452026,
      "num_tokens": 648098194.0,
      "step": 4067
    },
    {
      "epoch": 2.069175991861648,
      "grad_norm": 0.9107229113578796,
      "learning_rate": 1e-05,
      "loss": 0.4188,
      "mean_token_accuracy": 0.8632190227508545,
      "num_tokens": 648251173.0,
      "step": 4068
    },
    {
      "epoch": 2.0696846388606307,
      "grad_norm": 1.126113772392273,
      "learning_rate": 1e-05,
      "loss": 0.4806,
      "mean_token_accuracy": 0.845462441444397,
      "num_tokens": 648398627.0,
      "step": 4069
    },
    {
      "epoch": 2.0701932858596135,
      "grad_norm": 1.0377577543258667,
      "learning_rate": 1e-05,
      "loss": 0.4416,
      "mean_token_accuracy": 0.8560168743133545,
      "num_tokens": 648551725.0,
      "step": 4070
    },
    {
      "epoch": 2.0707019328585963,
      "grad_norm": 0.9613800048828125,
      "learning_rate": 1e-05,
      "loss": 0.4529,
      "mean_token_accuracy": 0.8547857999801636,
      "num_tokens": 648708382.0,
      "step": 4071
    },
    {
      "epoch": 2.0712105798575786,
      "grad_norm": 1.0431551933288574,
      "learning_rate": 1e-05,
      "loss": 0.4177,
      "mean_token_accuracy": 0.8637215495109558,
      "num_tokens": 648863508.0,
      "step": 4072
    },
    {
      "epoch": 2.0717192268565614,
      "grad_norm": 0.9346727132797241,
      "learning_rate": 1e-05,
      "loss": 0.468,
      "mean_token_accuracy": 0.8500462770462036,
      "num_tokens": 649012495.0,
      "step": 4073
    },
    {
      "epoch": 2.072227873855544,
      "grad_norm": 0.8879013061523438,
      "learning_rate": 1e-05,
      "loss": 0.3821,
      "mean_token_accuracy": 0.8734450340270996,
      "num_tokens": 649176493.0,
      "step": 4074
    },
    {
      "epoch": 2.072736520854527,
      "grad_norm": 0.966341495513916,
      "learning_rate": 1e-05,
      "loss": 0.415,
      "mean_token_accuracy": 0.8645215034484863,
      "num_tokens": 649334145.0,
      "step": 4075
    },
    {
      "epoch": 2.0732451678535098,
      "grad_norm": 0.8374088406562805,
      "learning_rate": 1e-05,
      "loss": 0.4019,
      "mean_token_accuracy": 0.8681143522262573,
      "num_tokens": 649497977.0,
      "step": 4076
    },
    {
      "epoch": 2.0737538148524925,
      "grad_norm": 0.9355109930038452,
      "learning_rate": 1e-05,
      "loss": 0.4626,
      "mean_token_accuracy": 0.8502155542373657,
      "num_tokens": 649665574.0,
      "step": 4077
    },
    {
      "epoch": 2.074262461851475,
      "grad_norm": 0.9582520723342896,
      "learning_rate": 1e-05,
      "loss": 0.4244,
      "mean_token_accuracy": 0.8623635768890381,
      "num_tokens": 649826804.0,
      "step": 4078
    },
    {
      "epoch": 2.0747711088504577,
      "grad_norm": 0.9194473624229431,
      "learning_rate": 1e-05,
      "loss": 0.4076,
      "mean_token_accuracy": 0.8666621446609497,
      "num_tokens": 649979766.0,
      "step": 4079
    },
    {
      "epoch": 2.0752797558494405,
      "grad_norm": 0.9332849383354187,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.857512891292572,
      "num_tokens": 650125023.0,
      "step": 4080
    },
    {
      "epoch": 2.0757884028484233,
      "grad_norm": 0.9183851480484009,
      "learning_rate": 1e-05,
      "loss": 0.4326,
      "mean_token_accuracy": 0.8598358631134033,
      "num_tokens": 650285291.0,
      "step": 4081
    },
    {
      "epoch": 2.076297049847406,
      "grad_norm": 0.8456012606620789,
      "learning_rate": 1e-05,
      "loss": 0.4216,
      "mean_token_accuracy": 0.8641607761383057,
      "num_tokens": 650457502.0,
      "step": 4082
    },
    {
      "epoch": 2.076805696846389,
      "grad_norm": 0.8970609903335571,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.859481155872345,
      "num_tokens": 650618262.0,
      "step": 4083
    },
    {
      "epoch": 2.077314343845371,
      "grad_norm": 0.8981460928916931,
      "learning_rate": 1e-05,
      "loss": 0.4358,
      "mean_token_accuracy": 0.8589435815811157,
      "num_tokens": 650774229.0,
      "step": 4084
    },
    {
      "epoch": 2.077822990844354,
      "grad_norm": 0.9190006852149963,
      "learning_rate": 1e-05,
      "loss": 0.4386,
      "mean_token_accuracy": 0.8579449653625488,
      "num_tokens": 650932968.0,
      "step": 4085
    },
    {
      "epoch": 2.0783316378433367,
      "grad_norm": 0.960010290145874,
      "learning_rate": 1e-05,
      "loss": 0.4584,
      "mean_token_accuracy": 0.8517380356788635,
      "num_tokens": 651094235.0,
      "step": 4086
    },
    {
      "epoch": 2.0788402848423195,
      "grad_norm": 1.0507935285568237,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8477388620376587,
      "num_tokens": 651247051.0,
      "step": 4087
    },
    {
      "epoch": 2.0793489318413023,
      "grad_norm": 0.8672595024108887,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.8619122505187988,
      "num_tokens": 651403530.0,
      "step": 4088
    },
    {
      "epoch": 2.0798575788402847,
      "grad_norm": 0.8593379855155945,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.8561261892318726,
      "num_tokens": 651579353.0,
      "step": 4089
    },
    {
      "epoch": 2.0803662258392674,
      "grad_norm": 0.9281165599822998,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.852559506893158,
      "num_tokens": 651737149.0,
      "step": 4090
    },
    {
      "epoch": 2.0808748728382502,
      "grad_norm": 0.9167788028717041,
      "learning_rate": 1e-05,
      "loss": 0.4347,
      "mean_token_accuracy": 0.8594133257865906,
      "num_tokens": 651896201.0,
      "step": 4091
    },
    {
      "epoch": 2.081383519837233,
      "grad_norm": 1.0042438507080078,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.854072630405426,
      "num_tokens": 652057532.0,
      "step": 4092
    },
    {
      "epoch": 2.081892166836216,
      "grad_norm": 0.8949097990989685,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8516091704368591,
      "num_tokens": 652209462.0,
      "step": 4093
    },
    {
      "epoch": 2.082400813835198,
      "grad_norm": 0.9609045386314392,
      "learning_rate": 1e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8596907258033752,
      "num_tokens": 652364933.0,
      "step": 4094
    },
    {
      "epoch": 2.082909460834181,
      "grad_norm": 0.902187705039978,
      "learning_rate": 1e-05,
      "loss": 0.4317,
      "mean_token_accuracy": 0.8602837324142456,
      "num_tokens": 652537736.0,
      "step": 4095
    },
    {
      "epoch": 2.0834181078331637,
      "grad_norm": 0.9226665496826172,
      "learning_rate": 1e-05,
      "loss": 0.4525,
      "mean_token_accuracy": 0.8551701307296753,
      "num_tokens": 652693646.0,
      "step": 4096
    },
    {
      "epoch": 2.0839267548321465,
      "grad_norm": 0.9428882002830505,
      "learning_rate": 1e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8616924285888672,
      "num_tokens": 652848681.0,
      "step": 4097
    },
    {
      "epoch": 2.0844354018311293,
      "grad_norm": 0.8364573121070862,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "mean_token_accuracy": 0.8645583391189575,
      "num_tokens": 653007977.0,
      "step": 4098
    },
    {
      "epoch": 2.084944048830112,
      "grad_norm": 0.9261143207550049,
      "learning_rate": 1e-05,
      "loss": 0.4133,
      "mean_token_accuracy": 0.866054117679596,
      "num_tokens": 653154060.0,
      "step": 4099
    },
    {
      "epoch": 2.0854526958290944,
      "grad_norm": 0.966831386089325,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8580394387245178,
      "num_tokens": 653315197.0,
      "step": 4100
    },
    {
      "epoch": 2.085961342828077,
      "grad_norm": 0.850862979888916,
      "learning_rate": 1e-05,
      "loss": 0.4063,
      "mean_token_accuracy": 0.8678941130638123,
      "num_tokens": 653474119.0,
      "step": 4101
    },
    {
      "epoch": 2.08646998982706,
      "grad_norm": 0.9313889741897583,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8533434867858887,
      "num_tokens": 653634130.0,
      "step": 4102
    },
    {
      "epoch": 2.086978636826043,
      "grad_norm": 0.8664780259132385,
      "learning_rate": 1e-05,
      "loss": 0.4169,
      "mean_token_accuracy": 0.8633689284324646,
      "num_tokens": 653796634.0,
      "step": 4103
    },
    {
      "epoch": 2.0874872838250256,
      "grad_norm": 0.9221550822257996,
      "learning_rate": 1e-05,
      "loss": 0.4875,
      "mean_token_accuracy": 0.8431581258773804,
      "num_tokens": 653956123.0,
      "step": 4104
    },
    {
      "epoch": 2.0879959308240084,
      "grad_norm": 0.9604993462562561,
      "learning_rate": 1e-05,
      "loss": 0.4909,
      "mean_token_accuracy": 0.8433293104171753,
      "num_tokens": 654110242.0,
      "step": 4105
    },
    {
      "epoch": 2.0885045778229907,
      "grad_norm": 0.8623951077461243,
      "learning_rate": 1e-05,
      "loss": 0.4236,
      "mean_token_accuracy": 0.8634408712387085,
      "num_tokens": 654284883.0,
      "step": 4106
    },
    {
      "epoch": 2.0890132248219735,
      "grad_norm": 0.893561840057373,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8487818837165833,
      "num_tokens": 654451207.0,
      "step": 4107
    },
    {
      "epoch": 2.0895218718209563,
      "grad_norm": 0.8882175087928772,
      "learning_rate": 1e-05,
      "loss": 0.4679,
      "mean_token_accuracy": 0.8507075905799866,
      "num_tokens": 654620877.0,
      "step": 4108
    },
    {
      "epoch": 2.090030518819939,
      "grad_norm": 0.8599258661270142,
      "learning_rate": 1e-05,
      "loss": 0.4693,
      "mean_token_accuracy": 0.8494659662246704,
      "num_tokens": 654789815.0,
      "step": 4109
    },
    {
      "epoch": 2.090539165818922,
      "grad_norm": 0.869370698928833,
      "learning_rate": 1e-05,
      "loss": 0.4545,
      "mean_token_accuracy": 0.8524019122123718,
      "num_tokens": 654955083.0,
      "step": 4110
    },
    {
      "epoch": 2.091047812817904,
      "grad_norm": 0.9025420546531677,
      "learning_rate": 1e-05,
      "loss": 0.4118,
      "mean_token_accuracy": 0.865307092666626,
      "num_tokens": 655105899.0,
      "step": 4111
    },
    {
      "epoch": 2.091556459816887,
      "grad_norm": 0.8740807175636292,
      "learning_rate": 1e-05,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8631594181060791,
      "num_tokens": 655262947.0,
      "step": 4112
    },
    {
      "epoch": 2.0920651068158698,
      "grad_norm": 0.8836954832077026,
      "learning_rate": 1e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.8628745079040527,
      "num_tokens": 655424948.0,
      "step": 4113
    },
    {
      "epoch": 2.0925737538148526,
      "grad_norm": 0.9837310314178467,
      "learning_rate": 1e-05,
      "loss": 0.4239,
      "mean_token_accuracy": 0.8626061677932739,
      "num_tokens": 655582982.0,
      "step": 4114
    },
    {
      "epoch": 2.0930824008138353,
      "grad_norm": 0.9199106097221375,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8546483516693115,
      "num_tokens": 655736818.0,
      "step": 4115
    },
    {
      "epoch": 2.0935910478128177,
      "grad_norm": 0.9366452097892761,
      "learning_rate": 1e-05,
      "loss": 0.4269,
      "mean_token_accuracy": 0.860552191734314,
      "num_tokens": 655892296.0,
      "step": 4116
    },
    {
      "epoch": 2.0940996948118005,
      "grad_norm": 0.9860832691192627,
      "learning_rate": 1e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.859500527381897,
      "num_tokens": 656042630.0,
      "step": 4117
    },
    {
      "epoch": 2.0946083418107833,
      "grad_norm": 0.8909306526184082,
      "learning_rate": 1e-05,
      "loss": 0.4208,
      "mean_token_accuracy": 0.8629220128059387,
      "num_tokens": 656199760.0,
      "step": 4118
    },
    {
      "epoch": 2.095116988809766,
      "grad_norm": 0.9349894523620605,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "mean_token_accuracy": 0.8533021211624146,
      "num_tokens": 656356098.0,
      "step": 4119
    },
    {
      "epoch": 2.095625635808749,
      "grad_norm": 0.9588991403579712,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8521022796630859,
      "num_tokens": 656514419.0,
      "step": 4120
    },
    {
      "epoch": 2.0961342828077316,
      "grad_norm": 0.8751233816146851,
      "learning_rate": 1e-05,
      "loss": 0.3911,
      "mean_token_accuracy": 0.8713945150375366,
      "num_tokens": 656661884.0,
      "step": 4121
    },
    {
      "epoch": 2.096642929806714,
      "grad_norm": 0.8489260077476501,
      "learning_rate": 1e-05,
      "loss": 0.4073,
      "mean_token_accuracy": 0.8669377565383911,
      "num_tokens": 656815650.0,
      "step": 4122
    },
    {
      "epoch": 2.0971515768056967,
      "grad_norm": 0.9509919881820679,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8608969449996948,
      "num_tokens": 656980239.0,
      "step": 4123
    },
    {
      "epoch": 2.0976602238046795,
      "grad_norm": 0.9359029531478882,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8602707982063293,
      "num_tokens": 657128763.0,
      "step": 4124
    },
    {
      "epoch": 2.0981688708036623,
      "grad_norm": 0.8939815759658813,
      "learning_rate": 1e-05,
      "loss": 0.4478,
      "mean_token_accuracy": 0.8559001088142395,
      "num_tokens": 657290891.0,
      "step": 4125
    },
    {
      "epoch": 2.098677517802645,
      "grad_norm": 0.9452753067016602,
      "learning_rate": 1e-05,
      "loss": 0.4483,
      "mean_token_accuracy": 0.8554445505142212,
      "num_tokens": 657442965.0,
      "step": 4126
    },
    {
      "epoch": 2.099186164801628,
      "grad_norm": 0.832610011100769,
      "learning_rate": 1e-05,
      "loss": 0.4376,
      "mean_token_accuracy": 0.8597967624664307,
      "num_tokens": 657627642.0,
      "step": 4127
    },
    {
      "epoch": 2.0996948118006102,
      "grad_norm": 0.9224095344543457,
      "learning_rate": 1e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8613498210906982,
      "num_tokens": 657787301.0,
      "step": 4128
    },
    {
      "epoch": 2.100203458799593,
      "grad_norm": 0.9500312805175781,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.859756350517273,
      "num_tokens": 657932879.0,
      "step": 4129
    },
    {
      "epoch": 2.100712105798576,
      "grad_norm": 0.9640877842903137,
      "learning_rate": 1e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8614565134048462,
      "num_tokens": 658086444.0,
      "step": 4130
    },
    {
      "epoch": 2.1012207527975586,
      "grad_norm": 0.9052125215530396,
      "learning_rate": 1e-05,
      "loss": 0.4639,
      "mean_token_accuracy": 0.8509527444839478,
      "num_tokens": 658255960.0,
      "step": 4131
    },
    {
      "epoch": 2.1017293997965414,
      "grad_norm": 0.9720301628112793,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8574552536010742,
      "num_tokens": 658417852.0,
      "step": 4132
    },
    {
      "epoch": 2.1022380467955237,
      "grad_norm": 0.9318115711212158,
      "learning_rate": 1e-05,
      "loss": 0.4523,
      "mean_token_accuracy": 0.8530559539794922,
      "num_tokens": 658574716.0,
      "step": 4133
    },
    {
      "epoch": 2.1027466937945065,
      "grad_norm": 0.9879447817802429,
      "learning_rate": 1e-05,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8617345094680786,
      "num_tokens": 658727698.0,
      "step": 4134
    },
    {
      "epoch": 2.1032553407934893,
      "grad_norm": 0.9011422991752625,
      "learning_rate": 1e-05,
      "loss": 0.4256,
      "mean_token_accuracy": 0.8616060018539429,
      "num_tokens": 658885875.0,
      "step": 4135
    },
    {
      "epoch": 2.103763987792472,
      "grad_norm": 0.9072966575622559,
      "learning_rate": 1e-05,
      "loss": 0.4321,
      "mean_token_accuracy": 0.860308825969696,
      "num_tokens": 659036589.0,
      "step": 4136
    },
    {
      "epoch": 2.104272634791455,
      "grad_norm": 0.9436938166618347,
      "learning_rate": 1e-05,
      "loss": 0.4455,
      "mean_token_accuracy": 0.8565719127655029,
      "num_tokens": 659196714.0,
      "step": 4137
    },
    {
      "epoch": 2.104781281790437,
      "grad_norm": 0.9317730069160461,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8525628447532654,
      "num_tokens": 659356387.0,
      "step": 4138
    },
    {
      "epoch": 2.10528992878942,
      "grad_norm": 0.8680718541145325,
      "learning_rate": 1e-05,
      "loss": 0.4282,
      "mean_token_accuracy": 0.8603150844573975,
      "num_tokens": 659511890.0,
      "step": 4139
    },
    {
      "epoch": 2.105798575788403,
      "grad_norm": 0.9945595264434814,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.8552508354187012,
      "num_tokens": 659665573.0,
      "step": 4140
    },
    {
      "epoch": 2.1063072227873856,
      "grad_norm": 1.0013288259506226,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8601479530334473,
      "num_tokens": 659823114.0,
      "step": 4141
    },
    {
      "epoch": 2.1068158697863684,
      "grad_norm": 0.91698157787323,
      "learning_rate": 1e-05,
      "loss": 0.4518,
      "mean_token_accuracy": 0.855778694152832,
      "num_tokens": 659986298.0,
      "step": 4142
    },
    {
      "epoch": 2.107324516785351,
      "grad_norm": 0.9010708332061768,
      "learning_rate": 1e-05,
      "loss": 0.4145,
      "mean_token_accuracy": 0.8645851612091064,
      "num_tokens": 660144890.0,
      "step": 4143
    },
    {
      "epoch": 2.1078331637843335,
      "grad_norm": 0.8915697932243347,
      "learning_rate": 1e-05,
      "loss": 0.4248,
      "mean_token_accuracy": 0.8608888387680054,
      "num_tokens": 660295004.0,
      "step": 4144
    },
    {
      "epoch": 2.1083418107833163,
      "grad_norm": 0.9084402322769165,
      "learning_rate": 1e-05,
      "loss": 0.4168,
      "mean_token_accuracy": 0.8644142150878906,
      "num_tokens": 660447649.0,
      "step": 4145
    },
    {
      "epoch": 2.108850457782299,
      "grad_norm": 0.9173476696014404,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8596066236495972,
      "num_tokens": 660607411.0,
      "step": 4146
    },
    {
      "epoch": 2.109359104781282,
      "grad_norm": 0.9064167737960815,
      "learning_rate": 1e-05,
      "loss": 0.4361,
      "mean_token_accuracy": 0.8588464260101318,
      "num_tokens": 660761477.0,
      "step": 4147
    },
    {
      "epoch": 2.1098677517802646,
      "grad_norm": 0.9110246896743774,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8663687705993652,
      "num_tokens": 660921469.0,
      "step": 4148
    },
    {
      "epoch": 2.1103763987792474,
      "grad_norm": 0.9070311784744263,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8485805988311768,
      "num_tokens": 661082876.0,
      "step": 4149
    },
    {
      "epoch": 2.1108850457782298,
      "grad_norm": 0.9548289179801941,
      "learning_rate": 1e-05,
      "loss": 0.4772,
      "mean_token_accuracy": 0.8454893231391907,
      "num_tokens": 661237468.0,
      "step": 4150
    },
    {
      "epoch": 2.1113936927772126,
      "grad_norm": 0.8961446285247803,
      "learning_rate": 1e-05,
      "loss": 0.435,
      "mean_token_accuracy": 0.859317421913147,
      "num_tokens": 661405065.0,
      "step": 4151
    },
    {
      "epoch": 2.1119023397761953,
      "grad_norm": 0.9669830203056335,
      "learning_rate": 1e-05,
      "loss": 0.4198,
      "mean_token_accuracy": 0.8659858703613281,
      "num_tokens": 661553561.0,
      "step": 4152
    },
    {
      "epoch": 2.112410986775178,
      "grad_norm": 0.9334442615509033,
      "learning_rate": 1e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.8505697250366211,
      "num_tokens": 661711645.0,
      "step": 4153
    },
    {
      "epoch": 2.112919633774161,
      "grad_norm": 1.0024757385253906,
      "learning_rate": 1e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.8586753606796265,
      "num_tokens": 661865294.0,
      "step": 4154
    },
    {
      "epoch": 2.1134282807731433,
      "grad_norm": 0.8926826119422913,
      "learning_rate": 1e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.8604754209518433,
      "num_tokens": 662026583.0,
      "step": 4155
    },
    {
      "epoch": 2.113936927772126,
      "grad_norm": 0.9363488554954529,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8561187982559204,
      "num_tokens": 662192533.0,
      "step": 4156
    },
    {
      "epoch": 2.114445574771109,
      "grad_norm": 0.9798843264579773,
      "learning_rate": 1e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.8614747524261475,
      "num_tokens": 662346698.0,
      "step": 4157
    },
    {
      "epoch": 2.1149542217700916,
      "grad_norm": 0.9556357264518738,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8487550616264343,
      "num_tokens": 662502572.0,
      "step": 4158
    },
    {
      "epoch": 2.1154628687690744,
      "grad_norm": 0.938748836517334,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.8568930625915527,
      "num_tokens": 662658294.0,
      "step": 4159
    },
    {
      "epoch": 2.1159715157680568,
      "grad_norm": 1.0129530429840088,
      "learning_rate": 1e-05,
      "loss": 0.4735,
      "mean_token_accuracy": 0.8483498692512512,
      "num_tokens": 662821554.0,
      "step": 4160
    },
    {
      "epoch": 2.1164801627670395,
      "grad_norm": 0.9006751179695129,
      "learning_rate": 1e-05,
      "loss": 0.4248,
      "mean_token_accuracy": 0.862185001373291,
      "num_tokens": 662974128.0,
      "step": 4161
    },
    {
      "epoch": 2.1169888097660223,
      "grad_norm": 1.0422261953353882,
      "learning_rate": 1e-05,
      "loss": 0.4582,
      "mean_token_accuracy": 0.8512689471244812,
      "num_tokens": 663123259.0,
      "step": 4162
    },
    {
      "epoch": 2.117497456765005,
      "grad_norm": 0.9526087641716003,
      "learning_rate": 1e-05,
      "loss": 0.4219,
      "mean_token_accuracy": 0.8615497350692749,
      "num_tokens": 663280117.0,
      "step": 4163
    },
    {
      "epoch": 2.118006103763988,
      "grad_norm": 0.9149175882339478,
      "learning_rate": 1e-05,
      "loss": 0.4161,
      "mean_token_accuracy": 0.8647677898406982,
      "num_tokens": 663445600.0,
      "step": 4164
    },
    {
      "epoch": 2.1185147507629707,
      "grad_norm": 1.089755892753601,
      "learning_rate": 1e-05,
      "loss": 0.4349,
      "mean_token_accuracy": 0.8590355515480042,
      "num_tokens": 663599402.0,
      "step": 4165
    },
    {
      "epoch": 2.119023397761953,
      "grad_norm": 1.0591999292373657,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8534281849861145,
      "num_tokens": 663747652.0,
      "step": 4166
    },
    {
      "epoch": 2.119532044760936,
      "grad_norm": 0.9893795847892761,
      "learning_rate": 1e-05,
      "loss": 0.4519,
      "mean_token_accuracy": 0.8535148501396179,
      "num_tokens": 663903015.0,
      "step": 4167
    },
    {
      "epoch": 2.1200406917599186,
      "grad_norm": 1.015095591545105,
      "learning_rate": 1e-05,
      "loss": 0.4977,
      "mean_token_accuracy": 0.8418223857879639,
      "num_tokens": 664069904.0,
      "step": 4168
    },
    {
      "epoch": 2.1205493387589014,
      "grad_norm": 0.9398860931396484,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8568176627159119,
      "num_tokens": 664228482.0,
      "step": 4169
    },
    {
      "epoch": 2.121057985757884,
      "grad_norm": 1.008710265159607,
      "learning_rate": 1e-05,
      "loss": 0.4427,
      "mean_token_accuracy": 0.8567774295806885,
      "num_tokens": 664386511.0,
      "step": 4170
    },
    {
      "epoch": 2.1215666327568665,
      "grad_norm": 0.9196237325668335,
      "learning_rate": 1e-05,
      "loss": 0.4461,
      "mean_token_accuracy": 0.8564049005508423,
      "num_tokens": 664551869.0,
      "step": 4171
    },
    {
      "epoch": 2.1220752797558493,
      "grad_norm": 0.9031176567077637,
      "learning_rate": 1e-05,
      "loss": 0.4236,
      "mean_token_accuracy": 0.8635759949684143,
      "num_tokens": 664712795.0,
      "step": 4172
    },
    {
      "epoch": 2.122583926754832,
      "grad_norm": 1.0113224983215332,
      "learning_rate": 1e-05,
      "loss": 0.4479,
      "mean_token_accuracy": 0.8555098176002502,
      "num_tokens": 664889059.0,
      "step": 4173
    },
    {
      "epoch": 2.123092573753815,
      "grad_norm": 0.9578308463096619,
      "learning_rate": 1e-05,
      "loss": 0.4197,
      "mean_token_accuracy": 0.8610748648643494,
      "num_tokens": 665036489.0,
      "step": 4174
    },
    {
      "epoch": 2.1236012207527977,
      "grad_norm": 0.9232265949249268,
      "learning_rate": 1e-05,
      "loss": 0.4169,
      "mean_token_accuracy": 0.864449143409729,
      "num_tokens": 665192739.0,
      "step": 4175
    },
    {
      "epoch": 2.1241098677517805,
      "grad_norm": 0.9943797588348389,
      "learning_rate": 1e-05,
      "loss": 0.4364,
      "mean_token_accuracy": 0.8594911098480225,
      "num_tokens": 665345764.0,
      "step": 4176
    },
    {
      "epoch": 2.124618514750763,
      "grad_norm": 0.9452236890792847,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.8575600385665894,
      "num_tokens": 665498670.0,
      "step": 4177
    },
    {
      "epoch": 2.1251271617497456,
      "grad_norm": 0.8776096701622009,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.8611912727355957,
      "num_tokens": 665650495.0,
      "step": 4178
    },
    {
      "epoch": 2.1256358087487284,
      "grad_norm": 0.9938535690307617,
      "learning_rate": 1e-05,
      "loss": 0.4524,
      "mean_token_accuracy": 0.8539614081382751,
      "num_tokens": 665814461.0,
      "step": 4179
    },
    {
      "epoch": 2.126144455747711,
      "grad_norm": 0.9521315097808838,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.8568743467330933,
      "num_tokens": 665976759.0,
      "step": 4180
    },
    {
      "epoch": 2.126653102746694,
      "grad_norm": 0.8951365947723389,
      "learning_rate": 1e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8524468541145325,
      "num_tokens": 666134195.0,
      "step": 4181
    },
    {
      "epoch": 2.1271617497456763,
      "grad_norm": 1.0694705247879028,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "mean_token_accuracy": 0.8601644039154053,
      "num_tokens": 666274229.0,
      "step": 4182
    },
    {
      "epoch": 2.127670396744659,
      "grad_norm": 0.9351718425750732,
      "learning_rate": 1e-05,
      "loss": 0.4448,
      "mean_token_accuracy": 0.8556416034698486,
      "num_tokens": 666445641.0,
      "step": 4183
    },
    {
      "epoch": 2.128179043743642,
      "grad_norm": 0.9262452125549316,
      "learning_rate": 1e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.851718544960022,
      "num_tokens": 666617768.0,
      "step": 4184
    },
    {
      "epoch": 2.1286876907426246,
      "grad_norm": 0.8724216222763062,
      "learning_rate": 1e-05,
      "loss": 0.4291,
      "mean_token_accuracy": 0.8605974912643433,
      "num_tokens": 666775755.0,
      "step": 4185
    },
    {
      "epoch": 2.1291963377416074,
      "grad_norm": 0.9080948233604431,
      "learning_rate": 1e-05,
      "loss": 0.4679,
      "mean_token_accuracy": 0.8505131006240845,
      "num_tokens": 666935081.0,
      "step": 4186
    },
    {
      "epoch": 2.1297049847405902,
      "grad_norm": 0.9036731719970703,
      "learning_rate": 1e-05,
      "loss": 0.4081,
      "mean_token_accuracy": 0.8680239915847778,
      "num_tokens": 667083302.0,
      "step": 4187
    },
    {
      "epoch": 2.1302136317395726,
      "grad_norm": 0.9800155758857727,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.8516967296600342,
      "num_tokens": 667248835.0,
      "step": 4188
    },
    {
      "epoch": 2.1307222787385554,
      "grad_norm": 0.9089070558547974,
      "learning_rate": 1e-05,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8470054864883423,
      "num_tokens": 667402091.0,
      "step": 4189
    },
    {
      "epoch": 2.131230925737538,
      "grad_norm": 0.9217497706413269,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8607636094093323,
      "num_tokens": 667556405.0,
      "step": 4190
    },
    {
      "epoch": 2.131739572736521,
      "grad_norm": 0.9196179509162903,
      "learning_rate": 1e-05,
      "loss": 0.4536,
      "mean_token_accuracy": 0.8530000448226929,
      "num_tokens": 667723355.0,
      "step": 4191
    },
    {
      "epoch": 2.1322482197355037,
      "grad_norm": 0.9593817591667175,
      "learning_rate": 1e-05,
      "loss": 0.4511,
      "mean_token_accuracy": 0.853350043296814,
      "num_tokens": 667894932.0,
      "step": 4192
    },
    {
      "epoch": 2.1327568667344865,
      "grad_norm": 0.8810621500015259,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "mean_token_accuracy": 0.8580273389816284,
      "num_tokens": 668054091.0,
      "step": 4193
    },
    {
      "epoch": 2.133265513733469,
      "grad_norm": 0.9316674470901489,
      "learning_rate": 1e-05,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8623379468917847,
      "num_tokens": 668202295.0,
      "step": 4194
    },
    {
      "epoch": 2.1337741607324516,
      "grad_norm": 0.8470092415809631,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "mean_token_accuracy": 0.8674142360687256,
      "num_tokens": 668373127.0,
      "step": 4195
    },
    {
      "epoch": 2.1342828077314344,
      "grad_norm": 0.9411569237709045,
      "learning_rate": 1e-05,
      "loss": 0.4912,
      "mean_token_accuracy": 0.8427003622055054,
      "num_tokens": 668537919.0,
      "step": 4196
    },
    {
      "epoch": 2.134791454730417,
      "grad_norm": 1.4109463691711426,
      "learning_rate": 1e-05,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8517212271690369,
      "num_tokens": 668698199.0,
      "step": 4197
    },
    {
      "epoch": 2.1353001017294,
      "grad_norm": 0.9155637621879578,
      "learning_rate": 1e-05,
      "loss": 0.4086,
      "mean_token_accuracy": 0.8653716444969177,
      "num_tokens": 668854088.0,
      "step": 4198
    },
    {
      "epoch": 2.1358087487283823,
      "grad_norm": 0.9064233303070068,
      "learning_rate": 1e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8617438673973083,
      "num_tokens": 669017211.0,
      "step": 4199
    },
    {
      "epoch": 2.136317395727365,
      "grad_norm": 0.9097160696983337,
      "learning_rate": 1e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.862013041973114,
      "num_tokens": 669192028.0,
      "step": 4200
    },
    {
      "epoch": 2.136826042726348,
      "grad_norm": 0.9515498876571655,
      "learning_rate": 1e-05,
      "loss": 0.4183,
      "mean_token_accuracy": 0.8625266551971436,
      "num_tokens": 669352646.0,
      "step": 4201
    },
    {
      "epoch": 2.1373346897253307,
      "grad_norm": 0.9216282963752747,
      "learning_rate": 1e-05,
      "loss": 0.4375,
      "mean_token_accuracy": 0.8592369556427002,
      "num_tokens": 669516310.0,
      "step": 4202
    },
    {
      "epoch": 2.1378433367243135,
      "grad_norm": 0.9496901035308838,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.8564037680625916,
      "num_tokens": 669669101.0,
      "step": 4203
    },
    {
      "epoch": 2.138351983723296,
      "grad_norm": 1.013985276222229,
      "learning_rate": 1e-05,
      "loss": 0.451,
      "mean_token_accuracy": 0.8551113605499268,
      "num_tokens": 669843789.0,
      "step": 4204
    },
    {
      "epoch": 2.1388606307222786,
      "grad_norm": 0.9502270221710205,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "mean_token_accuracy": 0.8561847805976868,
      "num_tokens": 670005303.0,
      "step": 4205
    },
    {
      "epoch": 2.1393692777212614,
      "grad_norm": 0.9519503116607666,
      "learning_rate": 1e-05,
      "loss": 0.4327,
      "mean_token_accuracy": 0.8593729734420776,
      "num_tokens": 670161044.0,
      "step": 4206
    },
    {
      "epoch": 2.139877924720244,
      "grad_norm": 0.9718447923660278,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8558679819107056,
      "num_tokens": 670314008.0,
      "step": 4207
    },
    {
      "epoch": 2.140386571719227,
      "grad_norm": 0.9260292053222656,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8572784066200256,
      "num_tokens": 670470095.0,
      "step": 4208
    },
    {
      "epoch": 2.1408952187182098,
      "grad_norm": 0.9734790921211243,
      "learning_rate": 1e-05,
      "loss": 0.4735,
      "mean_token_accuracy": 0.8485541343688965,
      "num_tokens": 670636949.0,
      "step": 4209
    },
    {
      "epoch": 2.141403865717192,
      "grad_norm": 0.8780142664909363,
      "learning_rate": 1e-05,
      "loss": 0.4157,
      "mean_token_accuracy": 0.864952564239502,
      "num_tokens": 670784785.0,
      "step": 4210
    },
    {
      "epoch": 2.141912512716175,
      "grad_norm": 1.1260484457015991,
      "learning_rate": 1e-05,
      "loss": 0.4157,
      "mean_token_accuracy": 0.864479660987854,
      "num_tokens": 670956463.0,
      "step": 4211
    },
    {
      "epoch": 2.1424211597151577,
      "grad_norm": 0.9373603463172913,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8588770031929016,
      "num_tokens": 671102909.0,
      "step": 4212
    },
    {
      "epoch": 2.1429298067141405,
      "grad_norm": 0.8711733818054199,
      "learning_rate": 1e-05,
      "loss": 0.4303,
      "mean_token_accuracy": 0.8594882488250732,
      "num_tokens": 671264081.0,
      "step": 4213
    },
    {
      "epoch": 2.1434384537131232,
      "grad_norm": 1.0101747512817383,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8565623164176941,
      "num_tokens": 671421218.0,
      "step": 4214
    },
    {
      "epoch": 2.1439471007121056,
      "grad_norm": 0.9408650994300842,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8543228507041931,
      "num_tokens": 671578732.0,
      "step": 4215
    },
    {
      "epoch": 2.1444557477110884,
      "grad_norm": 1.341538429260254,
      "learning_rate": 1e-05,
      "loss": 0.433,
      "mean_token_accuracy": 0.8586952090263367,
      "num_tokens": 671755061.0,
      "step": 4216
    },
    {
      "epoch": 2.144964394710071,
      "grad_norm": 0.937558114528656,
      "learning_rate": 1e-05,
      "loss": 0.3851,
      "mean_token_accuracy": 0.8737555146217346,
      "num_tokens": 671921101.0,
      "step": 4217
    },
    {
      "epoch": 2.145473041709054,
      "grad_norm": 0.8825535178184509,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.858366072177887,
      "num_tokens": 672085682.0,
      "step": 4218
    },
    {
      "epoch": 2.1459816887080367,
      "grad_norm": 0.9849353432655334,
      "learning_rate": 1e-05,
      "loss": 0.457,
      "mean_token_accuracy": 0.851929783821106,
      "num_tokens": 672242998.0,
      "step": 4219
    },
    {
      "epoch": 2.1464903357070195,
      "grad_norm": 0.9155281186103821,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8466140031814575,
      "num_tokens": 672406808.0,
      "step": 4220
    },
    {
      "epoch": 2.146998982706002,
      "grad_norm": 0.9637844562530518,
      "learning_rate": 1e-05,
      "loss": 0.4826,
      "mean_token_accuracy": 0.8453380465507507,
      "num_tokens": 672582458.0,
      "step": 4221
    },
    {
      "epoch": 2.1475076297049847,
      "grad_norm": 0.9090501070022583,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.857498288154602,
      "num_tokens": 672742756.0,
      "step": 4222
    },
    {
      "epoch": 2.1480162767039674,
      "grad_norm": 0.9212506413459778,
      "learning_rate": 1e-05,
      "loss": 0.429,
      "mean_token_accuracy": 0.860740065574646,
      "num_tokens": 672909847.0,
      "step": 4223
    },
    {
      "epoch": 2.1485249237029502,
      "grad_norm": 0.8985199332237244,
      "learning_rate": 1e-05,
      "loss": 0.4101,
      "mean_token_accuracy": 0.8661127090454102,
      "num_tokens": 673076346.0,
      "step": 4224
    },
    {
      "epoch": 2.149033570701933,
      "grad_norm": 0.9333046078681946,
      "learning_rate": 1e-05,
      "loss": 0.4646,
      "mean_token_accuracy": 0.8508179187774658,
      "num_tokens": 673225384.0,
      "step": 4225
    },
    {
      "epoch": 2.1495422177009154,
      "grad_norm": 0.874630868434906,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8571752309799194,
      "num_tokens": 673390482.0,
      "step": 4226
    },
    {
      "epoch": 2.150050864699898,
      "grad_norm": 0.8874839544296265,
      "learning_rate": 1e-05,
      "loss": 0.4285,
      "mean_token_accuracy": 0.8597501516342163,
      "num_tokens": 673557143.0,
      "step": 4227
    },
    {
      "epoch": 2.150559511698881,
      "grad_norm": 0.9544134736061096,
      "learning_rate": 1e-05,
      "loss": 0.4667,
      "mean_token_accuracy": 0.8492194414138794,
      "num_tokens": 673724830.0,
      "step": 4228
    },
    {
      "epoch": 2.1510681586978637,
      "grad_norm": 0.9367156028747559,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8530341386795044,
      "num_tokens": 673885742.0,
      "step": 4229
    },
    {
      "epoch": 2.1515768056968465,
      "grad_norm": 0.9363099932670593,
      "learning_rate": 1e-05,
      "loss": 0.446,
      "mean_token_accuracy": 0.8563543558120728,
      "num_tokens": 674041915.0,
      "step": 4230
    },
    {
      "epoch": 2.1520854526958293,
      "grad_norm": 0.9003242254257202,
      "learning_rate": 1e-05,
      "loss": 0.427,
      "mean_token_accuracy": 0.8603519797325134,
      "num_tokens": 674203202.0,
      "step": 4231
    },
    {
      "epoch": 2.1525940996948116,
      "grad_norm": 0.8814538717269897,
      "learning_rate": 1e-05,
      "loss": 0.4837,
      "mean_token_accuracy": 0.8455694317817688,
      "num_tokens": 674376067.0,
      "step": 4232
    },
    {
      "epoch": 2.1531027466937944,
      "grad_norm": 0.9837362766265869,
      "learning_rate": 1e-05,
      "loss": 0.4417,
      "mean_token_accuracy": 0.8560826182365417,
      "num_tokens": 674530535.0,
      "step": 4233
    },
    {
      "epoch": 2.153611393692777,
      "grad_norm": 0.9501599073410034,
      "learning_rate": 1e-05,
      "loss": 0.4345,
      "mean_token_accuracy": 0.8594794273376465,
      "num_tokens": 674683835.0,
      "step": 4234
    },
    {
      "epoch": 2.15412004069176,
      "grad_norm": 0.8494070768356323,
      "learning_rate": 1e-05,
      "loss": 0.4346,
      "mean_token_accuracy": 0.8582461476325989,
      "num_tokens": 674852193.0,
      "step": 4235
    },
    {
      "epoch": 2.154628687690743,
      "grad_norm": 0.884700357913971,
      "learning_rate": 1e-05,
      "loss": 0.4657,
      "mean_token_accuracy": 0.8502305746078491,
      "num_tokens": 675021716.0,
      "step": 4236
    },
    {
      "epoch": 2.155137334689725,
      "grad_norm": 0.9532796144485474,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8599141836166382,
      "num_tokens": 675172087.0,
      "step": 4237
    },
    {
      "epoch": 2.155645981688708,
      "grad_norm": 0.9468601942062378,
      "learning_rate": 1e-05,
      "loss": 0.4473,
      "mean_token_accuracy": 0.8549070954322815,
      "num_tokens": 675339134.0,
      "step": 4238
    },
    {
      "epoch": 2.1561546286876907,
      "grad_norm": 0.9046058058738708,
      "learning_rate": 1e-05,
      "loss": 0.4525,
      "mean_token_accuracy": 0.8540990352630615,
      "num_tokens": 675497933.0,
      "step": 4239
    },
    {
      "epoch": 2.1566632756866735,
      "grad_norm": 0.9532819986343384,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "mean_token_accuracy": 0.8545533418655396,
      "num_tokens": 675650088.0,
      "step": 4240
    },
    {
      "epoch": 2.1571719226856563,
      "grad_norm": 0.8591033220291138,
      "learning_rate": 1e-05,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8633793592453003,
      "num_tokens": 675815367.0,
      "step": 4241
    },
    {
      "epoch": 2.157680569684639,
      "grad_norm": 0.9060652256011963,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8582218289375305,
      "num_tokens": 675976634.0,
      "step": 4242
    },
    {
      "epoch": 2.1581892166836214,
      "grad_norm": 0.9459686279296875,
      "learning_rate": 1e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8501343131065369,
      "num_tokens": 676134723.0,
      "step": 4243
    },
    {
      "epoch": 2.158697863682604,
      "grad_norm": 0.886417031288147,
      "learning_rate": 1e-05,
      "loss": 0.4586,
      "mean_token_accuracy": 0.851806104183197,
      "num_tokens": 676301382.0,
      "step": 4244
    },
    {
      "epoch": 2.159206510681587,
      "grad_norm": 0.8885103464126587,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8584810495376587,
      "num_tokens": 676457791.0,
      "step": 4245
    },
    {
      "epoch": 2.1597151576805698,
      "grad_norm": 0.9095398783683777,
      "learning_rate": 1e-05,
      "loss": 0.4475,
      "mean_token_accuracy": 0.8562595844268799,
      "num_tokens": 676630331.0,
      "step": 4246
    },
    {
      "epoch": 2.1602238046795526,
      "grad_norm": 0.8835845589637756,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "mean_token_accuracy": 0.8514906167984009,
      "num_tokens": 676806010.0,
      "step": 4247
    },
    {
      "epoch": 2.160732451678535,
      "grad_norm": 0.9202662110328674,
      "learning_rate": 1e-05,
      "loss": 0.4659,
      "mean_token_accuracy": 0.850483775138855,
      "num_tokens": 676971441.0,
      "step": 4248
    },
    {
      "epoch": 2.1612410986775177,
      "grad_norm": 1.0267709493637085,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "mean_token_accuracy": 0.8617174625396729,
      "num_tokens": 677126808.0,
      "step": 4249
    },
    {
      "epoch": 2.1617497456765005,
      "grad_norm": 1.0010653734207153,
      "learning_rate": 1e-05,
      "loss": 0.4493,
      "mean_token_accuracy": 0.8573693037033081,
      "num_tokens": 677292539.0,
      "step": 4250
    },
    {
      "epoch": 2.1622583926754833,
      "grad_norm": 0.9092729091644287,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "mean_token_accuracy": 0.865876317024231,
      "num_tokens": 677442157.0,
      "step": 4251
    },
    {
      "epoch": 2.162767039674466,
      "grad_norm": 0.8797513246536255,
      "learning_rate": 1e-05,
      "loss": 0.4062,
      "mean_token_accuracy": 0.8681555986404419,
      "num_tokens": 677609509.0,
      "step": 4252
    },
    {
      "epoch": 2.163275686673449,
      "grad_norm": 0.9524892568588257,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "mean_token_accuracy": 0.8616195917129517,
      "num_tokens": 677762481.0,
      "step": 4253
    },
    {
      "epoch": 2.163784333672431,
      "grad_norm": 0.9220227003097534,
      "learning_rate": 1e-05,
      "loss": 0.4607,
      "mean_token_accuracy": 0.851956307888031,
      "num_tokens": 677937681.0,
      "step": 4254
    },
    {
      "epoch": 2.164292980671414,
      "grad_norm": 0.947273850440979,
      "learning_rate": 1e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8622714877128601,
      "num_tokens": 678092076.0,
      "step": 4255
    },
    {
      "epoch": 2.1648016276703967,
      "grad_norm": 0.9841005206108093,
      "learning_rate": 1e-05,
      "loss": 0.3996,
      "mean_token_accuracy": 0.8677623271942139,
      "num_tokens": 678236943.0,
      "step": 4256
    },
    {
      "epoch": 2.1653102746693795,
      "grad_norm": 0.9958181977272034,
      "learning_rate": 1e-05,
      "loss": 0.4666,
      "mean_token_accuracy": 0.8516644239425659,
      "num_tokens": 678377775.0,
      "step": 4257
    },
    {
      "epoch": 2.1658189216683623,
      "grad_norm": 0.9725253582000732,
      "learning_rate": 1e-05,
      "loss": 0.4431,
      "mean_token_accuracy": 0.8572109937667847,
      "num_tokens": 678533114.0,
      "step": 4258
    },
    {
      "epoch": 2.1663275686673447,
      "grad_norm": 0.9555931091308594,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8605384230613708,
      "num_tokens": 678694904.0,
      "step": 4259
    },
    {
      "epoch": 2.1668362156663274,
      "grad_norm": 1.3989323377609253,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8648083209991455,
      "num_tokens": 678854315.0,
      "step": 4260
    },
    {
      "epoch": 2.1673448626653102,
      "grad_norm": 0.9446583390235901,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8639305830001831,
      "num_tokens": 679013629.0,
      "step": 4261
    },
    {
      "epoch": 2.167853509664293,
      "grad_norm": 0.8738034963607788,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8571431636810303,
      "num_tokens": 679170763.0,
      "step": 4262
    },
    {
      "epoch": 2.168362156663276,
      "grad_norm": 1.0060641765594482,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.8516331911087036,
      "num_tokens": 679322506.0,
      "step": 4263
    },
    {
      "epoch": 2.1688708036622586,
      "grad_norm": 0.8976203203201294,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.8569817543029785,
      "num_tokens": 679480751.0,
      "step": 4264
    },
    {
      "epoch": 2.169379450661241,
      "grad_norm": 0.8971258997917175,
      "learning_rate": 1e-05,
      "loss": 0.4574,
      "mean_token_accuracy": 0.8524659872055054,
      "num_tokens": 679639537.0,
      "step": 4265
    },
    {
      "epoch": 2.1698880976602237,
      "grad_norm": 0.925749659538269,
      "learning_rate": 1e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8633953332901001,
      "num_tokens": 679799542.0,
      "step": 4266
    },
    {
      "epoch": 2.1703967446592065,
      "grad_norm": 0.8783673048019409,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8596926331520081,
      "num_tokens": 679969541.0,
      "step": 4267
    },
    {
      "epoch": 2.1709053916581893,
      "grad_norm": 0.9182345271110535,
      "learning_rate": 1e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.8597497940063477,
      "num_tokens": 680126987.0,
      "step": 4268
    },
    {
      "epoch": 2.171414038657172,
      "grad_norm": 0.8887823224067688,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8598739504814148,
      "num_tokens": 680294249.0,
      "step": 4269
    },
    {
      "epoch": 2.1719226856561544,
      "grad_norm": 0.8925853967666626,
      "learning_rate": 1e-05,
      "loss": 0.4499,
      "mean_token_accuracy": 0.8564271330833435,
      "num_tokens": 680445654.0,
      "step": 4270
    },
    {
      "epoch": 2.172431332655137,
      "grad_norm": 0.8978914022445679,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8519887924194336,
      "num_tokens": 680608877.0,
      "step": 4271
    },
    {
      "epoch": 2.17293997965412,
      "grad_norm": 0.8895770311355591,
      "learning_rate": 1e-05,
      "loss": 0.4516,
      "mean_token_accuracy": 0.855591356754303,
      "num_tokens": 680766149.0,
      "step": 4272
    },
    {
      "epoch": 2.173448626653103,
      "grad_norm": 0.8657394647598267,
      "learning_rate": 1e-05,
      "loss": 0.4091,
      "mean_token_accuracy": 0.8654162883758545,
      "num_tokens": 680924361.0,
      "step": 4273
    },
    {
      "epoch": 2.1739572736520856,
      "grad_norm": 0.9019215703010559,
      "learning_rate": 1e-05,
      "loss": 0.4399,
      "mean_token_accuracy": 0.8570451140403748,
      "num_tokens": 681081193.0,
      "step": 4274
    },
    {
      "epoch": 2.1744659206510684,
      "grad_norm": 0.8810397982597351,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8547847867012024,
      "num_tokens": 681233678.0,
      "step": 4275
    },
    {
      "epoch": 2.1749745676500507,
      "grad_norm": 0.9701651930809021,
      "learning_rate": 1e-05,
      "loss": 0.454,
      "mean_token_accuracy": 0.8537286520004272,
      "num_tokens": 681389768.0,
      "step": 4276
    },
    {
      "epoch": 2.1754832146490335,
      "grad_norm": 0.9187273979187012,
      "learning_rate": 1e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8543602228164673,
      "num_tokens": 681558379.0,
      "step": 4277
    },
    {
      "epoch": 2.1759918616480163,
      "grad_norm": 0.8782539367675781,
      "learning_rate": 1e-05,
      "loss": 0.4065,
      "mean_token_accuracy": 0.8672250509262085,
      "num_tokens": 681716150.0,
      "step": 4278
    },
    {
      "epoch": 2.176500508646999,
      "grad_norm": 0.9115551114082336,
      "learning_rate": 1e-05,
      "loss": 0.481,
      "mean_token_accuracy": 0.8463177680969238,
      "num_tokens": 681881029.0,
      "step": 4279
    },
    {
      "epoch": 2.177009155645982,
      "grad_norm": 0.9686064124107361,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8595022559165955,
      "num_tokens": 682034982.0,
      "step": 4280
    },
    {
      "epoch": 2.177517802644964,
      "grad_norm": 0.9756225943565369,
      "learning_rate": 1e-05,
      "loss": 0.4371,
      "mean_token_accuracy": 0.8567383289337158,
      "num_tokens": 682196326.0,
      "step": 4281
    },
    {
      "epoch": 2.178026449643947,
      "grad_norm": 0.9038667678833008,
      "learning_rate": 1e-05,
      "loss": 0.4279,
      "mean_token_accuracy": 0.8613846302032471,
      "num_tokens": 682360761.0,
      "step": 4282
    },
    {
      "epoch": 2.1785350966429298,
      "grad_norm": 0.9914163947105408,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8609004020690918,
      "num_tokens": 682525520.0,
      "step": 4283
    },
    {
      "epoch": 2.1790437436419126,
      "grad_norm": 0.8739263415336609,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8661122918128967,
      "num_tokens": 682683740.0,
      "step": 4284
    },
    {
      "epoch": 2.1795523906408953,
      "grad_norm": 1.0040781497955322,
      "learning_rate": 1e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8597602248191833,
      "num_tokens": 682851431.0,
      "step": 4285
    },
    {
      "epoch": 2.180061037639878,
      "grad_norm": 1.0031808614730835,
      "learning_rate": 1e-05,
      "loss": 0.421,
      "mean_token_accuracy": 0.8615937232971191,
      "num_tokens": 682998758.0,
      "step": 4286
    },
    {
      "epoch": 2.1805696846388605,
      "grad_norm": 1.015446662902832,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8581506013870239,
      "num_tokens": 683156422.0,
      "step": 4287
    },
    {
      "epoch": 2.1810783316378433,
      "grad_norm": 1.040065050125122,
      "learning_rate": 1e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.8622639179229736,
      "num_tokens": 683315540.0,
      "step": 4288
    },
    {
      "epoch": 2.181586978636826,
      "grad_norm": 0.889073371887207,
      "learning_rate": 1e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.8521596193313599,
      "num_tokens": 683475268.0,
      "step": 4289
    },
    {
      "epoch": 2.182095625635809,
      "grad_norm": 1.0887638330459595,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8617396950721741,
      "num_tokens": 683639710.0,
      "step": 4290
    },
    {
      "epoch": 2.1826042726347916,
      "grad_norm": 0.8876301646232605,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.8568639755249023,
      "num_tokens": 683807593.0,
      "step": 4291
    },
    {
      "epoch": 2.183112919633774,
      "grad_norm": 1.022648572921753,
      "learning_rate": 1e-05,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8475189208984375,
      "num_tokens": 683962253.0,
      "step": 4292
    },
    {
      "epoch": 2.1836215666327567,
      "grad_norm": 0.9099440574645996,
      "learning_rate": 1e-05,
      "loss": 0.4019,
      "mean_token_accuracy": 0.8683190941810608,
      "num_tokens": 684134385.0,
      "step": 4293
    },
    {
      "epoch": 2.1841302136317395,
      "grad_norm": 0.9514361619949341,
      "learning_rate": 1e-05,
      "loss": 0.4632,
      "mean_token_accuracy": 0.8512458801269531,
      "num_tokens": 684276805.0,
      "step": 4294
    },
    {
      "epoch": 2.1846388606307223,
      "grad_norm": 1.0674186944961548,
      "learning_rate": 1e-05,
      "loss": 0.4138,
      "mean_token_accuracy": 0.8648924827575684,
      "num_tokens": 684434079.0,
      "step": 4295
    },
    {
      "epoch": 2.185147507629705,
      "grad_norm": 0.9300355315208435,
      "learning_rate": 1e-05,
      "loss": 0.4212,
      "mean_token_accuracy": 0.8619647026062012,
      "num_tokens": 684600612.0,
      "step": 4296
    },
    {
      "epoch": 2.185656154628688,
      "grad_norm": 0.9934197664260864,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8561198711395264,
      "num_tokens": 684769275.0,
      "step": 4297
    },
    {
      "epoch": 2.1861648016276702,
      "grad_norm": 1.076294183731079,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.8571916222572327,
      "num_tokens": 684915465.0,
      "step": 4298
    },
    {
      "epoch": 2.186673448626653,
      "grad_norm": 0.9175110459327698,
      "learning_rate": 1e-05,
      "loss": 0.4641,
      "mean_token_accuracy": 0.851553738117218,
      "num_tokens": 685086328.0,
      "step": 4299
    },
    {
      "epoch": 2.187182095625636,
      "grad_norm": 1.1934930086135864,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.859502911567688,
      "num_tokens": 685243003.0,
      "step": 4300
    },
    {
      "epoch": 2.1876907426246186,
      "grad_norm": 0.9275957345962524,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8554542064666748,
      "num_tokens": 685418099.0,
      "step": 4301
    },
    {
      "epoch": 2.1881993896236014,
      "grad_norm": 0.9224422574043274,
      "learning_rate": 1e-05,
      "loss": 0.4245,
      "mean_token_accuracy": 0.8605654239654541,
      "num_tokens": 685586437.0,
      "step": 4302
    },
    {
      "epoch": 2.1887080366225837,
      "grad_norm": 0.966139018535614,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8592056035995483,
      "num_tokens": 685737580.0,
      "step": 4303
    },
    {
      "epoch": 2.1892166836215665,
      "grad_norm": 0.9049349427223206,
      "learning_rate": 1e-05,
      "loss": 0.4198,
      "mean_token_accuracy": 0.8639073371887207,
      "num_tokens": 685883071.0,
      "step": 4304
    },
    {
      "epoch": 2.1897253306205493,
      "grad_norm": 0.9150710701942444,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8586542010307312,
      "num_tokens": 686047266.0,
      "step": 4305
    },
    {
      "epoch": 2.190233977619532,
      "grad_norm": 0.9289497137069702,
      "learning_rate": 1e-05,
      "loss": 0.4143,
      "mean_token_accuracy": 0.8645882606506348,
      "num_tokens": 686203095.0,
      "step": 4306
    },
    {
      "epoch": 2.190742624618515,
      "grad_norm": 0.9255957007408142,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "mean_token_accuracy": 0.8672335147857666,
      "num_tokens": 686358907.0,
      "step": 4307
    },
    {
      "epoch": 2.1912512716174977,
      "grad_norm": 0.874142587184906,
      "learning_rate": 1e-05,
      "loss": 0.4394,
      "mean_token_accuracy": 0.8571419715881348,
      "num_tokens": 686515166.0,
      "step": 4308
    },
    {
      "epoch": 2.19175991861648,
      "grad_norm": 0.9326626658439636,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8514823317527771,
      "num_tokens": 686681595.0,
      "step": 4309
    },
    {
      "epoch": 2.192268565615463,
      "grad_norm": 1.4150363206863403,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.8597804307937622,
      "num_tokens": 686847911.0,
      "step": 4310
    },
    {
      "epoch": 2.1927772126144456,
      "grad_norm": 0.9543857574462891,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8587651252746582,
      "num_tokens": 686996951.0,
      "step": 4311
    },
    {
      "epoch": 2.1932858596134284,
      "grad_norm": 0.9119658470153809,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "mean_token_accuracy": 0.8558597564697266,
      "num_tokens": 687161028.0,
      "step": 4312
    },
    {
      "epoch": 2.193794506612411,
      "grad_norm": 0.8924853801727295,
      "learning_rate": 1e-05,
      "loss": 0.4274,
      "mean_token_accuracy": 0.8626085519790649,
      "num_tokens": 687314790.0,
      "step": 4313
    },
    {
      "epoch": 2.1943031536113935,
      "grad_norm": 0.9578880071640015,
      "learning_rate": 1e-05,
      "loss": 0.4478,
      "mean_token_accuracy": 0.8548815846443176,
      "num_tokens": 687462776.0,
      "step": 4314
    },
    {
      "epoch": 2.1948118006103763,
      "grad_norm": 0.8604456186294556,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8595722913742065,
      "num_tokens": 687627884.0,
      "step": 4315
    },
    {
      "epoch": 2.195320447609359,
      "grad_norm": 0.9253394603729248,
      "learning_rate": 1e-05,
      "loss": 0.4244,
      "mean_token_accuracy": 0.8634549975395203,
      "num_tokens": 687785512.0,
      "step": 4316
    },
    {
      "epoch": 2.195829094608342,
      "grad_norm": 1.0204318761825562,
      "learning_rate": 1e-05,
      "loss": 0.4369,
      "mean_token_accuracy": 0.8581687211990356,
      "num_tokens": 687955390.0,
      "step": 4317
    },
    {
      "epoch": 2.1963377416073246,
      "grad_norm": 0.8677594661712646,
      "learning_rate": 1e-05,
      "loss": 0.4241,
      "mean_token_accuracy": 0.8619781136512756,
      "num_tokens": 688112149.0,
      "step": 4318
    },
    {
      "epoch": 2.196846388606307,
      "grad_norm": 0.87612384557724,
      "learning_rate": 1e-05,
      "loss": 0.4144,
      "mean_token_accuracy": 0.8636021018028259,
      "num_tokens": 688264263.0,
      "step": 4319
    },
    {
      "epoch": 2.1973550356052898,
      "grad_norm": 0.8514624834060669,
      "learning_rate": 1e-05,
      "loss": 0.42,
      "mean_token_accuracy": 0.8636860847473145,
      "num_tokens": 688432958.0,
      "step": 4320
    },
    {
      "epoch": 2.1978636826042726,
      "grad_norm": 0.9815389513969421,
      "learning_rate": 1e-05,
      "loss": 0.4768,
      "mean_token_accuracy": 0.8477411270141602,
      "num_tokens": 688593908.0,
      "step": 4321
    },
    {
      "epoch": 2.1983723296032553,
      "grad_norm": 0.9428776502609253,
      "learning_rate": 1e-05,
      "loss": 0.4225,
      "mean_token_accuracy": 0.8623034954071045,
      "num_tokens": 688745272.0,
      "step": 4322
    },
    {
      "epoch": 2.198880976602238,
      "grad_norm": 0.9332963228225708,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "mean_token_accuracy": 0.8518528938293457,
      "num_tokens": 688903056.0,
      "step": 4323
    },
    {
      "epoch": 2.199389623601221,
      "grad_norm": 0.917495846748352,
      "learning_rate": 1e-05,
      "loss": 0.4289,
      "mean_token_accuracy": 0.8599016666412354,
      "num_tokens": 689058958.0,
      "step": 4324
    },
    {
      "epoch": 2.1998982706002033,
      "grad_norm": 0.939020037651062,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8560481071472168,
      "num_tokens": 689234401.0,
      "step": 4325
    },
    {
      "epoch": 2.200406917599186,
      "grad_norm": 0.9460886120796204,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8559812307357788,
      "num_tokens": 689387117.0,
      "step": 4326
    },
    {
      "epoch": 2.200915564598169,
      "grad_norm": 0.9061490893363953,
      "learning_rate": 1e-05,
      "loss": 0.4375,
      "mean_token_accuracy": 0.8584349155426025,
      "num_tokens": 689537247.0,
      "step": 4327
    },
    {
      "epoch": 2.2014242115971516,
      "grad_norm": 1.0042805671691895,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8566774725914001,
      "num_tokens": 689703523.0,
      "step": 4328
    },
    {
      "epoch": 2.2019328585961344,
      "grad_norm": 0.9766663312911987,
      "learning_rate": 1e-05,
      "loss": 0.463,
      "mean_token_accuracy": 0.849650502204895,
      "num_tokens": 689854306.0,
      "step": 4329
    },
    {
      "epoch": 2.202441505595117,
      "grad_norm": 0.9077035784721375,
      "learning_rate": 1e-05,
      "loss": 0.452,
      "mean_token_accuracy": 0.8559194803237915,
      "num_tokens": 690011320.0,
      "step": 4330
    },
    {
      "epoch": 2.2029501525940995,
      "grad_norm": 1.0141128301620483,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8605471849441528,
      "num_tokens": 690159881.0,
      "step": 4331
    },
    {
      "epoch": 2.2034587995930823,
      "grad_norm": 0.9698449969291687,
      "learning_rate": 1e-05,
      "loss": 0.4123,
      "mean_token_accuracy": 0.8654199838638306,
      "num_tokens": 690312474.0,
      "step": 4332
    },
    {
      "epoch": 2.203967446592065,
      "grad_norm": 0.950181245803833,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8600283861160278,
      "num_tokens": 690471811.0,
      "step": 4333
    },
    {
      "epoch": 2.204476093591048,
      "grad_norm": 1.8057748079299927,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8640917539596558,
      "num_tokens": 690629913.0,
      "step": 4334
    },
    {
      "epoch": 2.2049847405900307,
      "grad_norm": 0.9295744895935059,
      "learning_rate": 1e-05,
      "loss": 0.411,
      "mean_token_accuracy": 0.8665107488632202,
      "num_tokens": 690791021.0,
      "step": 4335
    },
    {
      "epoch": 2.205493387589013,
      "grad_norm": 0.9357677698135376,
      "learning_rate": 1e-05,
      "loss": 0.4188,
      "mean_token_accuracy": 0.8637399077415466,
      "num_tokens": 690946910.0,
      "step": 4336
    },
    {
      "epoch": 2.206002034587996,
      "grad_norm": 0.9426012635231018,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.8539572954177856,
      "num_tokens": 691114027.0,
      "step": 4337
    },
    {
      "epoch": 2.2065106815869786,
      "grad_norm": 0.8814846873283386,
      "learning_rate": 1e-05,
      "loss": 0.4558,
      "mean_token_accuracy": 0.8528778553009033,
      "num_tokens": 691287105.0,
      "step": 4338
    },
    {
      "epoch": 2.2070193285859614,
      "grad_norm": 0.9679071307182312,
      "learning_rate": 1e-05,
      "loss": 0.4367,
      "mean_token_accuracy": 0.8572427034378052,
      "num_tokens": 691433141.0,
      "step": 4339
    },
    {
      "epoch": 2.207527975584944,
      "grad_norm": 0.9318785667419434,
      "learning_rate": 1e-05,
      "loss": 0.4304,
      "mean_token_accuracy": 0.8598240613937378,
      "num_tokens": 691590813.0,
      "step": 4340
    },
    {
      "epoch": 2.2080366225839265,
      "grad_norm": 0.9134104251861572,
      "learning_rate": 1e-05,
      "loss": 0.4523,
      "mean_token_accuracy": 0.8552306890487671,
      "num_tokens": 691753280.0,
      "step": 4341
    },
    {
      "epoch": 2.2085452695829093,
      "grad_norm": 0.9650377631187439,
      "learning_rate": 1e-05,
      "loss": 0.4263,
      "mean_token_accuracy": 0.8613597750663757,
      "num_tokens": 691912782.0,
      "step": 4342
    },
    {
      "epoch": 2.209053916581892,
      "grad_norm": 0.8906856775283813,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8498727083206177,
      "num_tokens": 692077739.0,
      "step": 4343
    },
    {
      "epoch": 2.209562563580875,
      "grad_norm": 0.9788679480552673,
      "learning_rate": 1e-05,
      "loss": 0.4191,
      "mean_token_accuracy": 0.8626728057861328,
      "num_tokens": 692242294.0,
      "step": 4344
    },
    {
      "epoch": 2.2100712105798577,
      "grad_norm": 0.9652032256126404,
      "learning_rate": 1e-05,
      "loss": 0.4521,
      "mean_token_accuracy": 0.8537179231643677,
      "num_tokens": 692397707.0,
      "step": 4345
    },
    {
      "epoch": 2.2105798575788405,
      "grad_norm": 0.9755709767341614,
      "learning_rate": 1e-05,
      "loss": 0.4716,
      "mean_token_accuracy": 0.8492090702056885,
      "num_tokens": 692562541.0,
      "step": 4346
    },
    {
      "epoch": 2.211088504577823,
      "grad_norm": 0.971756637096405,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8570781350135803,
      "num_tokens": 692709163.0,
      "step": 4347
    },
    {
      "epoch": 2.2115971515768056,
      "grad_norm": 0.9199160933494568,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8610886335372925,
      "num_tokens": 692863692.0,
      "step": 4348
    },
    {
      "epoch": 2.2121057985757884,
      "grad_norm": 0.9759620428085327,
      "learning_rate": 1e-05,
      "loss": 0.4133,
      "mean_token_accuracy": 0.8647420406341553,
      "num_tokens": 693018522.0,
      "step": 4349
    },
    {
      "epoch": 2.212614445574771,
      "grad_norm": 0.872011125087738,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "mean_token_accuracy": 0.8554370403289795,
      "num_tokens": 693172309.0,
      "step": 4350
    },
    {
      "epoch": 2.213123092573754,
      "grad_norm": 0.9128387570381165,
      "learning_rate": 1e-05,
      "loss": 0.4629,
      "mean_token_accuracy": 0.8515660762786865,
      "num_tokens": 693334327.0,
      "step": 4351
    },
    {
      "epoch": 2.2136317395727367,
      "grad_norm": 0.946731686592102,
      "learning_rate": 1e-05,
      "loss": 0.4352,
      "mean_token_accuracy": 0.8582457304000854,
      "num_tokens": 693494304.0,
      "step": 4352
    },
    {
      "epoch": 2.214140386571719,
      "grad_norm": 0.9041470289230347,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.8558679819107056,
      "num_tokens": 693664029.0,
      "step": 4353
    },
    {
      "epoch": 2.214649033570702,
      "grad_norm": 1.008070945739746,
      "learning_rate": 1e-05,
      "loss": 0.42,
      "mean_token_accuracy": 0.8630983829498291,
      "num_tokens": 693827204.0,
      "step": 4354
    },
    {
      "epoch": 2.2151576805696847,
      "grad_norm": 0.8689998984336853,
      "learning_rate": 1e-05,
      "loss": 0.4728,
      "mean_token_accuracy": 0.848616361618042,
      "num_tokens": 694006803.0,
      "step": 4355
    },
    {
      "epoch": 2.2156663275686674,
      "grad_norm": 0.9089202880859375,
      "learning_rate": 1e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8560207486152649,
      "num_tokens": 694172486.0,
      "step": 4356
    },
    {
      "epoch": 2.2161749745676502,
      "grad_norm": 0.9482621550559998,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.8523670434951782,
      "num_tokens": 694346204.0,
      "step": 4357
    },
    {
      "epoch": 2.2166836215666326,
      "grad_norm": 1.0576685667037964,
      "learning_rate": 1e-05,
      "loss": 0.4839,
      "mean_token_accuracy": 0.8448648452758789,
      "num_tokens": 694504560.0,
      "step": 4358
    },
    {
      "epoch": 2.2171922685656154,
      "grad_norm": 0.9813886880874634,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8526740074157715,
      "num_tokens": 694652226.0,
      "step": 4359
    },
    {
      "epoch": 2.217700915564598,
      "grad_norm": 0.9225329756736755,
      "learning_rate": 1e-05,
      "loss": 0.4628,
      "mean_token_accuracy": 0.849311351776123,
      "num_tokens": 694817430.0,
      "step": 4360
    },
    {
      "epoch": 2.218209562563581,
      "grad_norm": 0.9040138721466064,
      "learning_rate": 1e-05,
      "loss": 0.4274,
      "mean_token_accuracy": 0.8618972301483154,
      "num_tokens": 694970575.0,
      "step": 4361
    },
    {
      "epoch": 2.2187182095625637,
      "grad_norm": 1.0440372228622437,
      "learning_rate": 1e-05,
      "loss": 0.4345,
      "mean_token_accuracy": 0.8597314953804016,
      "num_tokens": 695139610.0,
      "step": 4362
    },
    {
      "epoch": 2.219226856561546,
      "grad_norm": 0.9187480807304382,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "mean_token_accuracy": 0.8600801229476929,
      "num_tokens": 695290776.0,
      "step": 4363
    },
    {
      "epoch": 2.219735503560529,
      "grad_norm": 0.9653752446174622,
      "learning_rate": 1e-05,
      "loss": 0.4271,
      "mean_token_accuracy": 0.8609426021575928,
      "num_tokens": 695455043.0,
      "step": 4364
    },
    {
      "epoch": 2.2202441505595116,
      "grad_norm": 0.843532919883728,
      "learning_rate": 1e-05,
      "loss": 0.4309,
      "mean_token_accuracy": 0.8611189126968384,
      "num_tokens": 695627666.0,
      "step": 4365
    },
    {
      "epoch": 2.2207527975584944,
      "grad_norm": 0.9980239272117615,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8516084551811218,
      "num_tokens": 695788464.0,
      "step": 4366
    },
    {
      "epoch": 2.221261444557477,
      "grad_norm": 0.9190233945846558,
      "learning_rate": 1e-05,
      "loss": 0.4705,
      "mean_token_accuracy": 0.849383533000946,
      "num_tokens": 695943810.0,
      "step": 4367
    },
    {
      "epoch": 2.22177009155646,
      "grad_norm": 0.9599975943565369,
      "learning_rate": 1e-05,
      "loss": 0.4153,
      "mean_token_accuracy": 0.8644353151321411,
      "num_tokens": 696110054.0,
      "step": 4368
    },
    {
      "epoch": 2.2222787385554423,
      "grad_norm": 0.970862627029419,
      "learning_rate": 1e-05,
      "loss": 0.4151,
      "mean_token_accuracy": 0.8649633526802063,
      "num_tokens": 696264565.0,
      "step": 4369
    },
    {
      "epoch": 2.222787385554425,
      "grad_norm": 0.9327432513237,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "mean_token_accuracy": 0.8549506068229675,
      "num_tokens": 696419384.0,
      "step": 4370
    },
    {
      "epoch": 2.223296032553408,
      "grad_norm": 0.9206057190895081,
      "learning_rate": 1e-05,
      "loss": 0.4202,
      "mean_token_accuracy": 0.8625045418739319,
      "num_tokens": 696584696.0,
      "step": 4371
    },
    {
      "epoch": 2.2238046795523907,
      "grad_norm": 0.8772811889648438,
      "learning_rate": 1e-05,
      "loss": 0.4287,
      "mean_token_accuracy": 0.8601574301719666,
      "num_tokens": 696754841.0,
      "step": 4372
    },
    {
      "epoch": 2.2243133265513735,
      "grad_norm": 0.9989537596702576,
      "learning_rate": 1e-05,
      "loss": 0.445,
      "mean_token_accuracy": 0.8562902808189392,
      "num_tokens": 696901513.0,
      "step": 4373
    },
    {
      "epoch": 2.2248219735503563,
      "grad_norm": 0.8996737599372864,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "mean_token_accuracy": 0.8534572124481201,
      "num_tokens": 697066242.0,
      "step": 4374
    },
    {
      "epoch": 2.2253306205493386,
      "grad_norm": 0.9436988234519958,
      "learning_rate": 1e-05,
      "loss": 0.4468,
      "mean_token_accuracy": 0.8555812835693359,
      "num_tokens": 697217547.0,
      "step": 4375
    },
    {
      "epoch": 2.2258392675483214,
      "grad_norm": 0.859586775302887,
      "learning_rate": 1e-05,
      "loss": 0.4321,
      "mean_token_accuracy": 0.8602193593978882,
      "num_tokens": 697384006.0,
      "step": 4376
    },
    {
      "epoch": 2.226347914547304,
      "grad_norm": 1.0298329591751099,
      "learning_rate": 1e-05,
      "loss": 0.4831,
      "mean_token_accuracy": 0.8449706435203552,
      "num_tokens": 697534624.0,
      "step": 4377
    },
    {
      "epoch": 2.226856561546287,
      "grad_norm": 0.973120927810669,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "mean_token_accuracy": 0.8544663786888123,
      "num_tokens": 697695648.0,
      "step": 4378
    },
    {
      "epoch": 2.2273652085452698,
      "grad_norm": 0.8641661405563354,
      "learning_rate": 1e-05,
      "loss": 0.4363,
      "mean_token_accuracy": 0.8581136465072632,
      "num_tokens": 697856787.0,
      "step": 4379
    },
    {
      "epoch": 2.227873855544252,
      "grad_norm": 0.9121147394180298,
      "learning_rate": 1e-05,
      "loss": 0.4387,
      "mean_token_accuracy": 0.8577170968055725,
      "num_tokens": 698019018.0,
      "step": 4380
    },
    {
      "epoch": 2.228382502543235,
      "grad_norm": 0.9240990281105042,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8579505681991577,
      "num_tokens": 698180128.0,
      "step": 4381
    },
    {
      "epoch": 2.2288911495422177,
      "grad_norm": 0.8957146406173706,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8546497821807861,
      "num_tokens": 698339787.0,
      "step": 4382
    },
    {
      "epoch": 2.2293997965412005,
      "grad_norm": 0.9592480063438416,
      "learning_rate": 1e-05,
      "loss": 0.4234,
      "mean_token_accuracy": 0.8617354035377502,
      "num_tokens": 698490297.0,
      "step": 4383
    },
    {
      "epoch": 2.2299084435401832,
      "grad_norm": 0.9136049747467041,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "mean_token_accuracy": 0.8556954264640808,
      "num_tokens": 698644406.0,
      "step": 4384
    },
    {
      "epoch": 2.2304170905391656,
      "grad_norm": 0.9772505760192871,
      "learning_rate": 1e-05,
      "loss": 0.4451,
      "mean_token_accuracy": 0.8578481674194336,
      "num_tokens": 698794461.0,
      "step": 4385
    },
    {
      "epoch": 2.2309257375381484,
      "grad_norm": 0.8731710910797119,
      "learning_rate": 1e-05,
      "loss": 0.4063,
      "mean_token_accuracy": 0.8656190037727356,
      "num_tokens": 698949448.0,
      "step": 4386
    },
    {
      "epoch": 2.231434384537131,
      "grad_norm": 0.926904559135437,
      "learning_rate": 1e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8636926412582397,
      "num_tokens": 699113160.0,
      "step": 4387
    },
    {
      "epoch": 2.231943031536114,
      "grad_norm": 0.888241708278656,
      "learning_rate": 1e-05,
      "loss": 0.4762,
      "mean_token_accuracy": 0.847213864326477,
      "num_tokens": 699276343.0,
      "step": 4388
    },
    {
      "epoch": 2.2324516785350967,
      "grad_norm": 0.9436233639717102,
      "learning_rate": 1e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8591353893280029,
      "num_tokens": 699436640.0,
      "step": 4389
    },
    {
      "epoch": 2.2329603255340795,
      "grad_norm": 0.8924297094345093,
      "learning_rate": 1e-05,
      "loss": 0.4631,
      "mean_token_accuracy": 0.8514504432678223,
      "num_tokens": 699600680.0,
      "step": 4390
    },
    {
      "epoch": 2.233468972533062,
      "grad_norm": 0.9052316546440125,
      "learning_rate": 1e-05,
      "loss": 0.4125,
      "mean_token_accuracy": 0.8639260530471802,
      "num_tokens": 699758447.0,
      "step": 4391
    },
    {
      "epoch": 2.2339776195320447,
      "grad_norm": 0.8749424815177917,
      "learning_rate": 1e-05,
      "loss": 0.435,
      "mean_token_accuracy": 0.8595032691955566,
      "num_tokens": 699920412.0,
      "step": 4392
    },
    {
      "epoch": 2.2344862665310274,
      "grad_norm": 0.925260603427887,
      "learning_rate": 1e-05,
      "loss": 0.44,
      "mean_token_accuracy": 0.8578627109527588,
      "num_tokens": 700075114.0,
      "step": 4393
    },
    {
      "epoch": 2.2349949135300102,
      "grad_norm": 0.8851991891860962,
      "learning_rate": 1e-05,
      "loss": 0.4436,
      "mean_token_accuracy": 0.8560046553611755,
      "num_tokens": 700246268.0,
      "step": 4394
    },
    {
      "epoch": 2.235503560528993,
      "grad_norm": 0.8956559896469116,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8600329160690308,
      "num_tokens": 700400215.0,
      "step": 4395
    },
    {
      "epoch": 2.236012207527976,
      "grad_norm": 0.8950102925300598,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "mean_token_accuracy": 0.858242392539978,
      "num_tokens": 700561439.0,
      "step": 4396
    },
    {
      "epoch": 2.236520854526958,
      "grad_norm": 0.8950687646865845,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8612974882125854,
      "num_tokens": 700726538.0,
      "step": 4397
    },
    {
      "epoch": 2.237029501525941,
      "grad_norm": 0.9209567904472351,
      "learning_rate": 1e-05,
      "loss": 0.4342,
      "mean_token_accuracy": 0.8593178987503052,
      "num_tokens": 700904533.0,
      "step": 4398
    },
    {
      "epoch": 2.2375381485249237,
      "grad_norm": 0.9276517629623413,
      "learning_rate": 1e-05,
      "loss": 0.4568,
      "mean_token_accuracy": 0.8548173904418945,
      "num_tokens": 701063561.0,
      "step": 4399
    },
    {
      "epoch": 2.2380467955239065,
      "grad_norm": 0.8883805871009827,
      "learning_rate": 1e-05,
      "loss": 0.459,
      "mean_token_accuracy": 0.8530396223068237,
      "num_tokens": 701220420.0,
      "step": 4400
    },
    {
      "epoch": 2.2385554425228893,
      "grad_norm": 0.9363575577735901,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8569149971008301,
      "num_tokens": 701384057.0,
      "step": 4401
    },
    {
      "epoch": 2.2390640895218716,
      "grad_norm": 0.904596745967865,
      "learning_rate": 1e-05,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8628685474395752,
      "num_tokens": 701544168.0,
      "step": 4402
    },
    {
      "epoch": 2.2395727365208544,
      "grad_norm": 0.9615168571472168,
      "learning_rate": 1e-05,
      "loss": 0.4427,
      "mean_token_accuracy": 0.8559268712997437,
      "num_tokens": 701710466.0,
      "step": 4403
    },
    {
      "epoch": 2.240081383519837,
      "grad_norm": 0.9265811443328857,
      "learning_rate": 1e-05,
      "loss": 0.4579,
      "mean_token_accuracy": 0.8512325882911682,
      "num_tokens": 701866354.0,
      "step": 4404
    },
    {
      "epoch": 2.24059003051882,
      "grad_norm": 0.8753740787506104,
      "learning_rate": 1e-05,
      "loss": 0.4098,
      "mean_token_accuracy": 0.8662247657775879,
      "num_tokens": 702021535.0,
      "step": 4405
    },
    {
      "epoch": 2.241098677517803,
      "grad_norm": 0.9615698456764221,
      "learning_rate": 1e-05,
      "loss": 0.4578,
      "mean_token_accuracy": 0.8537775278091431,
      "num_tokens": 702180586.0,
      "step": 4406
    },
    {
      "epoch": 2.241607324516785,
      "grad_norm": 0.9781903624534607,
      "learning_rate": 1e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.860724925994873,
      "num_tokens": 702337600.0,
      "step": 4407
    },
    {
      "epoch": 2.242115971515768,
      "grad_norm": 0.8906936049461365,
      "learning_rate": 1e-05,
      "loss": 0.4554,
      "mean_token_accuracy": 0.8546968102455139,
      "num_tokens": 702507782.0,
      "step": 4408
    },
    {
      "epoch": 2.2426246185147507,
      "grad_norm": 0.9217578172683716,
      "learning_rate": 1e-05,
      "loss": 0.4464,
      "mean_token_accuracy": 0.8542501926422119,
      "num_tokens": 702673172.0,
      "step": 4409
    },
    {
      "epoch": 2.2431332655137335,
      "grad_norm": 0.9569871425628662,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8487017154693604,
      "num_tokens": 702834861.0,
      "step": 4410
    },
    {
      "epoch": 2.2436419125127163,
      "grad_norm": 0.8756091594696045,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8541702628135681,
      "num_tokens": 703001374.0,
      "step": 4411
    },
    {
      "epoch": 2.244150559511699,
      "grad_norm": 1.146317481994629,
      "learning_rate": 1e-05,
      "loss": 0.4394,
      "mean_token_accuracy": 0.8578425645828247,
      "num_tokens": 703173572.0,
      "step": 4412
    },
    {
      "epoch": 2.2446592065106814,
      "grad_norm": 0.9265860319137573,
      "learning_rate": 1e-05,
      "loss": 0.4521,
      "mean_token_accuracy": 0.8535090088844299,
      "num_tokens": 703335847.0,
      "step": 4413
    },
    {
      "epoch": 2.245167853509664,
      "grad_norm": 0.8638989925384521,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8653352856636047,
      "num_tokens": 703498507.0,
      "step": 4414
    },
    {
      "epoch": 2.245676500508647,
      "grad_norm": 0.8545963168144226,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8569827079772949,
      "num_tokens": 703665569.0,
      "step": 4415
    },
    {
      "epoch": 2.2461851475076298,
      "grad_norm": 0.9048717617988586,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8589744567871094,
      "num_tokens": 703823540.0,
      "step": 4416
    },
    {
      "epoch": 2.2466937945066126,
      "grad_norm": 0.9196427464485168,
      "learning_rate": 1e-05,
      "loss": 0.452,
      "mean_token_accuracy": 0.8551338911056519,
      "num_tokens": 703996737.0,
      "step": 4417
    },
    {
      "epoch": 2.2472024415055953,
      "grad_norm": 0.9343538880348206,
      "learning_rate": 1e-05,
      "loss": 0.4369,
      "mean_token_accuracy": 0.8585703372955322,
      "num_tokens": 704156689.0,
      "step": 4418
    },
    {
      "epoch": 2.2477110885045777,
      "grad_norm": 1.2548093795776367,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8610302209854126,
      "num_tokens": 704312861.0,
      "step": 4419
    },
    {
      "epoch": 2.2482197355035605,
      "grad_norm": 0.927617073059082,
      "learning_rate": 1e-05,
      "loss": 0.4424,
      "mean_token_accuracy": 0.8566126823425293,
      "num_tokens": 704467781.0,
      "step": 4420
    },
    {
      "epoch": 2.2487283825025433,
      "grad_norm": 0.8665664196014404,
      "learning_rate": 1e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8544788360595703,
      "num_tokens": 704643910.0,
      "step": 4421
    },
    {
      "epoch": 2.249237029501526,
      "grad_norm": 0.8504548668861389,
      "learning_rate": 1e-05,
      "loss": 0.4283,
      "mean_token_accuracy": 0.8599859476089478,
      "num_tokens": 704805338.0,
      "step": 4422
    },
    {
      "epoch": 2.249745676500509,
      "grad_norm": 0.8804091215133667,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8539838194847107,
      "num_tokens": 704969847.0,
      "step": 4423
    },
    {
      "epoch": 2.250254323499491,
      "grad_norm": 0.8922030925750732,
      "learning_rate": 1e-05,
      "loss": 0.4409,
      "mean_token_accuracy": 0.8573827743530273,
      "num_tokens": 705135627.0,
      "step": 4424
    },
    {
      "epoch": 2.250762970498474,
      "grad_norm": 0.9630568623542786,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8535498380661011,
      "num_tokens": 705287010.0,
      "step": 4425
    },
    {
      "epoch": 2.2512716174974567,
      "grad_norm": 0.888202428817749,
      "learning_rate": 1e-05,
      "loss": 0.4497,
      "mean_token_accuracy": 0.8549824357032776,
      "num_tokens": 705443703.0,
      "step": 4426
    },
    {
      "epoch": 2.2517802644964395,
      "grad_norm": 0.9108331799507141,
      "learning_rate": 1e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8600156903266907,
      "num_tokens": 705610804.0,
      "step": 4427
    },
    {
      "epoch": 2.2522889114954223,
      "grad_norm": 0.9033510684967041,
      "learning_rate": 1e-05,
      "loss": 0.4293,
      "mean_token_accuracy": 0.8614766597747803,
      "num_tokens": 705767052.0,
      "step": 4428
    },
    {
      "epoch": 2.2527975584944047,
      "grad_norm": 0.8618059158325195,
      "learning_rate": 1e-05,
      "loss": 0.4284,
      "mean_token_accuracy": 0.8605534434318542,
      "num_tokens": 705941182.0,
      "step": 4429
    },
    {
      "epoch": 2.2533062054933874,
      "grad_norm": 0.89496248960495,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8533060550689697,
      "num_tokens": 706101095.0,
      "step": 4430
    },
    {
      "epoch": 2.2538148524923702,
      "grad_norm": 0.8547977805137634,
      "learning_rate": 1e-05,
      "loss": 0.4347,
      "mean_token_accuracy": 0.8595601320266724,
      "num_tokens": 706262686.0,
      "step": 4431
    },
    {
      "epoch": 2.254323499491353,
      "grad_norm": 0.8869041204452515,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8575032949447632,
      "num_tokens": 706423567.0,
      "step": 4432
    },
    {
      "epoch": 2.254832146490336,
      "grad_norm": 0.9196506142616272,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8587489724159241,
      "num_tokens": 706581619.0,
      "step": 4433
    },
    {
      "epoch": 2.2553407934893186,
      "grad_norm": 0.8818204402923584,
      "learning_rate": 1e-05,
      "loss": 0.4347,
      "mean_token_accuracy": 0.8590331077575684,
      "num_tokens": 706747002.0,
      "step": 4434
    },
    {
      "epoch": 2.255849440488301,
      "grad_norm": 0.8719470500946045,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8537620306015015,
      "num_tokens": 706919479.0,
      "step": 4435
    },
    {
      "epoch": 2.2563580874872837,
      "grad_norm": 0.9123954176902771,
      "learning_rate": 1e-05,
      "loss": 0.4182,
      "mean_token_accuracy": 0.8630255460739136,
      "num_tokens": 707083769.0,
      "step": 4436
    },
    {
      "epoch": 2.2568667344862665,
      "grad_norm": 0.8835598826408386,
      "learning_rate": 1e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8604206442832947,
      "num_tokens": 707240251.0,
      "step": 4437
    },
    {
      "epoch": 2.2573753814852493,
      "grad_norm": 0.8867762684822083,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.859366774559021,
      "num_tokens": 707407475.0,
      "step": 4438
    },
    {
      "epoch": 2.257884028484232,
      "grad_norm": 0.941148042678833,
      "learning_rate": 1e-05,
      "loss": 0.4276,
      "mean_token_accuracy": 0.8595365285873413,
      "num_tokens": 707565685.0,
      "step": 4439
    },
    {
      "epoch": 2.258392675483215,
      "grad_norm": 0.8768989443778992,
      "learning_rate": 1e-05,
      "loss": 0.4773,
      "mean_token_accuracy": 0.8487832546234131,
      "num_tokens": 707731659.0,
      "step": 4440
    },
    {
      "epoch": 2.258901322482197,
      "grad_norm": 0.9788554906845093,
      "learning_rate": 1e-05,
      "loss": 0.421,
      "mean_token_accuracy": 0.8615604639053345,
      "num_tokens": 707888682.0,
      "step": 4441
    },
    {
      "epoch": 2.25940996948118,
      "grad_norm": 0.8423621654510498,
      "learning_rate": 1e-05,
      "loss": 0.4269,
      "mean_token_accuracy": 0.8617719411849976,
      "num_tokens": 708049516.0,
      "step": 4442
    },
    {
      "epoch": 2.259918616480163,
      "grad_norm": 0.9760943651199341,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.857210636138916,
      "num_tokens": 708204574.0,
      "step": 4443
    },
    {
      "epoch": 2.2604272634791456,
      "grad_norm": 0.9054985642433167,
      "learning_rate": 1e-05,
      "loss": 0.4141,
      "mean_token_accuracy": 0.8644254207611084,
      "num_tokens": 708365608.0,
      "step": 4444
    },
    {
      "epoch": 2.2609359104781284,
      "grad_norm": 0.8849431276321411,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.858661413192749,
      "num_tokens": 708519060.0,
      "step": 4445
    },
    {
      "epoch": 2.2614445574771107,
      "grad_norm": 0.948762059211731,
      "learning_rate": 1e-05,
      "loss": 0.4068,
      "mean_token_accuracy": 0.8667903542518616,
      "num_tokens": 708668919.0,
      "step": 4446
    },
    {
      "epoch": 2.2619532044760935,
      "grad_norm": 0.9438794255256653,
      "learning_rate": 1e-05,
      "loss": 0.4186,
      "mean_token_accuracy": 0.8631476759910583,
      "num_tokens": 708825717.0,
      "step": 4447
    },
    {
      "epoch": 2.2624618514750763,
      "grad_norm": 0.9877505302429199,
      "learning_rate": 1e-05,
      "loss": 0.384,
      "mean_token_accuracy": 0.8736969828605652,
      "num_tokens": 708978992.0,
      "step": 4448
    },
    {
      "epoch": 2.262970498474059,
      "grad_norm": 0.9099939465522766,
      "learning_rate": 1e-05,
      "loss": 0.4437,
      "mean_token_accuracy": 0.857690691947937,
      "num_tokens": 709146131.0,
      "step": 4449
    },
    {
      "epoch": 2.263479145473042,
      "grad_norm": 1.0143423080444336,
      "learning_rate": 1e-05,
      "loss": 0.3974,
      "mean_token_accuracy": 0.8684823513031006,
      "num_tokens": 709289990.0,
      "step": 4450
    },
    {
      "epoch": 2.263987792472024,
      "grad_norm": 0.9650892615318298,
      "learning_rate": 1e-05,
      "loss": 0.4684,
      "mean_token_accuracy": 0.8513463735580444,
      "num_tokens": 709454058.0,
      "step": 4451
    },
    {
      "epoch": 2.264496439471007,
      "grad_norm": 0.9049556851387024,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8524051904678345,
      "num_tokens": 709620493.0,
      "step": 4452
    },
    {
      "epoch": 2.2650050864699898,
      "grad_norm": 0.9698246121406555,
      "learning_rate": 1e-05,
      "loss": 0.4363,
      "mean_token_accuracy": 0.8589226007461548,
      "num_tokens": 709793625.0,
      "step": 4453
    },
    {
      "epoch": 2.2655137334689726,
      "grad_norm": 0.9414761066436768,
      "learning_rate": 1e-05,
      "loss": 0.4402,
      "mean_token_accuracy": 0.8582148551940918,
      "num_tokens": 709945354.0,
      "step": 4454
    },
    {
      "epoch": 2.2660223804679553,
      "grad_norm": 1.0288310050964355,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8637502193450928,
      "num_tokens": 710109134.0,
      "step": 4455
    },
    {
      "epoch": 2.266531027466938,
      "grad_norm": 0.9398514032363892,
      "learning_rate": 1e-05,
      "loss": 0.4224,
      "mean_token_accuracy": 0.8614974021911621,
      "num_tokens": 710261263.0,
      "step": 4456
    },
    {
      "epoch": 2.2670396744659205,
      "grad_norm": 1.0151559114456177,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8599753975868225,
      "num_tokens": 710421327.0,
      "step": 4457
    },
    {
      "epoch": 2.2675483214649033,
      "grad_norm": 0.9492122530937195,
      "learning_rate": 1e-05,
      "loss": 0.3898,
      "mean_token_accuracy": 0.8706491589546204,
      "num_tokens": 710579962.0,
      "step": 4458
    },
    {
      "epoch": 2.268056968463886,
      "grad_norm": 0.9367904663085938,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "mean_token_accuracy": 0.859420657157898,
      "num_tokens": 710740178.0,
      "step": 4459
    },
    {
      "epoch": 2.268565615462869,
      "grad_norm": 0.9847875237464905,
      "learning_rate": 1e-05,
      "loss": 0.4442,
      "mean_token_accuracy": 0.8568622469902039,
      "num_tokens": 710918771.0,
      "step": 4460
    },
    {
      "epoch": 2.2690742624618516,
      "grad_norm": 0.8942195773124695,
      "learning_rate": 1e-05,
      "loss": 0.4174,
      "mean_token_accuracy": 0.8642637133598328,
      "num_tokens": 711075031.0,
      "step": 4461
    },
    {
      "epoch": 2.2695829094608344,
      "grad_norm": 1.0235880613327026,
      "learning_rate": 1e-05,
      "loss": 0.4537,
      "mean_token_accuracy": 0.8542577028274536,
      "num_tokens": 711248181.0,
      "step": 4462
    },
    {
      "epoch": 2.2700915564598168,
      "grad_norm": 0.9366216063499451,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8548239469528198,
      "num_tokens": 711404363.0,
      "step": 4463
    },
    {
      "epoch": 2.2706002034587995,
      "grad_norm": 0.9664139151573181,
      "learning_rate": 1e-05,
      "loss": 0.4234,
      "mean_token_accuracy": 0.8624817728996277,
      "num_tokens": 711562515.0,
      "step": 4464
    },
    {
      "epoch": 2.2711088504577823,
      "grad_norm": 0.9802444577217102,
      "learning_rate": 1e-05,
      "loss": 0.4304,
      "mean_token_accuracy": 0.8602179884910583,
      "num_tokens": 711712262.0,
      "step": 4465
    },
    {
      "epoch": 2.271617497456765,
      "grad_norm": 0.8687217831611633,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8610938787460327,
      "num_tokens": 711864605.0,
      "step": 4466
    },
    {
      "epoch": 2.272126144455748,
      "grad_norm": 1.1002277135849,
      "learning_rate": 1e-05,
      "loss": 0.461,
      "mean_token_accuracy": 0.8507526516914368,
      "num_tokens": 712030169.0,
      "step": 4467
    },
    {
      "epoch": 2.2726347914547302,
      "grad_norm": 0.965670645236969,
      "learning_rate": 1e-05,
      "loss": 0.4781,
      "mean_token_accuracy": 0.8478188514709473,
      "num_tokens": 712202336.0,
      "step": 4468
    },
    {
      "epoch": 2.273143438453713,
      "grad_norm": 0.8851255774497986,
      "learning_rate": 1e-05,
      "loss": 0.4434,
      "mean_token_accuracy": 0.8568148612976074,
      "num_tokens": 712359950.0,
      "step": 4469
    },
    {
      "epoch": 2.273652085452696,
      "grad_norm": 1.0171219110488892,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8495321273803711,
      "num_tokens": 712510845.0,
      "step": 4470
    },
    {
      "epoch": 2.2741607324516786,
      "grad_norm": 0.9366046190261841,
      "learning_rate": 1e-05,
      "loss": 0.4215,
      "mean_token_accuracy": 0.8625251650810242,
      "num_tokens": 712667175.0,
      "step": 4471
    },
    {
      "epoch": 2.2746693794506614,
      "grad_norm": 0.949114203453064,
      "learning_rate": 1e-05,
      "loss": 0.4208,
      "mean_token_accuracy": 0.8625385761260986,
      "num_tokens": 712815226.0,
      "step": 4472
    },
    {
      "epoch": 2.2751780264496437,
      "grad_norm": 0.9524646401405334,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8614475727081299,
      "num_tokens": 712966666.0,
      "step": 4473
    },
    {
      "epoch": 2.2756866734486265,
      "grad_norm": 0.8781487941741943,
      "learning_rate": 1e-05,
      "loss": 0.4488,
      "mean_token_accuracy": 0.8545585870742798,
      "num_tokens": 713136477.0,
      "step": 4474
    },
    {
      "epoch": 2.2761953204476093,
      "grad_norm": 0.9310004711151123,
      "learning_rate": 1e-05,
      "loss": 0.4312,
      "mean_token_accuracy": 0.8581027984619141,
      "num_tokens": 713294399.0,
      "step": 4475
    },
    {
      "epoch": 2.276703967446592,
      "grad_norm": 0.8774072527885437,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "mean_token_accuracy": 0.8643208742141724,
      "num_tokens": 713456971.0,
      "step": 4476
    },
    {
      "epoch": 2.277212614445575,
      "grad_norm": 0.8634635806083679,
      "learning_rate": 1e-05,
      "loss": 0.4297,
      "mean_token_accuracy": 0.8596638441085815,
      "num_tokens": 713616120.0,
      "step": 4477
    },
    {
      "epoch": 2.2777212614445577,
      "grad_norm": 0.9329019784927368,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8485960364341736,
      "num_tokens": 713779256.0,
      "step": 4478
    },
    {
      "epoch": 2.27822990844354,
      "grad_norm": 0.9431872367858887,
      "learning_rate": 1e-05,
      "loss": 0.4389,
      "mean_token_accuracy": 0.8568741083145142,
      "num_tokens": 713937883.0,
      "step": 4479
    },
    {
      "epoch": 2.278738555442523,
      "grad_norm": 0.8608583807945251,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8544870615005493,
      "num_tokens": 714098535.0,
      "step": 4480
    },
    {
      "epoch": 2.2792472024415056,
      "grad_norm": 0.9586653113365173,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8556101322174072,
      "num_tokens": 714256873.0,
      "step": 4481
    },
    {
      "epoch": 2.2797558494404884,
      "grad_norm": 0.9250736236572266,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8535689115524292,
      "num_tokens": 714415903.0,
      "step": 4482
    },
    {
      "epoch": 2.280264496439471,
      "grad_norm": 1.0306283235549927,
      "learning_rate": 1e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8542593121528625,
      "num_tokens": 714569794.0,
      "step": 4483
    },
    {
      "epoch": 2.280773143438454,
      "grad_norm": 0.9855635166168213,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8572566509246826,
      "num_tokens": 714721963.0,
      "step": 4484
    },
    {
      "epoch": 2.2812817904374363,
      "grad_norm": 0.9278120994567871,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8632841110229492,
      "num_tokens": 714882142.0,
      "step": 4485
    },
    {
      "epoch": 2.281790437436419,
      "grad_norm": 0.9876662492752075,
      "learning_rate": 1e-05,
      "loss": 0.4712,
      "mean_token_accuracy": 0.8483551144599915,
      "num_tokens": 715036265.0,
      "step": 4486
    },
    {
      "epoch": 2.282299084435402,
      "grad_norm": 0.9069980382919312,
      "learning_rate": 1e-05,
      "loss": 0.4629,
      "mean_token_accuracy": 0.84922194480896,
      "num_tokens": 715196287.0,
      "step": 4487
    },
    {
      "epoch": 2.2828077314343846,
      "grad_norm": 0.907392680644989,
      "learning_rate": 1e-05,
      "loss": 0.426,
      "mean_token_accuracy": 0.8612542152404785,
      "num_tokens": 715351692.0,
      "step": 4488
    },
    {
      "epoch": 2.2833163784333674,
      "grad_norm": 0.9390474557876587,
      "learning_rate": 1e-05,
      "loss": 0.4124,
      "mean_token_accuracy": 0.8654056787490845,
      "num_tokens": 715509487.0,
      "step": 4489
    },
    {
      "epoch": 2.2838250254323498,
      "grad_norm": 0.8880806565284729,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.848419725894928,
      "num_tokens": 715667028.0,
      "step": 4490
    },
    {
      "epoch": 2.2843336724313326,
      "grad_norm": 0.8902788758277893,
      "learning_rate": 1e-05,
      "loss": 0.4285,
      "mean_token_accuracy": 0.8602020144462585,
      "num_tokens": 715828741.0,
      "step": 4491
    },
    {
      "epoch": 2.2848423194303153,
      "grad_norm": 0.9005802273750305,
      "learning_rate": 1e-05,
      "loss": 0.4398,
      "mean_token_accuracy": 0.857980489730835,
      "num_tokens": 715978519.0,
      "step": 4492
    },
    {
      "epoch": 2.285350966429298,
      "grad_norm": 0.8823654055595398,
      "learning_rate": 1e-05,
      "loss": 0.4389,
      "mean_token_accuracy": 0.8587138056755066,
      "num_tokens": 716138914.0,
      "step": 4493
    },
    {
      "epoch": 2.285859613428281,
      "grad_norm": 0.941698431968689,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8479748964309692,
      "num_tokens": 716294789.0,
      "step": 4494
    },
    {
      "epoch": 2.2863682604272633,
      "grad_norm": 0.9303914904594421,
      "learning_rate": 1e-05,
      "loss": 0.43,
      "mean_token_accuracy": 0.8607516288757324,
      "num_tokens": 716454119.0,
      "step": 4495
    },
    {
      "epoch": 2.286876907426246,
      "grad_norm": 0.9308568239212036,
      "learning_rate": 1e-05,
      "loss": 0.4225,
      "mean_token_accuracy": 0.8629610538482666,
      "num_tokens": 716601971.0,
      "step": 4496
    },
    {
      "epoch": 2.287385554425229,
      "grad_norm": 0.9544360041618347,
      "learning_rate": 1e-05,
      "loss": 0.4428,
      "mean_token_accuracy": 0.855931282043457,
      "num_tokens": 716762602.0,
      "step": 4497
    },
    {
      "epoch": 2.2878942014242116,
      "grad_norm": 0.8890806436538696,
      "learning_rate": 1e-05,
      "loss": 0.4297,
      "mean_token_accuracy": 0.8613079786300659,
      "num_tokens": 716916644.0,
      "step": 4498
    },
    {
      "epoch": 2.2884028484231944,
      "grad_norm": 0.8517684936523438,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8590466380119324,
      "num_tokens": 717073633.0,
      "step": 4499
    },
    {
      "epoch": 2.288911495422177,
      "grad_norm": 0.9687017202377319,
      "learning_rate": 1e-05,
      "loss": 0.4184,
      "mean_token_accuracy": 0.8640453219413757,
      "num_tokens": 717218108.0,
      "step": 4500
    },
    {
      "epoch": 2.2894201424211595,
      "grad_norm": 0.8814558386802673,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8642098903656006,
      "num_tokens": 717365072.0,
      "step": 4501
    },
    {
      "epoch": 2.2899287894201423,
      "grad_norm": 0.9428908228874207,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.855121374130249,
      "num_tokens": 717517090.0,
      "step": 4502
    },
    {
      "epoch": 2.290437436419125,
      "grad_norm": 0.9154829978942871,
      "learning_rate": 1e-05,
      "loss": 0.4381,
      "mean_token_accuracy": 0.857797384262085,
      "num_tokens": 717675365.0,
      "step": 4503
    },
    {
      "epoch": 2.290946083418108,
      "grad_norm": 0.8993266224861145,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.8522711992263794,
      "num_tokens": 717846642.0,
      "step": 4504
    },
    {
      "epoch": 2.2914547304170907,
      "grad_norm": 0.8822490572929382,
      "learning_rate": 1e-05,
      "loss": 0.4352,
      "mean_token_accuracy": 0.8591576814651489,
      "num_tokens": 718021255.0,
      "step": 4505
    },
    {
      "epoch": 2.2919633774160735,
      "grad_norm": 0.9621962308883667,
      "learning_rate": 1e-05,
      "loss": 0.4211,
      "mean_token_accuracy": 0.861912190914154,
      "num_tokens": 718173696.0,
      "step": 4506
    },
    {
      "epoch": 2.292472024415056,
      "grad_norm": 0.9636359810829163,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8525373935699463,
      "num_tokens": 718326779.0,
      "step": 4507
    },
    {
      "epoch": 2.2929806714140386,
      "grad_norm": 0.9130626320838928,
      "learning_rate": 1e-05,
      "loss": 0.4677,
      "mean_token_accuracy": 0.8482925891876221,
      "num_tokens": 718494164.0,
      "step": 4508
    },
    {
      "epoch": 2.2934893184130214,
      "grad_norm": 0.9200655221939087,
      "learning_rate": 1e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8560786247253418,
      "num_tokens": 718652480.0,
      "step": 4509
    },
    {
      "epoch": 2.293997965412004,
      "grad_norm": 0.8752954602241516,
      "learning_rate": 1e-05,
      "loss": 0.4355,
      "mean_token_accuracy": 0.8592425584793091,
      "num_tokens": 718815468.0,
      "step": 4510
    },
    {
      "epoch": 2.294506612410987,
      "grad_norm": 0.8763068914413452,
      "learning_rate": 1e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8521791696548462,
      "num_tokens": 718981560.0,
      "step": 4511
    },
    {
      "epoch": 2.2950152594099693,
      "grad_norm": 0.9308648109436035,
      "learning_rate": 1e-05,
      "loss": 0.4639,
      "mean_token_accuracy": 0.8517353534698486,
      "num_tokens": 719136494.0,
      "step": 4512
    },
    {
      "epoch": 2.295523906408952,
      "grad_norm": 0.9274827241897583,
      "learning_rate": 1e-05,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8475729823112488,
      "num_tokens": 719310020.0,
      "step": 4513
    },
    {
      "epoch": 2.296032553407935,
      "grad_norm": 0.8966953158378601,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8623745441436768,
      "num_tokens": 719474769.0,
      "step": 4514
    },
    {
      "epoch": 2.2965412004069177,
      "grad_norm": 0.9108511209487915,
      "learning_rate": 1e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.856505274772644,
      "num_tokens": 719627800.0,
      "step": 4515
    },
    {
      "epoch": 2.2970498474059005,
      "grad_norm": 0.9913238883018494,
      "learning_rate": 1e-05,
      "loss": 0.4565,
      "mean_token_accuracy": 0.8522071838378906,
      "num_tokens": 719775158.0,
      "step": 4516
    },
    {
      "epoch": 2.297558494404883,
      "grad_norm": 0.931964099407196,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8504119515419006,
      "num_tokens": 719928924.0,
      "step": 4517
    },
    {
      "epoch": 2.2980671414038656,
      "grad_norm": 0.9912185668945312,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8640331625938416,
      "num_tokens": 720075628.0,
      "step": 4518
    },
    {
      "epoch": 2.2985757884028484,
      "grad_norm": 1.0474159717559814,
      "learning_rate": 1e-05,
      "loss": 0.4772,
      "mean_token_accuracy": 0.8476606011390686,
      "num_tokens": 720233916.0,
      "step": 4519
    },
    {
      "epoch": 2.299084435401831,
      "grad_norm": 0.9056844711303711,
      "learning_rate": 1e-05,
      "loss": 0.4338,
      "mean_token_accuracy": 0.8600153923034668,
      "num_tokens": 720390860.0,
      "step": 4520
    },
    {
      "epoch": 2.299593082400814,
      "grad_norm": 0.9942348003387451,
      "learning_rate": 1e-05,
      "loss": 0.4526,
      "mean_token_accuracy": 0.856066882610321,
      "num_tokens": 720561398.0,
      "step": 4521
    },
    {
      "epoch": 2.3001017293997967,
      "grad_norm": 1.002570629119873,
      "learning_rate": 1e-05,
      "loss": 0.4321,
      "mean_token_accuracy": 0.8591198921203613,
      "num_tokens": 720717089.0,
      "step": 4522
    },
    {
      "epoch": 2.300610376398779,
      "grad_norm": 0.9176241755485535,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.8634848594665527,
      "num_tokens": 720878498.0,
      "step": 4523
    },
    {
      "epoch": 2.301119023397762,
      "grad_norm": 1.0189260244369507,
      "learning_rate": 1e-05,
      "loss": 0.435,
      "mean_token_accuracy": 0.8588002324104309,
      "num_tokens": 721053638.0,
      "step": 4524
    },
    {
      "epoch": 2.3016276703967447,
      "grad_norm": 0.8975340127944946,
      "learning_rate": 1e-05,
      "loss": 0.4415,
      "mean_token_accuracy": 0.8558616638183594,
      "num_tokens": 721223050.0,
      "step": 4525
    },
    {
      "epoch": 2.3021363173957274,
      "grad_norm": 0.9887248873710632,
      "learning_rate": 1e-05,
      "loss": 0.4332,
      "mean_token_accuracy": 0.857849657535553,
      "num_tokens": 721379440.0,
      "step": 4526
    },
    {
      "epoch": 2.3026449643947102,
      "grad_norm": 0.9114660024642944,
      "learning_rate": 1e-05,
      "loss": 0.4479,
      "mean_token_accuracy": 0.8539832234382629,
      "num_tokens": 721540215.0,
      "step": 4527
    },
    {
      "epoch": 2.303153611393693,
      "grad_norm": 1.0104368925094604,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8585807085037231,
      "num_tokens": 721691506.0,
      "step": 4528
    },
    {
      "epoch": 2.3036622583926754,
      "grad_norm": 0.8985596895217896,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8575284481048584,
      "num_tokens": 721850085.0,
      "step": 4529
    },
    {
      "epoch": 2.304170905391658,
      "grad_norm": 0.9269108176231384,
      "learning_rate": 1e-05,
      "loss": 0.4415,
      "mean_token_accuracy": 0.8557599782943726,
      "num_tokens": 721991138.0,
      "step": 4530
    },
    {
      "epoch": 2.304679552390641,
      "grad_norm": 0.898306131362915,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8564971685409546,
      "num_tokens": 722158468.0,
      "step": 4531
    },
    {
      "epoch": 2.3051881993896237,
      "grad_norm": 0.9220131039619446,
      "learning_rate": 1e-05,
      "loss": 0.4573,
      "mean_token_accuracy": 0.8531209826469421,
      "num_tokens": 722325029.0,
      "step": 4532
    },
    {
      "epoch": 2.3056968463886065,
      "grad_norm": 0.9147054553031921,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.8561792969703674,
      "num_tokens": 722478424.0,
      "step": 4533
    },
    {
      "epoch": 2.306205493387589,
      "grad_norm": 0.9304201006889343,
      "learning_rate": 1e-05,
      "loss": 0.4223,
      "mean_token_accuracy": 0.8611265420913696,
      "num_tokens": 722639326.0,
      "step": 4534
    },
    {
      "epoch": 2.3067141403865716,
      "grad_norm": 0.9701643586158752,
      "learning_rate": 1e-05,
      "loss": 0.4742,
      "mean_token_accuracy": 0.8480482697486877,
      "num_tokens": 722788846.0,
      "step": 4535
    },
    {
      "epoch": 2.3072227873855544,
      "grad_norm": 0.9791626930236816,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8625706434249878,
      "num_tokens": 722945492.0,
      "step": 4536
    },
    {
      "epoch": 2.307731434384537,
      "grad_norm": 1.2803678512573242,
      "learning_rate": 1e-05,
      "loss": 0.4417,
      "mean_token_accuracy": 0.8556251525878906,
      "num_tokens": 723112669.0,
      "step": 4537
    },
    {
      "epoch": 2.30824008138352,
      "grad_norm": 1.0095150470733643,
      "learning_rate": 1e-05,
      "loss": 0.5039,
      "mean_token_accuracy": 0.839377760887146,
      "num_tokens": 723262253.0,
      "step": 4538
    },
    {
      "epoch": 2.3087487283825023,
      "grad_norm": 0.8973872065544128,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.850966215133667,
      "num_tokens": 723418494.0,
      "step": 4539
    },
    {
      "epoch": 2.309257375381485,
      "grad_norm": 0.9098710417747498,
      "learning_rate": 1e-05,
      "loss": 0.4593,
      "mean_token_accuracy": 0.85355544090271,
      "num_tokens": 723575991.0,
      "step": 4540
    },
    {
      "epoch": 2.309766022380468,
      "grad_norm": 0.9440922737121582,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8534846305847168,
      "num_tokens": 723732068.0,
      "step": 4541
    },
    {
      "epoch": 2.3102746693794507,
      "grad_norm": 0.9420719742774963,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8521120548248291,
      "num_tokens": 723892353.0,
      "step": 4542
    },
    {
      "epoch": 2.3107833163784335,
      "grad_norm": 0.8923868536949158,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8646926283836365,
      "num_tokens": 724053937.0,
      "step": 4543
    },
    {
      "epoch": 2.311291963377416,
      "grad_norm": 0.9159566164016724,
      "learning_rate": 1e-05,
      "loss": 0.4475,
      "mean_token_accuracy": 0.855766773223877,
      "num_tokens": 724220225.0,
      "step": 4544
    },
    {
      "epoch": 2.3118006103763986,
      "grad_norm": 0.9505776762962341,
      "learning_rate": 1e-05,
      "loss": 0.4641,
      "mean_token_accuracy": 0.8512469530105591,
      "num_tokens": 724386972.0,
      "step": 4545
    },
    {
      "epoch": 2.3123092573753814,
      "grad_norm": 1.0132519006729126,
      "learning_rate": 1e-05,
      "loss": 0.4338,
      "mean_token_accuracy": 0.8611284494400024,
      "num_tokens": 724542617.0,
      "step": 4546
    },
    {
      "epoch": 2.312817904374364,
      "grad_norm": 0.8784273862838745,
      "learning_rate": 1e-05,
      "loss": 0.4289,
      "mean_token_accuracy": 0.8610193133354187,
      "num_tokens": 724718591.0,
      "step": 4547
    },
    {
      "epoch": 2.313326551373347,
      "grad_norm": 0.9775454998016357,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8595489859580994,
      "num_tokens": 724863471.0,
      "step": 4548
    },
    {
      "epoch": 2.3138351983723298,
      "grad_norm": 0.8906097412109375,
      "learning_rate": 1e-05,
      "loss": 0.4691,
      "mean_token_accuracy": 0.850866436958313,
      "num_tokens": 725027948.0,
      "step": 4549
    },
    {
      "epoch": 2.3143438453713125,
      "grad_norm": 0.9929198622703552,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "mean_token_accuracy": 0.8580214977264404,
      "num_tokens": 725205147.0,
      "step": 4550
    },
    {
      "epoch": 2.314852492370295,
      "grad_norm": 0.8838443756103516,
      "learning_rate": 1e-05,
      "loss": 0.4399,
      "mean_token_accuracy": 0.8587745428085327,
      "num_tokens": 725371502.0,
      "step": 4551
    },
    {
      "epoch": 2.3153611393692777,
      "grad_norm": 0.953943133354187,
      "learning_rate": 1e-05,
      "loss": 0.438,
      "mean_token_accuracy": 0.8588062524795532,
      "num_tokens": 725544049.0,
      "step": 4552
    },
    {
      "epoch": 2.3158697863682605,
      "grad_norm": 0.9285664558410645,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.857245922088623,
      "num_tokens": 725706763.0,
      "step": 4553
    },
    {
      "epoch": 2.3163784333672433,
      "grad_norm": 0.9023874998092651,
      "learning_rate": 1e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8561683893203735,
      "num_tokens": 725863765.0,
      "step": 4554
    },
    {
      "epoch": 2.316887080366226,
      "grad_norm": 1.0573559999465942,
      "learning_rate": 1e-05,
      "loss": 0.4623,
      "mean_token_accuracy": 0.851728081703186,
      "num_tokens": 726017034.0,
      "step": 4555
    },
    {
      "epoch": 2.3173957273652084,
      "grad_norm": 0.8584126830101013,
      "learning_rate": 1e-05,
      "loss": 0.437,
      "mean_token_accuracy": 0.8590702414512634,
      "num_tokens": 726172877.0,
      "step": 4556
    },
    {
      "epoch": 2.317904374364191,
      "grad_norm": 0.9466336369514465,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8593477010726929,
      "num_tokens": 726326358.0,
      "step": 4557
    },
    {
      "epoch": 2.318413021363174,
      "grad_norm": 0.9617027640342712,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8561615943908691,
      "num_tokens": 726475413.0,
      "step": 4558
    },
    {
      "epoch": 2.3189216683621567,
      "grad_norm": 0.9507430791854858,
      "learning_rate": 1e-05,
      "loss": 0.4893,
      "mean_token_accuracy": 0.8450244665145874,
      "num_tokens": 726644059.0,
      "step": 4559
    },
    {
      "epoch": 2.3194303153611395,
      "grad_norm": 0.9864686727523804,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.8587028980255127,
      "num_tokens": 726809779.0,
      "step": 4560
    },
    {
      "epoch": 2.319938962360122,
      "grad_norm": 0.9368869662284851,
      "learning_rate": 1e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8631230592727661,
      "num_tokens": 726970762.0,
      "step": 4561
    },
    {
      "epoch": 2.3204476093591047,
      "grad_norm": 0.9573301076889038,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8558048605918884,
      "num_tokens": 727125760.0,
      "step": 4562
    },
    {
      "epoch": 2.3209562563580874,
      "grad_norm": 0.9050489068031311,
      "learning_rate": 1e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.85848069190979,
      "num_tokens": 727284662.0,
      "step": 4563
    },
    {
      "epoch": 2.3214649033570702,
      "grad_norm": 1.0446038246154785,
      "learning_rate": 1e-05,
      "loss": 0.454,
      "mean_token_accuracy": 0.8540219664573669,
      "num_tokens": 727438075.0,
      "step": 4564
    },
    {
      "epoch": 2.321973550356053,
      "grad_norm": 0.9578843116760254,
      "learning_rate": 1e-05,
      "loss": 0.4557,
      "mean_token_accuracy": 0.8550673723220825,
      "num_tokens": 727597413.0,
      "step": 4565
    },
    {
      "epoch": 2.3224821973550354,
      "grad_norm": 0.8926106691360474,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8584423065185547,
      "num_tokens": 727764115.0,
      "step": 4566
    },
    {
      "epoch": 2.322990844354018,
      "grad_norm": 0.9425592422485352,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8564541339874268,
      "num_tokens": 727931105.0,
      "step": 4567
    },
    {
      "epoch": 2.323499491353001,
      "grad_norm": 0.858015775680542,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.857099175453186,
      "num_tokens": 728092211.0,
      "step": 4568
    },
    {
      "epoch": 2.3240081383519837,
      "grad_norm": 0.9361171126365662,
      "learning_rate": 1e-05,
      "loss": 0.4751,
      "mean_token_accuracy": 0.8496037721633911,
      "num_tokens": 728258373.0,
      "step": 4569
    },
    {
      "epoch": 2.3245167853509665,
      "grad_norm": 0.9430844187736511,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.8636689186096191,
      "num_tokens": 728425475.0,
      "step": 4570
    },
    {
      "epoch": 2.3250254323499493,
      "grad_norm": 0.9531563520431519,
      "learning_rate": 1e-05,
      "loss": 0.4712,
      "mean_token_accuracy": 0.8488687872886658,
      "num_tokens": 728590364.0,
      "step": 4571
    },
    {
      "epoch": 2.325534079348932,
      "grad_norm": 0.9843347668647766,
      "learning_rate": 1e-05,
      "loss": 0.4317,
      "mean_token_accuracy": 0.8596490025520325,
      "num_tokens": 728738679.0,
      "step": 4572
    },
    {
      "epoch": 2.3260427263479144,
      "grad_norm": 0.8593404293060303,
      "learning_rate": 1e-05,
      "loss": 0.4657,
      "mean_token_accuracy": 0.84918212890625,
      "num_tokens": 728901362.0,
      "step": 4573
    },
    {
      "epoch": 2.326551373346897,
      "grad_norm": 0.8554495573043823,
      "learning_rate": 1e-05,
      "loss": 0.4677,
      "mean_token_accuracy": 0.8513251543045044,
      "num_tokens": 729077322.0,
      "step": 4574
    },
    {
      "epoch": 2.32706002034588,
      "grad_norm": 0.9305860996246338,
      "learning_rate": 1e-05,
      "loss": 0.4228,
      "mean_token_accuracy": 0.8608434200286865,
      "num_tokens": 729222577.0,
      "step": 4575
    },
    {
      "epoch": 2.327568667344863,
      "grad_norm": 0.8833337426185608,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.8577058911323547,
      "num_tokens": 729376355.0,
      "step": 4576
    },
    {
      "epoch": 2.3280773143438456,
      "grad_norm": 0.8967020511627197,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8576448559761047,
      "num_tokens": 729528421.0,
      "step": 4577
    },
    {
      "epoch": 2.328585961342828,
      "grad_norm": 0.8985277414321899,
      "learning_rate": 1e-05,
      "loss": 0.4433,
      "mean_token_accuracy": 0.8569909930229187,
      "num_tokens": 729697233.0,
      "step": 4578
    },
    {
      "epoch": 2.3290946083418107,
      "grad_norm": 0.8995670676231384,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8587366938591003,
      "num_tokens": 729850719.0,
      "step": 4579
    },
    {
      "epoch": 2.3296032553407935,
      "grad_norm": 0.9012477397918701,
      "learning_rate": 1e-05,
      "loss": 0.4673,
      "mean_token_accuracy": 0.8507951498031616,
      "num_tokens": 730009711.0,
      "step": 4580
    },
    {
      "epoch": 2.3301119023397763,
      "grad_norm": 0.8857806921005249,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8554847240447998,
      "num_tokens": 730162457.0,
      "step": 4581
    },
    {
      "epoch": 2.330620549338759,
      "grad_norm": 0.9476555585861206,
      "learning_rate": 1e-05,
      "loss": 0.4753,
      "mean_token_accuracy": 0.8466869592666626,
      "num_tokens": 730315273.0,
      "step": 4582
    },
    {
      "epoch": 2.3311291963377414,
      "grad_norm": 0.8659005165100098,
      "learning_rate": 1e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.8618274927139282,
      "num_tokens": 730464801.0,
      "step": 4583
    },
    {
      "epoch": 2.331637843336724,
      "grad_norm": 0.8836120963096619,
      "learning_rate": 1e-05,
      "loss": 0.4251,
      "mean_token_accuracy": 0.8608165979385376,
      "num_tokens": 730622310.0,
      "step": 4584
    },
    {
      "epoch": 2.332146490335707,
      "grad_norm": 0.8895693421363831,
      "learning_rate": 1e-05,
      "loss": 0.4398,
      "mean_token_accuracy": 0.8577582836151123,
      "num_tokens": 730778206.0,
      "step": 4585
    },
    {
      "epoch": 2.3326551373346898,
      "grad_norm": 0.9397143721580505,
      "learning_rate": 1e-05,
      "loss": 0.4631,
      "mean_token_accuracy": 0.8512009382247925,
      "num_tokens": 730938482.0,
      "step": 4586
    },
    {
      "epoch": 2.3331637843336726,
      "grad_norm": 0.9026806950569153,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "mean_token_accuracy": 0.8560445308685303,
      "num_tokens": 731100793.0,
      "step": 4587
    },
    {
      "epoch": 2.333672431332655,
      "grad_norm": 0.9613650441169739,
      "learning_rate": 1e-05,
      "loss": 0.4775,
      "mean_token_accuracy": 0.8481220602989197,
      "num_tokens": 731256654.0,
      "step": 4588
    },
    {
      "epoch": 2.3341810783316377,
      "grad_norm": 0.8520517349243164,
      "learning_rate": 1e-05,
      "loss": 0.4388,
      "mean_token_accuracy": 0.8573417067527771,
      "num_tokens": 731413316.0,
      "step": 4589
    },
    {
      "epoch": 2.3346897253306205,
      "grad_norm": 0.9405336976051331,
      "learning_rate": 1e-05,
      "loss": 0.4106,
      "mean_token_accuracy": 0.864769458770752,
      "num_tokens": 731560273.0,
      "step": 4590
    },
    {
      "epoch": 2.3351983723296033,
      "grad_norm": 0.879772424697876,
      "learning_rate": 1e-05,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8663976788520813,
      "num_tokens": 731716793.0,
      "step": 4591
    },
    {
      "epoch": 2.335707019328586,
      "grad_norm": 0.9004299640655518,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8643734455108643,
      "num_tokens": 731882252.0,
      "step": 4592
    },
    {
      "epoch": 2.336215666327569,
      "grad_norm": 0.8501989245414734,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8574635982513428,
      "num_tokens": 732050123.0,
      "step": 4593
    },
    {
      "epoch": 2.3367243133265516,
      "grad_norm": 0.8961136937141418,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.8601347208023071,
      "num_tokens": 732205142.0,
      "step": 4594
    },
    {
      "epoch": 2.337232960325534,
      "grad_norm": 0.9832904934883118,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8558573722839355,
      "num_tokens": 732361135.0,
      "step": 4595
    },
    {
      "epoch": 2.3377416073245167,
      "grad_norm": 0.9565850496292114,
      "learning_rate": 1e-05,
      "loss": 0.4633,
      "mean_token_accuracy": 0.8502834439277649,
      "num_tokens": 732523658.0,
      "step": 4596
    },
    {
      "epoch": 2.3382502543234995,
      "grad_norm": 0.9144861102104187,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.8545467853546143,
      "num_tokens": 732677365.0,
      "step": 4597
    },
    {
      "epoch": 2.3387589013224823,
      "grad_norm": 0.8910345435142517,
      "learning_rate": 1e-05,
      "loss": 0.438,
      "mean_token_accuracy": 0.857315182685852,
      "num_tokens": 732842797.0,
      "step": 4598
    },
    {
      "epoch": 2.339267548321465,
      "grad_norm": 0.9919068217277527,
      "learning_rate": 1e-05,
      "loss": 0.4601,
      "mean_token_accuracy": 0.8525848984718323,
      "num_tokens": 733002613.0,
      "step": 4599
    },
    {
      "epoch": 2.3397761953204474,
      "grad_norm": 0.9106639623641968,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8479093909263611,
      "num_tokens": 733168024.0,
      "step": 4600
    },
    {
      "epoch": 2.3402848423194302,
      "grad_norm": 0.9065938591957092,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8635804653167725,
      "num_tokens": 733327001.0,
      "step": 4601
    },
    {
      "epoch": 2.340793489318413,
      "grad_norm": 0.8956307172775269,
      "learning_rate": 1e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.859512984752655,
      "num_tokens": 733495169.0,
      "step": 4602
    },
    {
      "epoch": 2.341302136317396,
      "grad_norm": 0.9894669055938721,
      "learning_rate": 1e-05,
      "loss": 0.4321,
      "mean_token_accuracy": 0.8590751886367798,
      "num_tokens": 733633586.0,
      "step": 4603
    },
    {
      "epoch": 2.3418107833163786,
      "grad_norm": 0.9020395278930664,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.8553770184516907,
      "num_tokens": 733783519.0,
      "step": 4604
    },
    {
      "epoch": 2.342319430315361,
      "grad_norm": 1.063341736793518,
      "learning_rate": 1e-05,
      "loss": 0.4437,
      "mean_token_accuracy": 0.8559243679046631,
      "num_tokens": 733947060.0,
      "step": 4605
    },
    {
      "epoch": 2.3428280773143437,
      "grad_norm": 0.9173327684402466,
      "learning_rate": 1e-05,
      "loss": 0.4185,
      "mean_token_accuracy": 0.8639484643936157,
      "num_tokens": 734114541.0,
      "step": 4606
    },
    {
      "epoch": 2.3433367243133265,
      "grad_norm": 1.05625319480896,
      "learning_rate": 1e-05,
      "loss": 0.4475,
      "mean_token_accuracy": 0.8549398183822632,
      "num_tokens": 734282987.0,
      "step": 4607
    },
    {
      "epoch": 2.3438453713123093,
      "grad_norm": 0.9332970380783081,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8487000465393066,
      "num_tokens": 734444293.0,
      "step": 4608
    },
    {
      "epoch": 2.344354018311292,
      "grad_norm": 0.9928670525550842,
      "learning_rate": 1e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8533625602722168,
      "num_tokens": 734591939.0,
      "step": 4609
    },
    {
      "epoch": 2.3448626653102744,
      "grad_norm": 0.8672624230384827,
      "learning_rate": 1e-05,
      "loss": 0.4033,
      "mean_token_accuracy": 0.8678281903266907,
      "num_tokens": 734747917.0,
      "step": 4610
    },
    {
      "epoch": 2.345371312309257,
      "grad_norm": 0.969215989112854,
      "learning_rate": 1e-05,
      "loss": 0.4205,
      "mean_token_accuracy": 0.8641879558563232,
      "num_tokens": 734912507.0,
      "step": 4611
    },
    {
      "epoch": 2.34587995930824,
      "grad_norm": 0.9247586131095886,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8647682666778564,
      "num_tokens": 735069410.0,
      "step": 4612
    },
    {
      "epoch": 2.346388606307223,
      "grad_norm": 0.8570278882980347,
      "learning_rate": 1e-05,
      "loss": 0.4173,
      "mean_token_accuracy": 0.8636913299560547,
      "num_tokens": 735225465.0,
      "step": 4613
    },
    {
      "epoch": 2.3468972533062056,
      "grad_norm": 0.9816693663597107,
      "learning_rate": 1e-05,
      "loss": 0.4198,
      "mean_token_accuracy": 0.8631116151809692,
      "num_tokens": 735380532.0,
      "step": 4614
    },
    {
      "epoch": 2.3474059003051884,
      "grad_norm": 0.8836074471473694,
      "learning_rate": 1e-05,
      "loss": 0.4263,
      "mean_token_accuracy": 0.8616890907287598,
      "num_tokens": 735530373.0,
      "step": 4615
    },
    {
      "epoch": 2.347914547304171,
      "grad_norm": 0.9004747867584229,
      "learning_rate": 1e-05,
      "loss": 0.4675,
      "mean_token_accuracy": 0.8508669137954712,
      "num_tokens": 735689763.0,
      "step": 4616
    },
    {
      "epoch": 2.3484231943031535,
      "grad_norm": 0.9340812563896179,
      "learning_rate": 1e-05,
      "loss": 0.4009,
      "mean_token_accuracy": 0.8688367009162903,
      "num_tokens": 735841618.0,
      "step": 4617
    },
    {
      "epoch": 2.3489318413021363,
      "grad_norm": 0.9207538962364197,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "mean_token_accuracy": 0.8663789629936218,
      "num_tokens": 736000433.0,
      "step": 4618
    },
    {
      "epoch": 2.349440488301119,
      "grad_norm": 0.9837210178375244,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "mean_token_accuracy": 0.857857346534729,
      "num_tokens": 736157243.0,
      "step": 4619
    },
    {
      "epoch": 2.349949135300102,
      "grad_norm": 0.8667672872543335,
      "learning_rate": 1e-05,
      "loss": 0.4309,
      "mean_token_accuracy": 0.8590836524963379,
      "num_tokens": 736329409.0,
      "step": 4620
    },
    {
      "epoch": 2.3504577822990846,
      "grad_norm": 1.0480256080627441,
      "learning_rate": 1e-05,
      "loss": 0.4376,
      "mean_token_accuracy": 0.856756329536438,
      "num_tokens": 736483322.0,
      "step": 4621
    },
    {
      "epoch": 2.350966429298067,
      "grad_norm": 0.9391472935676575,
      "learning_rate": 1e-05,
      "loss": 0.4443,
      "mean_token_accuracy": 0.8562041521072388,
      "num_tokens": 736642712.0,
      "step": 4622
    },
    {
      "epoch": 2.3514750762970498,
      "grad_norm": 0.927949845790863,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "mean_token_accuracy": 0.8579003214836121,
      "num_tokens": 736795254.0,
      "step": 4623
    },
    {
      "epoch": 2.3519837232960326,
      "grad_norm": 0.9654059410095215,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "mean_token_accuracy": 0.855154812335968,
      "num_tokens": 736956462.0,
      "step": 4624
    },
    {
      "epoch": 2.3524923702950153,
      "grad_norm": 0.9332446455955505,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "mean_token_accuracy": 0.8547961115837097,
      "num_tokens": 737113166.0,
      "step": 4625
    },
    {
      "epoch": 2.353001017293998,
      "grad_norm": 0.8744509816169739,
      "learning_rate": 1e-05,
      "loss": 0.42,
      "mean_token_accuracy": 0.8633664846420288,
      "num_tokens": 737279301.0,
      "step": 4626
    },
    {
      "epoch": 2.3535096642929805,
      "grad_norm": 0.8209827542304993,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8543109893798828,
      "num_tokens": 737455486.0,
      "step": 4627
    },
    {
      "epoch": 2.3540183112919633,
      "grad_norm": 0.8656889796257019,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "mean_token_accuracy": 0.8588730692863464,
      "num_tokens": 737617291.0,
      "step": 4628
    },
    {
      "epoch": 2.354526958290946,
      "grad_norm": 0.9394766092300415,
      "learning_rate": 1e-05,
      "loss": 0.448,
      "mean_token_accuracy": 0.8561714291572571,
      "num_tokens": 737775312.0,
      "step": 4629
    },
    {
      "epoch": 2.355035605289929,
      "grad_norm": 0.8893435597419739,
      "learning_rate": 1e-05,
      "loss": 0.4203,
      "mean_token_accuracy": 0.862876296043396,
      "num_tokens": 737938881.0,
      "step": 4630
    },
    {
      "epoch": 2.3555442522889116,
      "grad_norm": 0.9624534249305725,
      "learning_rate": 1e-05,
      "loss": 0.4545,
      "mean_token_accuracy": 0.8524142503738403,
      "num_tokens": 738091397.0,
      "step": 4631
    },
    {
      "epoch": 2.356052899287894,
      "grad_norm": 0.8696402311325073,
      "learning_rate": 1e-05,
      "loss": 0.4214,
      "mean_token_accuracy": 0.8624419569969177,
      "num_tokens": 738254476.0,
      "step": 4632
    },
    {
      "epoch": 2.3565615462868768,
      "grad_norm": 0.9505226016044617,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8569638133049011,
      "num_tokens": 738427280.0,
      "step": 4633
    },
    {
      "epoch": 2.3570701932858595,
      "grad_norm": 0.9444771409034729,
      "learning_rate": 1e-05,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8611069917678833,
      "num_tokens": 738587235.0,
      "step": 4634
    },
    {
      "epoch": 2.3575788402848423,
      "grad_norm": 0.8936687707901001,
      "learning_rate": 1e-05,
      "loss": 0.4012,
      "mean_token_accuracy": 0.8680918216705322,
      "num_tokens": 738736548.0,
      "step": 4635
    },
    {
      "epoch": 2.358087487283825,
      "grad_norm": 1.0075676441192627,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8534650802612305,
      "num_tokens": 738903821.0,
      "step": 4636
    },
    {
      "epoch": 2.358596134282808,
      "grad_norm": 0.9175559878349304,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8470964431762695,
      "num_tokens": 739056824.0,
      "step": 4637
    },
    {
      "epoch": 2.3591047812817907,
      "grad_norm": 1.0028626918792725,
      "learning_rate": 1e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8617447018623352,
      "num_tokens": 739208932.0,
      "step": 4638
    },
    {
      "epoch": 2.359613428280773,
      "grad_norm": 0.9204167723655701,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8545858860015869,
      "num_tokens": 739369912.0,
      "step": 4639
    },
    {
      "epoch": 2.360122075279756,
      "grad_norm": 1.0421146154403687,
      "learning_rate": 1e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.8625353574752808,
      "num_tokens": 739529429.0,
      "step": 4640
    },
    {
      "epoch": 2.3606307222787386,
      "grad_norm": 1.040346622467041,
      "learning_rate": 1e-05,
      "loss": 0.4208,
      "mean_token_accuracy": 0.8633586168289185,
      "num_tokens": 739682071.0,
      "step": 4641
    },
    {
      "epoch": 2.3611393692777214,
      "grad_norm": 0.914192795753479,
      "learning_rate": 1e-05,
      "loss": 0.4338,
      "mean_token_accuracy": 0.8609576225280762,
      "num_tokens": 739826514.0,
      "step": 4642
    },
    {
      "epoch": 2.361648016276704,
      "grad_norm": 0.9608312249183655,
      "learning_rate": 1e-05,
      "loss": 0.4699,
      "mean_token_accuracy": 0.8505814671516418,
      "num_tokens": 739987484.0,
      "step": 4643
    },
    {
      "epoch": 2.3621566632756865,
      "grad_norm": 0.9846158027648926,
      "learning_rate": 1e-05,
      "loss": 0.4235,
      "mean_token_accuracy": 0.8619943857192993,
      "num_tokens": 740139289.0,
      "step": 4644
    },
    {
      "epoch": 2.3626653102746693,
      "grad_norm": 0.9244502782821655,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8589105606079102,
      "num_tokens": 740295292.0,
      "step": 4645
    },
    {
      "epoch": 2.363173957273652,
      "grad_norm": 0.967990517616272,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8616007566452026,
      "num_tokens": 740449970.0,
      "step": 4646
    },
    {
      "epoch": 2.363682604272635,
      "grad_norm": 0.8939878940582275,
      "learning_rate": 1e-05,
      "loss": 0.4469,
      "mean_token_accuracy": 0.8555958271026611,
      "num_tokens": 740607621.0,
      "step": 4647
    },
    {
      "epoch": 2.3641912512716177,
      "grad_norm": 0.9425458908081055,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8575696349143982,
      "num_tokens": 740754163.0,
      "step": 4648
    },
    {
      "epoch": 2.3646998982706,
      "grad_norm": 0.9591701626777649,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8612913489341736,
      "num_tokens": 740909565.0,
      "step": 4649
    },
    {
      "epoch": 2.365208545269583,
      "grad_norm": 0.9903981685638428,
      "learning_rate": 1e-05,
      "loss": 0.4322,
      "mean_token_accuracy": 0.8616185188293457,
      "num_tokens": 741074703.0,
      "step": 4650
    },
    {
      "epoch": 2.3657171922685656,
      "grad_norm": 0.9808692932128906,
      "learning_rate": 1e-05,
      "loss": 0.4508,
      "mean_token_accuracy": 0.854599118232727,
      "num_tokens": 741232219.0,
      "step": 4651
    },
    {
      "epoch": 2.3662258392675484,
      "grad_norm": 0.874530017375946,
      "learning_rate": 1e-05,
      "loss": 0.4125,
      "mean_token_accuracy": 0.8656905889511108,
      "num_tokens": 741403525.0,
      "step": 4652
    },
    {
      "epoch": 2.366734486266531,
      "grad_norm": 0.9854841828346252,
      "learning_rate": 1e-05,
      "loss": 0.4002,
      "mean_token_accuracy": 0.8691383600234985,
      "num_tokens": 741564479.0,
      "step": 4653
    },
    {
      "epoch": 2.3672431332655135,
      "grad_norm": 0.9224373698234558,
      "learning_rate": 1e-05,
      "loss": 0.4258,
      "mean_token_accuracy": 0.8609287738800049,
      "num_tokens": 741728453.0,
      "step": 4654
    },
    {
      "epoch": 2.3677517802644963,
      "grad_norm": 0.9162703156471252,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "mean_token_accuracy": 0.8567160367965698,
      "num_tokens": 741877478.0,
      "step": 4655
    },
    {
      "epoch": 2.368260427263479,
      "grad_norm": 1.03668212890625,
      "learning_rate": 1e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8631287217140198,
      "num_tokens": 742045389.0,
      "step": 4656
    },
    {
      "epoch": 2.368769074262462,
      "grad_norm": 0.8980605006217957,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8585948348045349,
      "num_tokens": 742197956.0,
      "step": 4657
    },
    {
      "epoch": 2.3692777212614446,
      "grad_norm": 0.9165881872177124,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8621605038642883,
      "num_tokens": 742355845.0,
      "step": 4658
    },
    {
      "epoch": 2.3697863682604274,
      "grad_norm": 0.8757665753364563,
      "learning_rate": 1e-05,
      "loss": 0.4593,
      "mean_token_accuracy": 0.851668119430542,
      "num_tokens": 742516329.0,
      "step": 4659
    },
    {
      "epoch": 2.3702950152594098,
      "grad_norm": 0.935166597366333,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8546200394630432,
      "num_tokens": 742671682.0,
      "step": 4660
    },
    {
      "epoch": 2.3708036622583926,
      "grad_norm": 1.0466597080230713,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8509244918823242,
      "num_tokens": 742824141.0,
      "step": 4661
    },
    {
      "epoch": 2.3713123092573754,
      "grad_norm": 0.8813978433609009,
      "learning_rate": 1e-05,
      "loss": 0.4489,
      "mean_token_accuracy": 0.8555780649185181,
      "num_tokens": 742999389.0,
      "step": 4662
    },
    {
      "epoch": 2.371820956256358,
      "grad_norm": 0.8976166248321533,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8543903827667236,
      "num_tokens": 743157150.0,
      "step": 4663
    },
    {
      "epoch": 2.372329603255341,
      "grad_norm": 0.9232761859893799,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "mean_token_accuracy": 0.8589328527450562,
      "num_tokens": 743328436.0,
      "step": 4664
    },
    {
      "epoch": 2.3728382502543237,
      "grad_norm": 0.8817173838615417,
      "learning_rate": 1e-05,
      "loss": 0.4234,
      "mean_token_accuracy": 0.8617080450057983,
      "num_tokens": 743481174.0,
      "step": 4665
    },
    {
      "epoch": 2.373346897253306,
      "grad_norm": 1.13491952419281,
      "learning_rate": 1e-05,
      "loss": 0.4597,
      "mean_token_accuracy": 0.8530697822570801,
      "num_tokens": 743642776.0,
      "step": 4666
    },
    {
      "epoch": 2.373855544252289,
      "grad_norm": 1.4366291761398315,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.8627820014953613,
      "num_tokens": 743809658.0,
      "step": 4667
    },
    {
      "epoch": 2.3743641912512716,
      "grad_norm": 0.9859811067581177,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8563073873519897,
      "num_tokens": 743962306.0,
      "step": 4668
    },
    {
      "epoch": 2.3748728382502544,
      "grad_norm": 0.9425814151763916,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8557022213935852,
      "num_tokens": 744117518.0,
      "step": 4669
    },
    {
      "epoch": 2.375381485249237,
      "grad_norm": 0.9705214500427246,
      "learning_rate": 1e-05,
      "loss": 0.4738,
      "mean_token_accuracy": 0.850643515586853,
      "num_tokens": 744277189.0,
      "step": 4670
    },
    {
      "epoch": 2.3758901322482195,
      "grad_norm": 0.9333838820457458,
      "learning_rate": 1e-05,
      "loss": 0.4614,
      "mean_token_accuracy": 0.8518938422203064,
      "num_tokens": 744432354.0,
      "step": 4671
    },
    {
      "epoch": 2.3763987792472023,
      "grad_norm": 0.8503909111022949,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "mean_token_accuracy": 0.8648768067359924,
      "num_tokens": 744602450.0,
      "step": 4672
    },
    {
      "epoch": 2.376907426246185,
      "grad_norm": 0.9568102955818176,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8558696508407593,
      "num_tokens": 744763687.0,
      "step": 4673
    },
    {
      "epoch": 2.377416073245168,
      "grad_norm": 0.9368235468864441,
      "learning_rate": 1e-05,
      "loss": 0.4323,
      "mean_token_accuracy": 0.859917938709259,
      "num_tokens": 744941184.0,
      "step": 4674
    },
    {
      "epoch": 2.3779247202441507,
      "grad_norm": 1.0416042804718018,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8539679050445557,
      "num_tokens": 745110653.0,
      "step": 4675
    },
    {
      "epoch": 2.378433367243133,
      "grad_norm": 0.9496248364448547,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.8581831455230713,
      "num_tokens": 745273816.0,
      "step": 4676
    },
    {
      "epoch": 2.378942014242116,
      "grad_norm": 0.9366162419319153,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.854411780834198,
      "num_tokens": 745424921.0,
      "step": 4677
    },
    {
      "epoch": 2.3794506612410986,
      "grad_norm": 0.8841323852539062,
      "learning_rate": 1e-05,
      "loss": 0.466,
      "mean_token_accuracy": 0.8519430756568909,
      "num_tokens": 745600889.0,
      "step": 4678
    },
    {
      "epoch": 2.3799593082400814,
      "grad_norm": 0.9119141101837158,
      "learning_rate": 1e-05,
      "loss": 0.4206,
      "mean_token_accuracy": 0.8637368083000183,
      "num_tokens": 745766797.0,
      "step": 4679
    },
    {
      "epoch": 2.380467955239064,
      "grad_norm": 1.0425060987472534,
      "learning_rate": 1e-05,
      "loss": 0.4193,
      "mean_token_accuracy": 0.8617416024208069,
      "num_tokens": 745931100.0,
      "step": 4680
    },
    {
      "epoch": 2.380976602238047,
      "grad_norm": 0.9240050315856934,
      "learning_rate": 1e-05,
      "loss": 0.4364,
      "mean_token_accuracy": 0.8603978157043457,
      "num_tokens": 746090202.0,
      "step": 4681
    },
    {
      "epoch": 2.3814852492370293,
      "grad_norm": 0.9670682549476624,
      "learning_rate": 1e-05,
      "loss": 0.4434,
      "mean_token_accuracy": 0.8564561605453491,
      "num_tokens": 746255957.0,
      "step": 4682
    },
    {
      "epoch": 2.381993896236012,
      "grad_norm": 0.9642400145530701,
      "learning_rate": 1e-05,
      "loss": 0.4271,
      "mean_token_accuracy": 0.8614307641983032,
      "num_tokens": 746420006.0,
      "step": 4683
    },
    {
      "epoch": 2.382502543234995,
      "grad_norm": 0.9733436107635498,
      "learning_rate": 1e-05,
      "loss": 0.4519,
      "mean_token_accuracy": 0.8539668321609497,
      "num_tokens": 746589548.0,
      "step": 4684
    },
    {
      "epoch": 2.3830111902339777,
      "grad_norm": 0.9867450594902039,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.8615120649337769,
      "num_tokens": 746746846.0,
      "step": 4685
    },
    {
      "epoch": 2.3835198372329605,
      "grad_norm": 0.9444828629493713,
      "learning_rate": 1e-05,
      "loss": 0.4395,
      "mean_token_accuracy": 0.8575142025947571,
      "num_tokens": 746894866.0,
      "step": 4686
    },
    {
      "epoch": 2.3840284842319432,
      "grad_norm": 1.0291507244110107,
      "learning_rate": 1e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8585408926010132,
      "num_tokens": 747055985.0,
      "step": 4687
    },
    {
      "epoch": 2.3845371312309256,
      "grad_norm": 0.850340723991394,
      "learning_rate": 1e-05,
      "loss": 0.4064,
      "mean_token_accuracy": 0.8683385848999023,
      "num_tokens": 747223779.0,
      "step": 4688
    },
    {
      "epoch": 2.3850457782299084,
      "grad_norm": 1.0372014045715332,
      "learning_rate": 1e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.863206148147583,
      "num_tokens": 747378261.0,
      "step": 4689
    },
    {
      "epoch": 2.385554425228891,
      "grad_norm": 0.9969184994697571,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8603659868240356,
      "num_tokens": 747533998.0,
      "step": 4690
    },
    {
      "epoch": 2.386063072227874,
      "grad_norm": 0.9529933929443359,
      "learning_rate": 1e-05,
      "loss": 0.4821,
      "mean_token_accuracy": 0.8441590070724487,
      "num_tokens": 747695636.0,
      "step": 4691
    },
    {
      "epoch": 2.3865717192268567,
      "grad_norm": 1.0091328620910645,
      "learning_rate": 1e-05,
      "loss": 0.4624,
      "mean_token_accuracy": 0.851618230342865,
      "num_tokens": 747850725.0,
      "step": 4692
    },
    {
      "epoch": 2.387080366225839,
      "grad_norm": 0.9514685273170471,
      "learning_rate": 1e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.8622990846633911,
      "num_tokens": 748011157.0,
      "step": 4693
    },
    {
      "epoch": 2.387589013224822,
      "grad_norm": 0.961980402469635,
      "learning_rate": 1e-05,
      "loss": 0.4373,
      "mean_token_accuracy": 0.8580829501152039,
      "num_tokens": 748172245.0,
      "step": 4694
    },
    {
      "epoch": 2.3880976602238047,
      "grad_norm": 0.967473566532135,
      "learning_rate": 1e-05,
      "loss": 0.4206,
      "mean_token_accuracy": 0.8628791570663452,
      "num_tokens": 748325050.0,
      "step": 4695
    },
    {
      "epoch": 2.3886063072227874,
      "grad_norm": 0.9420827031135559,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8563902974128723,
      "num_tokens": 748476090.0,
      "step": 4696
    },
    {
      "epoch": 2.3891149542217702,
      "grad_norm": 0.9462102055549622,
      "learning_rate": 1e-05,
      "loss": 0.4593,
      "mean_token_accuracy": 0.8517200350761414,
      "num_tokens": 748638219.0,
      "step": 4697
    },
    {
      "epoch": 2.3896236012207526,
      "grad_norm": 0.9324440956115723,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8563007712364197,
      "num_tokens": 748802917.0,
      "step": 4698
    },
    {
      "epoch": 2.3901322482197354,
      "grad_norm": 0.9615978598594666,
      "learning_rate": 1e-05,
      "loss": 0.4408,
      "mean_token_accuracy": 0.8580653667449951,
      "num_tokens": 748955626.0,
      "step": 4699
    },
    {
      "epoch": 2.390640895218718,
      "grad_norm": 0.9752944707870483,
      "learning_rate": 1e-05,
      "loss": 0.4816,
      "mean_token_accuracy": 0.8470532894134521,
      "num_tokens": 749117584.0,
      "step": 4700
    },
    {
      "epoch": 2.391149542217701,
      "grad_norm": 0.9216363430023193,
      "learning_rate": 1e-05,
      "loss": 0.4566,
      "mean_token_accuracy": 0.854560375213623,
      "num_tokens": 749275558.0,
      "step": 4701
    },
    {
      "epoch": 2.3916581892166837,
      "grad_norm": 0.9103134274482727,
      "learning_rate": 1e-05,
      "loss": 0.4513,
      "mean_token_accuracy": 0.8529074192047119,
      "num_tokens": 749440693.0,
      "step": 4702
    },
    {
      "epoch": 2.3921668362156665,
      "grad_norm": 0.9516760110855103,
      "learning_rate": 1e-05,
      "loss": 0.4196,
      "mean_token_accuracy": 0.8639135360717773,
      "num_tokens": 749592575.0,
      "step": 4703
    },
    {
      "epoch": 2.392675483214649,
      "grad_norm": 0.9901017546653748,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "mean_token_accuracy": 0.8616241812705994,
      "num_tokens": 749746118.0,
      "step": 4704
    },
    {
      "epoch": 2.3931841302136316,
      "grad_norm": 0.8785843849182129,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "mean_token_accuracy": 0.8633971214294434,
      "num_tokens": 749904411.0,
      "step": 4705
    },
    {
      "epoch": 2.3936927772126144,
      "grad_norm": 0.88527512550354,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8575128316879272,
      "num_tokens": 750064386.0,
      "step": 4706
    },
    {
      "epoch": 2.394201424211597,
      "grad_norm": 0.9171162843704224,
      "learning_rate": 1e-05,
      "loss": 0.4385,
      "mean_token_accuracy": 0.8575048446655273,
      "num_tokens": 750223529.0,
      "step": 4707
    },
    {
      "epoch": 2.39471007121058,
      "grad_norm": 0.8819383382797241,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8574961423873901,
      "num_tokens": 750392408.0,
      "step": 4708
    },
    {
      "epoch": 2.395218718209563,
      "grad_norm": 0.8735357522964478,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8608437776565552,
      "num_tokens": 750558566.0,
      "step": 4709
    },
    {
      "epoch": 2.395727365208545,
      "grad_norm": 0.954378068447113,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "mean_token_accuracy": 0.8561128377914429,
      "num_tokens": 750725594.0,
      "step": 4710
    },
    {
      "epoch": 2.396236012207528,
      "grad_norm": 0.9977381229400635,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "mean_token_accuracy": 0.8524102568626404,
      "num_tokens": 750864466.0,
      "step": 4711
    },
    {
      "epoch": 2.3967446592065107,
      "grad_norm": 0.9387546181678772,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "mean_token_accuracy": 0.8657817840576172,
      "num_tokens": 751019674.0,
      "step": 4712
    },
    {
      "epoch": 2.3972533062054935,
      "grad_norm": 0.9221512079238892,
      "learning_rate": 1e-05,
      "loss": 0.4186,
      "mean_token_accuracy": 0.8629974126815796,
      "num_tokens": 751181685.0,
      "step": 4713
    },
    {
      "epoch": 2.3977619532044763,
      "grad_norm": 0.8834879994392395,
      "learning_rate": 1e-05,
      "loss": 0.4368,
      "mean_token_accuracy": 0.85898756980896,
      "num_tokens": 751341611.0,
      "step": 4714
    },
    {
      "epoch": 2.3982706002034586,
      "grad_norm": 0.865632176399231,
      "learning_rate": 1e-05,
      "loss": 0.4394,
      "mean_token_accuracy": 0.8603726029396057,
      "num_tokens": 751497507.0,
      "step": 4715
    },
    {
      "epoch": 2.3987792472024414,
      "grad_norm": 0.9302217364311218,
      "learning_rate": 1e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8591216206550598,
      "num_tokens": 751658357.0,
      "step": 4716
    },
    {
      "epoch": 2.399287894201424,
      "grad_norm": 0.8736921548843384,
      "learning_rate": 1e-05,
      "loss": 0.4151,
      "mean_token_accuracy": 0.8649030923843384,
      "num_tokens": 751819220.0,
      "step": 4717
    },
    {
      "epoch": 2.399796541200407,
      "grad_norm": 0.8840773105621338,
      "learning_rate": 1e-05,
      "loss": 0.4596,
      "mean_token_accuracy": 0.8527463674545288,
      "num_tokens": 751985925.0,
      "step": 4718
    },
    {
      "epoch": 2.4003051881993898,
      "grad_norm": 1.068041443824768,
      "learning_rate": 1e-05,
      "loss": 0.4467,
      "mean_token_accuracy": 0.8539555072784424,
      "num_tokens": 752152586.0,
      "step": 4719
    },
    {
      "epoch": 2.400813835198372,
      "grad_norm": 0.9924994111061096,
      "learning_rate": 1e-05,
      "loss": 0.4491,
      "mean_token_accuracy": 0.8539040684700012,
      "num_tokens": 752297184.0,
      "step": 4720
    },
    {
      "epoch": 2.401322482197355,
      "grad_norm": 0.9711671471595764,
      "learning_rate": 1e-05,
      "loss": 0.4473,
      "mean_token_accuracy": 0.8555803298950195,
      "num_tokens": 752461486.0,
      "step": 4721
    },
    {
      "epoch": 2.4018311291963377,
      "grad_norm": 0.9455640912055969,
      "learning_rate": 1e-05,
      "loss": 0.4323,
      "mean_token_accuracy": 0.8586920499801636,
      "num_tokens": 752623708.0,
      "step": 4722
    },
    {
      "epoch": 2.4023397761953205,
      "grad_norm": 0.8912090063095093,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8601329922676086,
      "num_tokens": 752783092.0,
      "step": 4723
    },
    {
      "epoch": 2.4028484231943033,
      "grad_norm": 0.933478593826294,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8575913906097412,
      "num_tokens": 752932908.0,
      "step": 4724
    },
    {
      "epoch": 2.403357070193286,
      "grad_norm": 0.9062263369560242,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.858018696308136,
      "num_tokens": 753098396.0,
      "step": 4725
    },
    {
      "epoch": 2.4038657171922684,
      "grad_norm": 0.9479162096977234,
      "learning_rate": 1e-05,
      "loss": 0.4456,
      "mean_token_accuracy": 0.8552176356315613,
      "num_tokens": 753256285.0,
      "step": 4726
    },
    {
      "epoch": 2.404374364191251,
      "grad_norm": 0.9442487955093384,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8518790006637573,
      "num_tokens": 753412905.0,
      "step": 4727
    },
    {
      "epoch": 2.404883011190234,
      "grad_norm": 0.9494585990905762,
      "learning_rate": 1e-05,
      "loss": 0.3956,
      "mean_token_accuracy": 0.8692808151245117,
      "num_tokens": 753554563.0,
      "step": 4728
    },
    {
      "epoch": 2.4053916581892167,
      "grad_norm": 0.9336660504341125,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "mean_token_accuracy": 0.8570419549942017,
      "num_tokens": 753724603.0,
      "step": 4729
    },
    {
      "epoch": 2.4059003051881995,
      "grad_norm": 0.8968808054924011,
      "learning_rate": 1e-05,
      "loss": 0.4184,
      "mean_token_accuracy": 0.8642021417617798,
      "num_tokens": 753883153.0,
      "step": 4730
    },
    {
      "epoch": 2.4064089521871823,
      "grad_norm": 0.9284188151359558,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8559017777442932,
      "num_tokens": 754043284.0,
      "step": 4731
    },
    {
      "epoch": 2.4069175991861647,
      "grad_norm": 0.8959974050521851,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8548084497451782,
      "num_tokens": 754197393.0,
      "step": 4732
    },
    {
      "epoch": 2.4074262461851474,
      "grad_norm": 0.9350323677062988,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8600445985794067,
      "num_tokens": 754351162.0,
      "step": 4733
    },
    {
      "epoch": 2.4079348931841302,
      "grad_norm": 0.8921173214912415,
      "learning_rate": 1e-05,
      "loss": 0.4671,
      "mean_token_accuracy": 0.8496970534324646,
      "num_tokens": 754512562.0,
      "step": 4734
    },
    {
      "epoch": 2.408443540183113,
      "grad_norm": 0.8848469257354736,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8504832983016968,
      "num_tokens": 754672287.0,
      "step": 4735
    },
    {
      "epoch": 2.408952187182096,
      "grad_norm": 0.8982890844345093,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8612314462661743,
      "num_tokens": 754823935.0,
      "step": 4736
    },
    {
      "epoch": 2.409460834181078,
      "grad_norm": 0.9384156465530396,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8550631999969482,
      "num_tokens": 754983096.0,
      "step": 4737
    },
    {
      "epoch": 2.409969481180061,
      "grad_norm": 0.8527581095695496,
      "learning_rate": 1e-05,
      "loss": 0.4282,
      "mean_token_accuracy": 0.8617085218429565,
      "num_tokens": 755155364.0,
      "step": 4738
    },
    {
      "epoch": 2.4104781281790437,
      "grad_norm": 0.8351427316665649,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "mean_token_accuracy": 0.8600209951400757,
      "num_tokens": 755318398.0,
      "step": 4739
    },
    {
      "epoch": 2.4109867751780265,
      "grad_norm": 0.9485787153244019,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8549201488494873,
      "num_tokens": 755467594.0,
      "step": 4740
    },
    {
      "epoch": 2.4114954221770093,
      "grad_norm": 0.8743390440940857,
      "learning_rate": 1e-05,
      "loss": 0.417,
      "mean_token_accuracy": 0.8640428185462952,
      "num_tokens": 755619810.0,
      "step": 4741
    },
    {
      "epoch": 2.4120040691759916,
      "grad_norm": 0.8932185173034668,
      "learning_rate": 1e-05,
      "loss": 0.4158,
      "mean_token_accuracy": 0.8637555241584778,
      "num_tokens": 755776980.0,
      "step": 4742
    },
    {
      "epoch": 2.4125127161749744,
      "grad_norm": 0.8570172786712646,
      "learning_rate": 1e-05,
      "loss": 0.4125,
      "mean_token_accuracy": 0.8660873770713806,
      "num_tokens": 755934665.0,
      "step": 4743
    },
    {
      "epoch": 2.413021363173957,
      "grad_norm": 0.8673945069313049,
      "learning_rate": 1e-05,
      "loss": 0.4447,
      "mean_token_accuracy": 0.855694055557251,
      "num_tokens": 756086651.0,
      "step": 4744
    },
    {
      "epoch": 2.41353001017294,
      "grad_norm": 0.9503878355026245,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.852497935295105,
      "num_tokens": 756237065.0,
      "step": 4745
    },
    {
      "epoch": 2.414038657171923,
      "grad_norm": 0.856009840965271,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8569983839988708,
      "num_tokens": 756399704.0,
      "step": 4746
    },
    {
      "epoch": 2.4145473041709056,
      "grad_norm": 0.8498873710632324,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "mean_token_accuracy": 0.8537609577178955,
      "num_tokens": 756552790.0,
      "step": 4747
    },
    {
      "epoch": 2.415055951169888,
      "grad_norm": 0.8596876859664917,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.8616723418235779,
      "num_tokens": 756711634.0,
      "step": 4748
    },
    {
      "epoch": 2.4155645981688707,
      "grad_norm": 0.8566173315048218,
      "learning_rate": 1e-05,
      "loss": 0.4145,
      "mean_token_accuracy": 0.8648356199264526,
      "num_tokens": 756873060.0,
      "step": 4749
    },
    {
      "epoch": 2.4160732451678535,
      "grad_norm": 0.8351802825927734,
      "learning_rate": 1e-05,
      "loss": 0.4289,
      "mean_token_accuracy": 0.861006498336792,
      "num_tokens": 757032898.0,
      "step": 4750
    },
    {
      "epoch": 2.4165818921668363,
      "grad_norm": 0.849201500415802,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8551095128059387,
      "num_tokens": 757194206.0,
      "step": 4751
    },
    {
      "epoch": 2.417090539165819,
      "grad_norm": 0.8686506152153015,
      "learning_rate": 1e-05,
      "loss": 0.4171,
      "mean_token_accuracy": 0.8662176132202148,
      "num_tokens": 757357972.0,
      "step": 4752
    },
    {
      "epoch": 2.417599186164802,
      "grad_norm": 0.8447994589805603,
      "learning_rate": 1e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.8616929054260254,
      "num_tokens": 757525531.0,
      "step": 4753
    },
    {
      "epoch": 2.418107833163784,
      "grad_norm": 0.916158139705658,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8554221987724304,
      "num_tokens": 757677296.0,
      "step": 4754
    },
    {
      "epoch": 2.418616480162767,
      "grad_norm": 0.8697807192802429,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "mean_token_accuracy": 0.8546140789985657,
      "num_tokens": 757836573.0,
      "step": 4755
    },
    {
      "epoch": 2.4191251271617498,
      "grad_norm": 0.8870343565940857,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8591193556785583,
      "num_tokens": 757995955.0,
      "step": 4756
    },
    {
      "epoch": 2.4196337741607326,
      "grad_norm": 0.9044178128242493,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8570597171783447,
      "num_tokens": 758148208.0,
      "step": 4757
    },
    {
      "epoch": 2.4201424211597153,
      "grad_norm": 1.0102097988128662,
      "learning_rate": 1e-05,
      "loss": 0.4547,
      "mean_token_accuracy": 0.8534799218177795,
      "num_tokens": 758312347.0,
      "step": 4758
    },
    {
      "epoch": 2.4206510681586977,
      "grad_norm": 0.8721146583557129,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8564634323120117,
      "num_tokens": 758474081.0,
      "step": 4759
    },
    {
      "epoch": 2.4211597151576805,
      "grad_norm": 0.8786713480949402,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "mean_token_accuracy": 0.8591622114181519,
      "num_tokens": 758639550.0,
      "step": 4760
    },
    {
      "epoch": 2.4216683621566633,
      "grad_norm": 0.8972764611244202,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8601624965667725,
      "num_tokens": 758802055.0,
      "step": 4761
    },
    {
      "epoch": 2.422177009155646,
      "grad_norm": 0.8989319205284119,
      "learning_rate": 1e-05,
      "loss": 0.4546,
      "mean_token_accuracy": 0.8546927571296692,
      "num_tokens": 758958913.0,
      "step": 4762
    },
    {
      "epoch": 2.422685656154629,
      "grad_norm": 0.9417235851287842,
      "learning_rate": 1e-05,
      "loss": 0.4857,
      "mean_token_accuracy": 0.8451066613197327,
      "num_tokens": 759120435.0,
      "step": 4763
    },
    {
      "epoch": 2.423194303153611,
      "grad_norm": 0.9240588545799255,
      "learning_rate": 1e-05,
      "loss": 0.412,
      "mean_token_accuracy": 0.8654221296310425,
      "num_tokens": 759270507.0,
      "step": 4764
    },
    {
      "epoch": 2.423702950152594,
      "grad_norm": 0.8854421377182007,
      "learning_rate": 1e-05,
      "loss": 0.4117,
      "mean_token_accuracy": 0.8663169145584106,
      "num_tokens": 759432853.0,
      "step": 4765
    },
    {
      "epoch": 2.4242115971515767,
      "grad_norm": 0.928834080696106,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8653715252876282,
      "num_tokens": 759579946.0,
      "step": 4766
    },
    {
      "epoch": 2.4247202441505595,
      "grad_norm": 0.8773136138916016,
      "learning_rate": 1e-05,
      "loss": 0.4433,
      "mean_token_accuracy": 0.8564538955688477,
      "num_tokens": 759749341.0,
      "step": 4767
    },
    {
      "epoch": 2.4252288911495423,
      "grad_norm": 0.8476248979568481,
      "learning_rate": 1e-05,
      "loss": 0.4262,
      "mean_token_accuracy": 0.8624737858772278,
      "num_tokens": 759919471.0,
      "step": 4768
    },
    {
      "epoch": 2.425737538148525,
      "grad_norm": 0.899351954460144,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8652138710021973,
      "num_tokens": 760063245.0,
      "step": 4769
    },
    {
      "epoch": 2.4262461851475075,
      "grad_norm": 0.8362252116203308,
      "learning_rate": 1e-05,
      "loss": 0.4098,
      "mean_token_accuracy": 0.8650784492492676,
      "num_tokens": 760224828.0,
      "step": 4770
    },
    {
      "epoch": 2.4267548321464902,
      "grad_norm": 0.9169917106628418,
      "learning_rate": 1e-05,
      "loss": 0.4341,
      "mean_token_accuracy": 0.8595162630081177,
      "num_tokens": 760378809.0,
      "step": 4771
    },
    {
      "epoch": 2.427263479145473,
      "grad_norm": 0.8968492746353149,
      "learning_rate": 1e-05,
      "loss": 0.4195,
      "mean_token_accuracy": 0.8629690408706665,
      "num_tokens": 760536153.0,
      "step": 4772
    },
    {
      "epoch": 2.427772126144456,
      "grad_norm": 0.8856859803199768,
      "learning_rate": 1e-05,
      "loss": 0.4623,
      "mean_token_accuracy": 0.8506142497062683,
      "num_tokens": 760694351.0,
      "step": 4773
    },
    {
      "epoch": 2.4282807731434386,
      "grad_norm": 0.946141242980957,
      "learning_rate": 1e-05,
      "loss": 0.4255,
      "mean_token_accuracy": 0.8618991374969482,
      "num_tokens": 760838181.0,
      "step": 4774
    },
    {
      "epoch": 2.4287894201424214,
      "grad_norm": 0.9400628805160522,
      "learning_rate": 1e-05,
      "loss": 0.4555,
      "mean_token_accuracy": 0.8526998162269592,
      "num_tokens": 761009751.0,
      "step": 4775
    },
    {
      "epoch": 2.4292980671414037,
      "grad_norm": 0.9153785705566406,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8605570197105408,
      "num_tokens": 761170210.0,
      "step": 4776
    },
    {
      "epoch": 2.4298067141403865,
      "grad_norm": 0.8960493803024292,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8575877547264099,
      "num_tokens": 761330303.0,
      "step": 4777
    },
    {
      "epoch": 2.4303153611393693,
      "grad_norm": 1.01410973072052,
      "learning_rate": 1e-05,
      "loss": 0.4456,
      "mean_token_accuracy": 0.8566583395004272,
      "num_tokens": 761491740.0,
      "step": 4778
    },
    {
      "epoch": 2.430824008138352,
      "grad_norm": 0.9823146462440491,
      "learning_rate": 1e-05,
      "loss": 0.4175,
      "mean_token_accuracy": 0.8643597960472107,
      "num_tokens": 761645233.0,
      "step": 4779
    },
    {
      "epoch": 2.431332655137335,
      "grad_norm": 0.8986417651176453,
      "learning_rate": 1e-05,
      "loss": 0.458,
      "mean_token_accuracy": 0.8521144390106201,
      "num_tokens": 761813341.0,
      "step": 4780
    },
    {
      "epoch": 2.431841302136317,
      "grad_norm": 0.9008432030677795,
      "learning_rate": 1e-05,
      "loss": 0.4265,
      "mean_token_accuracy": 0.8616411685943604,
      "num_tokens": 761972013.0,
      "step": 4781
    },
    {
      "epoch": 2.4323499491353,
      "grad_norm": 0.8586981296539307,
      "learning_rate": 1e-05,
      "loss": 0.4282,
      "mean_token_accuracy": 0.8597036600112915,
      "num_tokens": 762139354.0,
      "step": 4782
    },
    {
      "epoch": 2.432858596134283,
      "grad_norm": 0.9051372408866882,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.858922004699707,
      "num_tokens": 762302578.0,
      "step": 4783
    },
    {
      "epoch": 2.4333672431332656,
      "grad_norm": 0.9037993550300598,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "mean_token_accuracy": 0.8665629625320435,
      "num_tokens": 762460461.0,
      "step": 4784
    },
    {
      "epoch": 2.4338758901322484,
      "grad_norm": 0.8953592777252197,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.8560519218444824,
      "num_tokens": 762626751.0,
      "step": 4785
    },
    {
      "epoch": 2.4343845371312307,
      "grad_norm": 0.8536945581436157,
      "learning_rate": 1e-05,
      "loss": 0.431,
      "mean_token_accuracy": 0.8601256608963013,
      "num_tokens": 762803568.0,
      "step": 4786
    },
    {
      "epoch": 2.4348931841302135,
      "grad_norm": 0.875242292881012,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8583277463912964,
      "num_tokens": 762977204.0,
      "step": 4787
    },
    {
      "epoch": 2.4354018311291963,
      "grad_norm": 1.011696457862854,
      "learning_rate": 1e-05,
      "loss": 0.4299,
      "mean_token_accuracy": 0.8596693873405457,
      "num_tokens": 763121607.0,
      "step": 4788
    },
    {
      "epoch": 2.435910478128179,
      "grad_norm": 0.8985397815704346,
      "learning_rate": 1e-05,
      "loss": 0.454,
      "mean_token_accuracy": 0.8534476161003113,
      "num_tokens": 763287031.0,
      "step": 4789
    },
    {
      "epoch": 2.436419125127162,
      "grad_norm": 1.7359514236450195,
      "learning_rate": 1e-05,
      "loss": 0.4517,
      "mean_token_accuracy": 0.8541314005851746,
      "num_tokens": 763440200.0,
      "step": 4790
    },
    {
      "epoch": 2.4369277721261446,
      "grad_norm": 1.0107104778289795,
      "learning_rate": 1e-05,
      "loss": 0.4237,
      "mean_token_accuracy": 0.863031268119812,
      "num_tokens": 763592000.0,
      "step": 4791
    },
    {
      "epoch": 2.437436419125127,
      "grad_norm": 0.944232702255249,
      "learning_rate": 1e-05,
      "loss": 0.4501,
      "mean_token_accuracy": 0.854409396648407,
      "num_tokens": 763765662.0,
      "step": 4792
    },
    {
      "epoch": 2.4379450661241098,
      "grad_norm": 0.8871685862541199,
      "learning_rate": 1e-05,
      "loss": 0.4333,
      "mean_token_accuracy": 0.8585347533226013,
      "num_tokens": 763939945.0,
      "step": 4793
    },
    {
      "epoch": 2.4384537131230926,
      "grad_norm": 1.0742279291152954,
      "learning_rate": 1e-05,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8643221855163574,
      "num_tokens": 764082350.0,
      "step": 4794
    },
    {
      "epoch": 2.4389623601220753,
      "grad_norm": 0.9682660698890686,
      "learning_rate": 1e-05,
      "loss": 0.4371,
      "mean_token_accuracy": 0.8597748279571533,
      "num_tokens": 764237823.0,
      "step": 4795
    },
    {
      "epoch": 2.439471007121058,
      "grad_norm": 0.9521532654762268,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8522917032241821,
      "num_tokens": 764395519.0,
      "step": 4796
    },
    {
      "epoch": 2.439979654120041,
      "grad_norm": 0.9624428153038025,
      "learning_rate": 1e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8598757386207581,
      "num_tokens": 764558321.0,
      "step": 4797
    },
    {
      "epoch": 2.4404883011190233,
      "grad_norm": 0.9017758965492249,
      "learning_rate": 1e-05,
      "loss": 0.4597,
      "mean_token_accuracy": 0.8528575897216797,
      "num_tokens": 764720615.0,
      "step": 4798
    },
    {
      "epoch": 2.440996948118006,
      "grad_norm": 0.8903970718383789,
      "learning_rate": 1e-05,
      "loss": 0.447,
      "mean_token_accuracy": 0.8557796478271484,
      "num_tokens": 764886477.0,
      "step": 4799
    },
    {
      "epoch": 2.441505595116989,
      "grad_norm": 0.975334644317627,
      "learning_rate": 1e-05,
      "loss": 0.479,
      "mean_token_accuracy": 0.8487585783004761,
      "num_tokens": 765064375.0,
      "step": 4800
    },
    {
      "epoch": 2.4420142421159716,
      "grad_norm": 1.2733622789382935,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8532027006149292,
      "num_tokens": 765232031.0,
      "step": 4801
    },
    {
      "epoch": 2.4425228891149544,
      "grad_norm": 0.996767520904541,
      "learning_rate": 1e-05,
      "loss": 0.4795,
      "mean_token_accuracy": 0.8473283648490906,
      "num_tokens": 765405280.0,
      "step": 4802
    },
    {
      "epoch": 2.4430315361139368,
      "grad_norm": 0.981573224067688,
      "learning_rate": 1e-05,
      "loss": 0.4441,
      "mean_token_accuracy": 0.8556265234947205,
      "num_tokens": 765557818.0,
      "step": 4803
    },
    {
      "epoch": 2.4435401831129195,
      "grad_norm": 0.9518735408782959,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8591273427009583,
      "num_tokens": 765704128.0,
      "step": 4804
    },
    {
      "epoch": 2.4440488301119023,
      "grad_norm": 0.9897494316101074,
      "learning_rate": 1e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8603610396385193,
      "num_tokens": 765860896.0,
      "step": 4805
    },
    {
      "epoch": 2.444557477110885,
      "grad_norm": 1.0344433784484863,
      "learning_rate": 1e-05,
      "loss": 0.4528,
      "mean_token_accuracy": 0.853305995464325,
      "num_tokens": 766021963.0,
      "step": 4806
    },
    {
      "epoch": 2.445066124109868,
      "grad_norm": 0.8656305074691772,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.856003999710083,
      "num_tokens": 766184156.0,
      "step": 4807
    },
    {
      "epoch": 2.4455747711088502,
      "grad_norm": 0.9636114835739136,
      "learning_rate": 1e-05,
      "loss": 0.4492,
      "mean_token_accuracy": 0.855746328830719,
      "num_tokens": 766348235.0,
      "step": 4808
    },
    {
      "epoch": 2.446083418107833,
      "grad_norm": 1.0234853029251099,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.860809862613678,
      "num_tokens": 766508467.0,
      "step": 4809
    },
    {
      "epoch": 2.446592065106816,
      "grad_norm": 0.8884817957878113,
      "learning_rate": 1e-05,
      "loss": 0.4203,
      "mean_token_accuracy": 0.8626583218574524,
      "num_tokens": 766664722.0,
      "step": 4810
    },
    {
      "epoch": 2.4471007121057986,
      "grad_norm": 0.9701591730117798,
      "learning_rate": 1e-05,
      "loss": 0.3896,
      "mean_token_accuracy": 0.8732377290725708,
      "num_tokens": 766821416.0,
      "step": 4811
    },
    {
      "epoch": 2.4476093591047814,
      "grad_norm": 0.9683111310005188,
      "learning_rate": 1e-05,
      "loss": 0.4517,
      "mean_token_accuracy": 0.8539290428161621,
      "num_tokens": 766973521.0,
      "step": 4812
    },
    {
      "epoch": 2.448118006103764,
      "grad_norm": 1.375157117843628,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8657113909721375,
      "num_tokens": 767140971.0,
      "step": 4813
    },
    {
      "epoch": 2.4486266531027465,
      "grad_norm": 0.9833841919898987,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8539447784423828,
      "num_tokens": 767301821.0,
      "step": 4814
    },
    {
      "epoch": 2.4491353001017293,
      "grad_norm": 0.9484559297561646,
      "learning_rate": 1e-05,
      "loss": 0.4266,
      "mean_token_accuracy": 0.8614406585693359,
      "num_tokens": 767467792.0,
      "step": 4815
    },
    {
      "epoch": 2.449643947100712,
      "grad_norm": 0.9295511841773987,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8562976121902466,
      "num_tokens": 767630278.0,
      "step": 4816
    },
    {
      "epoch": 2.450152594099695,
      "grad_norm": 0.8951289653778076,
      "learning_rate": 1e-05,
      "loss": 0.437,
      "mean_token_accuracy": 0.8576017618179321,
      "num_tokens": 767800184.0,
      "step": 4817
    },
    {
      "epoch": 2.4506612410986777,
      "grad_norm": 0.9002968072891235,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8604477643966675,
      "num_tokens": 767961963.0,
      "step": 4818
    },
    {
      "epoch": 2.4511698880976605,
      "grad_norm": 0.9128016829490662,
      "learning_rate": 1e-05,
      "loss": 0.4169,
      "mean_token_accuracy": 0.8654111623764038,
      "num_tokens": 768111977.0,
      "step": 4819
    },
    {
      "epoch": 2.451678535096643,
      "grad_norm": 0.9885953068733215,
      "learning_rate": 1e-05,
      "loss": 0.407,
      "mean_token_accuracy": 0.8666447401046753,
      "num_tokens": 768274598.0,
      "step": 4820
    },
    {
      "epoch": 2.4521871820956256,
      "grad_norm": 0.8686884045600891,
      "learning_rate": 1e-05,
      "loss": 0.4509,
      "mean_token_accuracy": 0.8539066314697266,
      "num_tokens": 768427058.0,
      "step": 4821
    },
    {
      "epoch": 2.4526958290946084,
      "grad_norm": 0.8789154887199402,
      "learning_rate": 1e-05,
      "loss": 0.465,
      "mean_token_accuracy": 0.8505491614341736,
      "num_tokens": 768596525.0,
      "step": 4822
    },
    {
      "epoch": 2.453204476093591,
      "grad_norm": 0.8778037428855896,
      "learning_rate": 1e-05,
      "loss": 0.4192,
      "mean_token_accuracy": 0.8648917078971863,
      "num_tokens": 768763320.0,
      "step": 4823
    },
    {
      "epoch": 2.453713123092574,
      "grad_norm": 0.8817370533943176,
      "learning_rate": 1e-05,
      "loss": 0.4141,
      "mean_token_accuracy": 0.8660046458244324,
      "num_tokens": 768931098.0,
      "step": 4824
    },
    {
      "epoch": 2.4542217700915563,
      "grad_norm": 0.9536473751068115,
      "learning_rate": 1e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8574396371841431,
      "num_tokens": 769083639.0,
      "step": 4825
    },
    {
      "epoch": 2.454730417090539,
      "grad_norm": 0.9757430553436279,
      "learning_rate": 1e-05,
      "loss": 0.4644,
      "mean_token_accuracy": 0.8526658415794373,
      "num_tokens": 769236726.0,
      "step": 4826
    },
    {
      "epoch": 2.455239064089522,
      "grad_norm": 0.9364215135574341,
      "learning_rate": 1e-05,
      "loss": 0.4254,
      "mean_token_accuracy": 0.8613604307174683,
      "num_tokens": 769387035.0,
      "step": 4827
    },
    {
      "epoch": 2.4557477110885046,
      "grad_norm": 0.9096567630767822,
      "learning_rate": 1e-05,
      "loss": 0.4513,
      "mean_token_accuracy": 0.8547468185424805,
      "num_tokens": 769546860.0,
      "step": 4828
    },
    {
      "epoch": 2.4562563580874874,
      "grad_norm": 0.8673074245452881,
      "learning_rate": 1e-05,
      "loss": 0.4281,
      "mean_token_accuracy": 0.8612393140792847,
      "num_tokens": 769696427.0,
      "step": 4829
    },
    {
      "epoch": 2.4567650050864698,
      "grad_norm": 1.0168750286102295,
      "learning_rate": 1e-05,
      "loss": 0.4054,
      "mean_token_accuracy": 0.866659939289093,
      "num_tokens": 769852870.0,
      "step": 4830
    },
    {
      "epoch": 2.4572736520854526,
      "grad_norm": 0.9404606819152832,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.858262836933136,
      "num_tokens": 769992679.0,
      "step": 4831
    },
    {
      "epoch": 2.4577822990844354,
      "grad_norm": 0.9464917778968811,
      "learning_rate": 1e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8508808016777039,
      "num_tokens": 770157889.0,
      "step": 4832
    },
    {
      "epoch": 2.458290946083418,
      "grad_norm": 0.9580070376396179,
      "learning_rate": 1e-05,
      "loss": 0.4251,
      "mean_token_accuracy": 0.8595504760742188,
      "num_tokens": 770327479.0,
      "step": 4833
    },
    {
      "epoch": 2.458799593082401,
      "grad_norm": 0.8783279061317444,
      "learning_rate": 1e-05,
      "loss": 0.4391,
      "mean_token_accuracy": 0.8582003116607666,
      "num_tokens": 770495718.0,
      "step": 4834
    },
    {
      "epoch": 2.4593082400813837,
      "grad_norm": 0.9397633671760559,
      "learning_rate": 1e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8579865097999573,
      "num_tokens": 770649211.0,
      "step": 4835
    },
    {
      "epoch": 2.459816887080366,
      "grad_norm": 0.9210466146469116,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8545470237731934,
      "num_tokens": 770807568.0,
      "step": 4836
    },
    {
      "epoch": 2.460325534079349,
      "grad_norm": 0.9139899611473083,
      "learning_rate": 1e-05,
      "loss": 0.4461,
      "mean_token_accuracy": 0.8555503487586975,
      "num_tokens": 770972725.0,
      "step": 4837
    },
    {
      "epoch": 2.4608341810783316,
      "grad_norm": 0.9260302186012268,
      "learning_rate": 1e-05,
      "loss": 0.3865,
      "mean_token_accuracy": 0.8711071610450745,
      "num_tokens": 771115947.0,
      "step": 4838
    },
    {
      "epoch": 2.4613428280773144,
      "grad_norm": 0.8994468450546265,
      "learning_rate": 1e-05,
      "loss": 0.4165,
      "mean_token_accuracy": 0.8638403415679932,
      "num_tokens": 771281930.0,
      "step": 4839
    },
    {
      "epoch": 2.461851475076297,
      "grad_norm": 1.0170361995697021,
      "learning_rate": 1e-05,
      "loss": 0.4987,
      "mean_token_accuracy": 0.8415489196777344,
      "num_tokens": 771443575.0,
      "step": 4840
    },
    {
      "epoch": 2.46236012207528,
      "grad_norm": 0.9376793503761292,
      "learning_rate": 1e-05,
      "loss": 0.4624,
      "mean_token_accuracy": 0.8517590761184692,
      "num_tokens": 771610548.0,
      "step": 4841
    },
    {
      "epoch": 2.4628687690742623,
      "grad_norm": 0.9882082939147949,
      "learning_rate": 1e-05,
      "loss": 0.4661,
      "mean_token_accuracy": 0.8496295809745789,
      "num_tokens": 771758233.0,
      "step": 4842
    },
    {
      "epoch": 2.463377416073245,
      "grad_norm": 0.9411653876304626,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "mean_token_accuracy": 0.8629002571105957,
      "num_tokens": 771915010.0,
      "step": 4843
    },
    {
      "epoch": 2.463886063072228,
      "grad_norm": 0.9319836497306824,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8531183004379272,
      "num_tokens": 772093055.0,
      "step": 4844
    },
    {
      "epoch": 2.4643947100712107,
      "grad_norm": 0.9602798223495483,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8609582185745239,
      "num_tokens": 772251316.0,
      "step": 4845
    },
    {
      "epoch": 2.4649033570701935,
      "grad_norm": 0.8279889225959778,
      "learning_rate": 1e-05,
      "loss": 0.4164,
      "mean_token_accuracy": 0.865665853023529,
      "num_tokens": 772411962.0,
      "step": 4846
    },
    {
      "epoch": 2.465412004069176,
      "grad_norm": 0.9248610734939575,
      "learning_rate": 1e-05,
      "loss": 0.4441,
      "mean_token_accuracy": 0.8555432558059692,
      "num_tokens": 772579215.0,
      "step": 4847
    },
    {
      "epoch": 2.4659206510681586,
      "grad_norm": 0.9257016181945801,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8550276160240173,
      "num_tokens": 772734084.0,
      "step": 4848
    },
    {
      "epoch": 2.4664292980671414,
      "grad_norm": 0.8871474862098694,
      "learning_rate": 1e-05,
      "loss": 0.4221,
      "mean_token_accuracy": 0.8642134666442871,
      "num_tokens": 772893088.0,
      "step": 4849
    },
    {
      "epoch": 2.466937945066124,
      "grad_norm": 0.8337686657905579,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8641759157180786,
      "num_tokens": 773069483.0,
      "step": 4850
    },
    {
      "epoch": 2.467446592065107,
      "grad_norm": 0.9676657915115356,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8470906615257263,
      "num_tokens": 773230998.0,
      "step": 4851
    },
    {
      "epoch": 2.4679552390640893,
      "grad_norm": 0.8760485649108887,
      "learning_rate": 1e-05,
      "loss": 0.4338,
      "mean_token_accuracy": 0.8574041724205017,
      "num_tokens": 773388030.0,
      "step": 4852
    },
    {
      "epoch": 2.468463886063072,
      "grad_norm": 0.8590541481971741,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8576443791389465,
      "num_tokens": 773551588.0,
      "step": 4853
    },
    {
      "epoch": 2.468972533062055,
      "grad_norm": 0.9119069576263428,
      "learning_rate": 1e-05,
      "loss": 0.4392,
      "mean_token_accuracy": 0.8565086126327515,
      "num_tokens": 773708733.0,
      "step": 4854
    },
    {
      "epoch": 2.4694811800610377,
      "grad_norm": 1.0006945133209229,
      "learning_rate": 1e-05,
      "loss": 0.4404,
      "mean_token_accuracy": 0.8568620681762695,
      "num_tokens": 773861934.0,
      "step": 4855
    },
    {
      "epoch": 2.4699898270600205,
      "grad_norm": 0.8730751872062683,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8648988604545593,
      "num_tokens": 774027412.0,
      "step": 4856
    },
    {
      "epoch": 2.470498474059003,
      "grad_norm": 0.9297431707382202,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8567512035369873,
      "num_tokens": 774187145.0,
      "step": 4857
    },
    {
      "epoch": 2.4710071210579856,
      "grad_norm": 0.8286118507385254,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "mean_token_accuracy": 0.8668421506881714,
      "num_tokens": 774354417.0,
      "step": 4858
    },
    {
      "epoch": 2.4715157680569684,
      "grad_norm": 0.9265725612640381,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.8573986887931824,
      "num_tokens": 774500419.0,
      "step": 4859
    },
    {
      "epoch": 2.472024415055951,
      "grad_norm": 0.9389397501945496,
      "learning_rate": 1e-05,
      "loss": 0.4202,
      "mean_token_accuracy": 0.8619729280471802,
      "num_tokens": 774658382.0,
      "step": 4860
    },
    {
      "epoch": 2.472533062054934,
      "grad_norm": 0.8721439242362976,
      "learning_rate": 1e-05,
      "loss": 0.4589,
      "mean_token_accuracy": 0.8539974689483643,
      "num_tokens": 774814169.0,
      "step": 4861
    },
    {
      "epoch": 2.4730417090539167,
      "grad_norm": 0.9134073257446289,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8584816455841064,
      "num_tokens": 774978306.0,
      "step": 4862
    },
    {
      "epoch": 2.4735503560528995,
      "grad_norm": 0.9060777425765991,
      "learning_rate": 1e-05,
      "loss": 0.4389,
      "mean_token_accuracy": 0.8568652272224426,
      "num_tokens": 775144801.0,
      "step": 4863
    },
    {
      "epoch": 2.474059003051882,
      "grad_norm": 0.9013589024543762,
      "learning_rate": 1e-05,
      "loss": 0.4283,
      "mean_token_accuracy": 0.8597061634063721,
      "num_tokens": 775301131.0,
      "step": 4864
    },
    {
      "epoch": 2.4745676500508647,
      "grad_norm": 0.9063507318496704,
      "learning_rate": 1e-05,
      "loss": 0.4302,
      "mean_token_accuracy": 0.8605985641479492,
      "num_tokens": 775464800.0,
      "step": 4865
    },
    {
      "epoch": 2.4750762970498474,
      "grad_norm": 0.95588618516922,
      "learning_rate": 1e-05,
      "loss": 0.4545,
      "mean_token_accuracy": 0.8535224795341492,
      "num_tokens": 775609674.0,
      "step": 4866
    },
    {
      "epoch": 2.4755849440488302,
      "grad_norm": 0.9267025589942932,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8562873601913452,
      "num_tokens": 775770167.0,
      "step": 4867
    },
    {
      "epoch": 2.476093591047813,
      "grad_norm": 0.9138270616531372,
      "learning_rate": 1e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.8506574630737305,
      "num_tokens": 775937278.0,
      "step": 4868
    },
    {
      "epoch": 2.4766022380467954,
      "grad_norm": 0.8666245341300964,
      "learning_rate": 1e-05,
      "loss": 0.4197,
      "mean_token_accuracy": 0.8650186657905579,
      "num_tokens": 776102454.0,
      "step": 4869
    },
    {
      "epoch": 2.477110885045778,
      "grad_norm": 0.961727499961853,
      "learning_rate": 1e-05,
      "loss": 0.421,
      "mean_token_accuracy": 0.8643406629562378,
      "num_tokens": 776256891.0,
      "step": 4870
    },
    {
      "epoch": 2.477619532044761,
      "grad_norm": 0.876415491104126,
      "learning_rate": 1e-05,
      "loss": 0.4201,
      "mean_token_accuracy": 0.8636149764060974,
      "num_tokens": 776419578.0,
      "step": 4871
    },
    {
      "epoch": 2.4781281790437437,
      "grad_norm": 0.8858059048652649,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8590368032455444,
      "num_tokens": 776580024.0,
      "step": 4872
    },
    {
      "epoch": 2.4786368260427265,
      "grad_norm": 0.8576182126998901,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.856072187423706,
      "num_tokens": 776728328.0,
      "step": 4873
    },
    {
      "epoch": 2.479145473041709,
      "grad_norm": 0.927025318145752,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8582096099853516,
      "num_tokens": 776881190.0,
      "step": 4874
    },
    {
      "epoch": 2.4796541200406916,
      "grad_norm": 0.8863288760185242,
      "learning_rate": 1e-05,
      "loss": 0.4363,
      "mean_token_accuracy": 0.8576536178588867,
      "num_tokens": 777045533.0,
      "step": 4875
    },
    {
      "epoch": 2.4801627670396744,
      "grad_norm": 0.8556225895881653,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "mean_token_accuracy": 0.8564656972885132,
      "num_tokens": 777210571.0,
      "step": 4876
    },
    {
      "epoch": 2.480671414038657,
      "grad_norm": 0.9408009052276611,
      "learning_rate": 1e-05,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8474354147911072,
      "num_tokens": 777387909.0,
      "step": 4877
    },
    {
      "epoch": 2.48118006103764,
      "grad_norm": 0.920619010925293,
      "learning_rate": 1e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.8599849343299866,
      "num_tokens": 777541394.0,
      "step": 4878
    },
    {
      "epoch": 2.4816887080366223,
      "grad_norm": 0.9394387602806091,
      "learning_rate": 1e-05,
      "loss": 0.4473,
      "mean_token_accuracy": 0.8554525375366211,
      "num_tokens": 777690828.0,
      "step": 4879
    },
    {
      "epoch": 2.482197355035605,
      "grad_norm": 0.9379386305809021,
      "learning_rate": 1e-05,
      "loss": 0.4195,
      "mean_token_accuracy": 0.8626469969749451,
      "num_tokens": 777839792.0,
      "step": 4880
    },
    {
      "epoch": 2.482706002034588,
      "grad_norm": 0.9391974210739136,
      "learning_rate": 1e-05,
      "loss": 0.4221,
      "mean_token_accuracy": 0.862157940864563,
      "num_tokens": 777986004.0,
      "step": 4881
    },
    {
      "epoch": 2.4832146490335707,
      "grad_norm": 0.967427670955658,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8584979176521301,
      "num_tokens": 778138868.0,
      "step": 4882
    },
    {
      "epoch": 2.4837232960325535,
      "grad_norm": 0.9181756377220154,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8566079139709473,
      "num_tokens": 778297635.0,
      "step": 4883
    },
    {
      "epoch": 2.4842319430315363,
      "grad_norm": 0.8523744344711304,
      "learning_rate": 1e-05,
      "loss": 0.4136,
      "mean_token_accuracy": 0.8662214279174805,
      "num_tokens": 778464246.0,
      "step": 4884
    },
    {
      "epoch": 2.484740590030519,
      "grad_norm": 0.902087390422821,
      "learning_rate": 1e-05,
      "loss": 0.4701,
      "mean_token_accuracy": 0.8492181301116943,
      "num_tokens": 778634588.0,
      "step": 4885
    },
    {
      "epoch": 2.4852492370295014,
      "grad_norm": 0.9841026067733765,
      "learning_rate": 1e-05,
      "loss": 0.4887,
      "mean_token_accuracy": 0.844761848449707,
      "num_tokens": 778782743.0,
      "step": 4886
    },
    {
      "epoch": 2.485757884028484,
      "grad_norm": 0.9505780339241028,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8542821407318115,
      "num_tokens": 778933303.0,
      "step": 4887
    },
    {
      "epoch": 2.486266531027467,
      "grad_norm": 0.8608629703521729,
      "learning_rate": 1e-05,
      "loss": 0.3945,
      "mean_token_accuracy": 0.8706498146057129,
      "num_tokens": 779106415.0,
      "step": 4888
    },
    {
      "epoch": 2.4867751780264498,
      "grad_norm": 0.9805450439453125,
      "learning_rate": 1e-05,
      "loss": 0.4436,
      "mean_token_accuracy": 0.8559221029281616,
      "num_tokens": 779272445.0,
      "step": 4889
    },
    {
      "epoch": 2.4872838250254325,
      "grad_norm": 0.938648521900177,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8558464646339417,
      "num_tokens": 779428841.0,
      "step": 4890
    },
    {
      "epoch": 2.487792472024415,
      "grad_norm": 0.8914933800697327,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8562479615211487,
      "num_tokens": 779587772.0,
      "step": 4891
    },
    {
      "epoch": 2.4883011190233977,
      "grad_norm": 0.8571580052375793,
      "learning_rate": 1e-05,
      "loss": 0.4248,
      "mean_token_accuracy": 0.8619630932807922,
      "num_tokens": 779752096.0,
      "step": 4892
    },
    {
      "epoch": 2.4888097660223805,
      "grad_norm": 0.9890848398208618,
      "learning_rate": 1e-05,
      "loss": 0.4556,
      "mean_token_accuracy": 0.8534312844276428,
      "num_tokens": 779902168.0,
      "step": 4893
    },
    {
      "epoch": 2.4893184130213633,
      "grad_norm": 0.8465467095375061,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.862485408782959,
      "num_tokens": 780066182.0,
      "step": 4894
    },
    {
      "epoch": 2.489827060020346,
      "grad_norm": 0.8678783178329468,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8596822023391724,
      "num_tokens": 780230835.0,
      "step": 4895
    },
    {
      "epoch": 2.4903357070193284,
      "grad_norm": 0.9244816899299622,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "mean_token_accuracy": 0.8552453517913818,
      "num_tokens": 780379379.0,
      "step": 4896
    },
    {
      "epoch": 2.490844354018311,
      "grad_norm": 0.8347296714782715,
      "learning_rate": 1e-05,
      "loss": 0.4657,
      "mean_token_accuracy": 0.8511793613433838,
      "num_tokens": 780545850.0,
      "step": 4897
    },
    {
      "epoch": 2.491353001017294,
      "grad_norm": 0.9213253855705261,
      "learning_rate": 1e-05,
      "loss": 0.4194,
      "mean_token_accuracy": 0.8629751205444336,
      "num_tokens": 780709618.0,
      "step": 4898
    },
    {
      "epoch": 2.4918616480162767,
      "grad_norm": 0.9033819437026978,
      "learning_rate": 1e-05,
      "loss": 0.4575,
      "mean_token_accuracy": 0.8512227535247803,
      "num_tokens": 780875436.0,
      "step": 4899
    },
    {
      "epoch": 2.4923702950152595,
      "grad_norm": 0.9290204048156738,
      "learning_rate": 1e-05,
      "loss": 0.4886,
      "mean_token_accuracy": 0.8433731198310852,
      "num_tokens": 781038418.0,
      "step": 4900
    },
    {
      "epoch": 2.492878942014242,
      "grad_norm": 0.9555561542510986,
      "learning_rate": 1e-05,
      "loss": 0.4368,
      "mean_token_accuracy": 0.8578329086303711,
      "num_tokens": 781200813.0,
      "step": 4901
    },
    {
      "epoch": 2.4933875890132247,
      "grad_norm": 0.9693750739097595,
      "learning_rate": 1e-05,
      "loss": 0.4396,
      "mean_token_accuracy": 0.8561158180236816,
      "num_tokens": 781342015.0,
      "step": 4902
    },
    {
      "epoch": 2.4938962360122074,
      "grad_norm": 0.990487813949585,
      "learning_rate": 1e-05,
      "loss": 0.4415,
      "mean_token_accuracy": 0.8567934632301331,
      "num_tokens": 781494626.0,
      "step": 4903
    },
    {
      "epoch": 2.4944048830111902,
      "grad_norm": 0.9190848469734192,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8608559370040894,
      "num_tokens": 781654668.0,
      "step": 4904
    },
    {
      "epoch": 2.494913530010173,
      "grad_norm": 1.000533103942871,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8600908517837524,
      "num_tokens": 781814957.0,
      "step": 4905
    },
    {
      "epoch": 2.495422177009156,
      "grad_norm": 0.9340627789497375,
      "learning_rate": 1e-05,
      "loss": 0.4701,
      "mean_token_accuracy": 0.8493572473526001,
      "num_tokens": 781972092.0,
      "step": 4906
    },
    {
      "epoch": 2.4959308240081386,
      "grad_norm": 1.0018067359924316,
      "learning_rate": 1e-05,
      "loss": 0.4677,
      "mean_token_accuracy": 0.8498542904853821,
      "num_tokens": 782131669.0,
      "step": 4907
    },
    {
      "epoch": 2.496439471007121,
      "grad_norm": 0.9221603870391846,
      "learning_rate": 1e-05,
      "loss": 0.4871,
      "mean_token_accuracy": 0.8443587422370911,
      "num_tokens": 782293763.0,
      "step": 4908
    },
    {
      "epoch": 2.4969481180061037,
      "grad_norm": 0.9541207551956177,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.854601263999939,
      "num_tokens": 782439748.0,
      "step": 4909
    },
    {
      "epoch": 2.4974567650050865,
      "grad_norm": 0.989997923374176,
      "learning_rate": 1e-05,
      "loss": 0.4621,
      "mean_token_accuracy": 0.8509840369224548,
      "num_tokens": 782606788.0,
      "step": 4910
    },
    {
      "epoch": 2.4979654120040693,
      "grad_norm": 0.992000162601471,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8571023344993591,
      "num_tokens": 782753839.0,
      "step": 4911
    },
    {
      "epoch": 2.498474059003052,
      "grad_norm": 0.934873104095459,
      "learning_rate": 1e-05,
      "loss": 0.4297,
      "mean_token_accuracy": 0.8594983816146851,
      "num_tokens": 782909993.0,
      "step": 4912
    },
    {
      "epoch": 2.4989827060020344,
      "grad_norm": 1.005946159362793,
      "learning_rate": 1e-05,
      "loss": 0.4373,
      "mean_token_accuracy": 0.8588556051254272,
      "num_tokens": 783060995.0,
      "step": 4913
    },
    {
      "epoch": 2.499491353001017,
      "grad_norm": 0.9700726866722107,
      "learning_rate": 1e-05,
      "loss": 0.4427,
      "mean_token_accuracy": 0.8575286865234375,
      "num_tokens": 783215610.0,
      "step": 4914
    },
    {
      "epoch": 2.5,
      "grad_norm": 1.0107392072677612,
      "learning_rate": 1e-05,
      "loss": 0.4256,
      "mean_token_accuracy": 0.861768364906311,
      "num_tokens": 783383780.0,
      "step": 4915
    },
    {
      "epoch": 2.500508646998983,
      "grad_norm": 0.9195471405982971,
      "learning_rate": 1e-05,
      "loss": 0.4475,
      "mean_token_accuracy": 0.8552471399307251,
      "num_tokens": 783555679.0,
      "step": 4916
    },
    {
      "epoch": 2.5010172939979656,
      "grad_norm": 0.9699219465255737,
      "learning_rate": 1e-05,
      "loss": 0.445,
      "mean_token_accuracy": 0.8576300740242004,
      "num_tokens": 783719996.0,
      "step": 4917
    },
    {
      "epoch": 2.501525940996948,
      "grad_norm": 1.0189080238342285,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8536948561668396,
      "num_tokens": 783880886.0,
      "step": 4918
    },
    {
      "epoch": 2.5020345879959307,
      "grad_norm": 0.8958556652069092,
      "learning_rate": 1e-05,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8617076873779297,
      "num_tokens": 784052705.0,
      "step": 4919
    },
    {
      "epoch": 2.5025432349949135,
      "grad_norm": 1.017082691192627,
      "learning_rate": 1e-05,
      "loss": 0.4234,
      "mean_token_accuracy": 0.8614131212234497,
      "num_tokens": 784198844.0,
      "step": 4920
    },
    {
      "epoch": 2.5030518819938963,
      "grad_norm": 1.014042854309082,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8566722273826599,
      "num_tokens": 784355610.0,
      "step": 4921
    },
    {
      "epoch": 2.503560528992879,
      "grad_norm": 0.8815516829490662,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8559888005256653,
      "num_tokens": 784524245.0,
      "step": 4922
    },
    {
      "epoch": 2.5040691759918614,
      "grad_norm": 0.9433804750442505,
      "learning_rate": 1e-05,
      "loss": 0.4227,
      "mean_token_accuracy": 0.8626022338867188,
      "num_tokens": 784687981.0,
      "step": 4923
    },
    {
      "epoch": 2.504577822990844,
      "grad_norm": 1.0432592630386353,
      "learning_rate": 1e-05,
      "loss": 0.4577,
      "mean_token_accuracy": 0.8526281118392944,
      "num_tokens": 784850952.0,
      "step": 4924
    },
    {
      "epoch": 2.505086469989827,
      "grad_norm": 1.0025300979614258,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.858517587184906,
      "num_tokens": 785007092.0,
      "step": 4925
    },
    {
      "epoch": 2.5055951169888098,
      "grad_norm": 0.9529726505279541,
      "learning_rate": 1e-05,
      "loss": 0.4492,
      "mean_token_accuracy": 0.8542094826698303,
      "num_tokens": 785151045.0,
      "step": 4926
    },
    {
      "epoch": 2.5061037639877926,
      "grad_norm": 0.9628111124038696,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.85976243019104,
      "num_tokens": 785308226.0,
      "step": 4927
    },
    {
      "epoch": 2.5066124109867753,
      "grad_norm": 0.9270199537277222,
      "learning_rate": 1e-05,
      "loss": 0.4462,
      "mean_token_accuracy": 0.8552361130714417,
      "num_tokens": 785470010.0,
      "step": 4928
    },
    {
      "epoch": 2.507121057985758,
      "grad_norm": 0.9673618078231812,
      "learning_rate": 1e-05,
      "loss": 0.4523,
      "mean_token_accuracy": 0.8539666533470154,
      "num_tokens": 785644727.0,
      "step": 4929
    },
    {
      "epoch": 2.5076297049847405,
      "grad_norm": 1.0008258819580078,
      "learning_rate": 1e-05,
      "loss": 0.4666,
      "mean_token_accuracy": 0.8490034341812134,
      "num_tokens": 785801195.0,
      "step": 4930
    },
    {
      "epoch": 2.5081383519837233,
      "grad_norm": 0.925003707408905,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8556567430496216,
      "num_tokens": 785966207.0,
      "step": 4931
    },
    {
      "epoch": 2.508646998982706,
      "grad_norm": 1.0037144422531128,
      "learning_rate": 1e-05,
      "loss": 0.458,
      "mean_token_accuracy": 0.8509320616722107,
      "num_tokens": 786115203.0,
      "step": 4932
    },
    {
      "epoch": 2.509155645981689,
      "grad_norm": 1.1053309440612793,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8615651726722717,
      "num_tokens": 786264456.0,
      "step": 4933
    },
    {
      "epoch": 2.5096642929806716,
      "grad_norm": 0.8932700157165527,
      "learning_rate": 1e-05,
      "loss": 0.4242,
      "mean_token_accuracy": 0.8625380992889404,
      "num_tokens": 786424319.0,
      "step": 4934
    },
    {
      "epoch": 2.510172939979654,
      "grad_norm": 0.9791507720947266,
      "learning_rate": 1e-05,
      "loss": 0.4764,
      "mean_token_accuracy": 0.8476613759994507,
      "num_tokens": 786578284.0,
      "step": 4935
    },
    {
      "epoch": 2.5106815869786367,
      "grad_norm": 1.005222201347351,
      "learning_rate": 1e-05,
      "loss": 0.427,
      "mean_token_accuracy": 0.8623913526535034,
      "num_tokens": 786732176.0,
      "step": 4936
    },
    {
      "epoch": 2.5111902339776195,
      "grad_norm": 0.8984588980674744,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.8568511009216309,
      "num_tokens": 786889777.0,
      "step": 4937
    },
    {
      "epoch": 2.5116988809766023,
      "grad_norm": 0.8648860454559326,
      "learning_rate": 1e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8603200912475586,
      "num_tokens": 787060121.0,
      "step": 4938
    },
    {
      "epoch": 2.512207527975585,
      "grad_norm": 0.9340547323226929,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8607314229011536,
      "num_tokens": 787224425.0,
      "step": 4939
    },
    {
      "epoch": 2.5127161749745675,
      "grad_norm": 0.8825793266296387,
      "learning_rate": 1e-05,
      "loss": 0.4869,
      "mean_token_accuracy": 0.8437527418136597,
      "num_tokens": 787389440.0,
      "step": 4940
    },
    {
      "epoch": 2.5132248219735502,
      "grad_norm": 0.9117286801338196,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8571785092353821,
      "num_tokens": 787550693.0,
      "step": 4941
    },
    {
      "epoch": 2.513733468972533,
      "grad_norm": 0.9391953945159912,
      "learning_rate": 1e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.8555510640144348,
      "num_tokens": 787711904.0,
      "step": 4942
    },
    {
      "epoch": 2.514242115971516,
      "grad_norm": 0.9750176668167114,
      "learning_rate": 1e-05,
      "loss": 0.4197,
      "mean_token_accuracy": 0.8627287149429321,
      "num_tokens": 787876921.0,
      "step": 4943
    },
    {
      "epoch": 2.5147507629704986,
      "grad_norm": 0.8940181136131287,
      "learning_rate": 1e-05,
      "loss": 0.448,
      "mean_token_accuracy": 0.8568694591522217,
      "num_tokens": 788040834.0,
      "step": 4944
    },
    {
      "epoch": 2.515259409969481,
      "grad_norm": 0.9667310118675232,
      "learning_rate": 1e-05,
      "loss": 0.4598,
      "mean_token_accuracy": 0.8532164096832275,
      "num_tokens": 788203339.0,
      "step": 4945
    },
    {
      "epoch": 2.5157680569684637,
      "grad_norm": 0.9718950986862183,
      "learning_rate": 1e-05,
      "loss": 0.4442,
      "mean_token_accuracy": 0.8562138080596924,
      "num_tokens": 788365031.0,
      "step": 4946
    },
    {
      "epoch": 2.5162767039674465,
      "grad_norm": 0.8439437747001648,
      "learning_rate": 1e-05,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8619179725646973,
      "num_tokens": 788536411.0,
      "step": 4947
    },
    {
      "epoch": 2.5167853509664293,
      "grad_norm": 0.9090779423713684,
      "learning_rate": 1e-05,
      "loss": 0.4205,
      "mean_token_accuracy": 0.8635702133178711,
      "num_tokens": 788684184.0,
      "step": 4948
    },
    {
      "epoch": 2.517293997965412,
      "grad_norm": 0.9026861786842346,
      "learning_rate": 1e-05,
      "loss": 0.4558,
      "mean_token_accuracy": 0.8554911017417908,
      "num_tokens": 788855318.0,
      "step": 4949
    },
    {
      "epoch": 2.517802644964395,
      "grad_norm": 0.8946784138679504,
      "learning_rate": 1e-05,
      "loss": 0.408,
      "mean_token_accuracy": 0.8685429692268372,
      "num_tokens": 789007564.0,
      "step": 4950
    },
    {
      "epoch": 2.5183112919633777,
      "grad_norm": 0.8506482243537903,
      "learning_rate": 1e-05,
      "loss": 0.4509,
      "mean_token_accuracy": 0.8538142442703247,
      "num_tokens": 789180703.0,
      "step": 4951
    },
    {
      "epoch": 2.51881993896236,
      "grad_norm": 0.9500423669815063,
      "learning_rate": 1e-05,
      "loss": 0.4557,
      "mean_token_accuracy": 0.8525198698043823,
      "num_tokens": 789338711.0,
      "step": 4952
    },
    {
      "epoch": 2.519328585961343,
      "grad_norm": 0.9335389733314514,
      "learning_rate": 1e-05,
      "loss": 0.4172,
      "mean_token_accuracy": 0.8638827204704285,
      "num_tokens": 789483931.0,
      "step": 4953
    },
    {
      "epoch": 2.5198372329603256,
      "grad_norm": 0.9226436018943787,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8589118719100952,
      "num_tokens": 789651554.0,
      "step": 4954
    },
    {
      "epoch": 2.5203458799593084,
      "grad_norm": 0.9027571082115173,
      "learning_rate": 1e-05,
      "loss": 0.451,
      "mean_token_accuracy": 0.8539646863937378,
      "num_tokens": 789806734.0,
      "step": 4955
    },
    {
      "epoch": 2.520854526958291,
      "grad_norm": 0.8884085416793823,
      "learning_rate": 1e-05,
      "loss": 0.4523,
      "mean_token_accuracy": 0.8536780476570129,
      "num_tokens": 789980144.0,
      "step": 4956
    },
    {
      "epoch": 2.5213631739572735,
      "grad_norm": 0.9327627420425415,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.8556212186813354,
      "num_tokens": 790137138.0,
      "step": 4957
    },
    {
      "epoch": 2.5218718209562563,
      "grad_norm": 0.858979344367981,
      "learning_rate": 1e-05,
      "loss": 0.396,
      "mean_token_accuracy": 0.8699868321418762,
      "num_tokens": 790289686.0,
      "step": 4958
    },
    {
      "epoch": 2.522380467955239,
      "grad_norm": 0.893985390663147,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8482808470726013,
      "num_tokens": 790455729.0,
      "step": 4959
    },
    {
      "epoch": 2.522889114954222,
      "grad_norm": 0.8696373105049133,
      "learning_rate": 1e-05,
      "loss": 0.4467,
      "mean_token_accuracy": 0.855807900428772,
      "num_tokens": 790617809.0,
      "step": 4960
    },
    {
      "epoch": 2.5233977619532046,
      "grad_norm": 0.9078160524368286,
      "learning_rate": 1e-05,
      "loss": 0.4755,
      "mean_token_accuracy": 0.8464839458465576,
      "num_tokens": 790769406.0,
      "step": 4961
    },
    {
      "epoch": 2.523906408952187,
      "grad_norm": 0.8878223299980164,
      "learning_rate": 1e-05,
      "loss": 0.4702,
      "mean_token_accuracy": 0.8494907021522522,
      "num_tokens": 790931867.0,
      "step": 4962
    },
    {
      "epoch": 2.5244150559511698,
      "grad_norm": 0.9027721881866455,
      "learning_rate": 1e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8584170937538147,
      "num_tokens": 791097268.0,
      "step": 4963
    },
    {
      "epoch": 2.5249237029501526,
      "grad_norm": 0.8924083113670349,
      "learning_rate": 1e-05,
      "loss": 0.455,
      "mean_token_accuracy": 0.8540330529212952,
      "num_tokens": 791257431.0,
      "step": 4964
    },
    {
      "epoch": 2.5254323499491353,
      "grad_norm": 0.906243085861206,
      "learning_rate": 1e-05,
      "loss": 0.4098,
      "mean_token_accuracy": 0.865562915802002,
      "num_tokens": 791407837.0,
      "step": 4965
    },
    {
      "epoch": 2.525940996948118,
      "grad_norm": 0.8967453837394714,
      "learning_rate": 1e-05,
      "loss": 0.4614,
      "mean_token_accuracy": 0.8541016578674316,
      "num_tokens": 791571096.0,
      "step": 4966
    },
    {
      "epoch": 2.5264496439471005,
      "grad_norm": 0.9120975732803345,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.851302444934845,
      "num_tokens": 791733488.0,
      "step": 4967
    },
    {
      "epoch": 2.5269582909460833,
      "grad_norm": 0.896980881690979,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.8537946939468384,
      "num_tokens": 791889891.0,
      "step": 4968
    },
    {
      "epoch": 2.527466937945066,
      "grad_norm": 0.9022660851478577,
      "learning_rate": 1e-05,
      "loss": 0.4172,
      "mean_token_accuracy": 0.8633976578712463,
      "num_tokens": 792056142.0,
      "step": 4969
    },
    {
      "epoch": 2.527975584944049,
      "grad_norm": 0.9533856511116028,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "mean_token_accuracy": 0.8529731631278992,
      "num_tokens": 792214100.0,
      "step": 4970
    },
    {
      "epoch": 2.5284842319430316,
      "grad_norm": 0.8792077302932739,
      "learning_rate": 1e-05,
      "loss": 0.4335,
      "mean_token_accuracy": 0.859175443649292,
      "num_tokens": 792375627.0,
      "step": 4971
    },
    {
      "epoch": 2.528992878942014,
      "grad_norm": 0.872978150844574,
      "learning_rate": 1e-05,
      "loss": 0.413,
      "mean_token_accuracy": 0.8643496632575989,
      "num_tokens": 792530725.0,
      "step": 4972
    },
    {
      "epoch": 2.529501525940997,
      "grad_norm": 0.962568461894989,
      "learning_rate": 1e-05,
      "loss": 0.4567,
      "mean_token_accuracy": 0.852573037147522,
      "num_tokens": 792679522.0,
      "step": 4973
    },
    {
      "epoch": 2.5300101729399795,
      "grad_norm": 0.8625645637512207,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8584206700325012,
      "num_tokens": 792847903.0,
      "step": 4974
    },
    {
      "epoch": 2.5305188199389623,
      "grad_norm": 0.9123122096061707,
      "learning_rate": 1e-05,
      "loss": 0.4319,
      "mean_token_accuracy": 0.8611040115356445,
      "num_tokens": 793001198.0,
      "step": 4975
    },
    {
      "epoch": 2.531027466937945,
      "grad_norm": 0.913103461265564,
      "learning_rate": 1e-05,
      "loss": 0.4193,
      "mean_token_accuracy": 0.8622370362281799,
      "num_tokens": 793151668.0,
      "step": 4976
    },
    {
      "epoch": 2.531536113936928,
      "grad_norm": 0.8801509737968445,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.8534757494926453,
      "num_tokens": 793307932.0,
      "step": 4977
    },
    {
      "epoch": 2.5320447609359107,
      "grad_norm": 0.9614366292953491,
      "learning_rate": 1e-05,
      "loss": 0.44,
      "mean_token_accuracy": 0.8562948703765869,
      "num_tokens": 793460024.0,
      "step": 4978
    },
    {
      "epoch": 2.532553407934893,
      "grad_norm": 0.9331154227256775,
      "learning_rate": 1e-05,
      "loss": 0.4377,
      "mean_token_accuracy": 0.8571763634681702,
      "num_tokens": 793624026.0,
      "step": 4979
    },
    {
      "epoch": 2.533062054933876,
      "grad_norm": 0.8993594646453857,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.864728569984436,
      "num_tokens": 793777015.0,
      "step": 4980
    },
    {
      "epoch": 2.5335707019328586,
      "grad_norm": 1.0761821269989014,
      "learning_rate": 1e-05,
      "loss": 0.4628,
      "mean_token_accuracy": 0.8507185578346252,
      "num_tokens": 793940262.0,
      "step": 4981
    },
    {
      "epoch": 2.5340793489318414,
      "grad_norm": 0.8796427249908447,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8514412045478821,
      "num_tokens": 794106369.0,
      "step": 4982
    },
    {
      "epoch": 2.534587995930824,
      "grad_norm": 0.9506366848945618,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8596986532211304,
      "num_tokens": 794254361.0,
      "step": 4983
    },
    {
      "epoch": 2.5350966429298065,
      "grad_norm": 0.8823890686035156,
      "learning_rate": 1e-05,
      "loss": 0.4358,
      "mean_token_accuracy": 0.8592212200164795,
      "num_tokens": 794419893.0,
      "step": 4984
    },
    {
      "epoch": 2.5356052899287893,
      "grad_norm": 0.9204660058021545,
      "learning_rate": 1e-05,
      "loss": 0.3984,
      "mean_token_accuracy": 0.8693076372146606,
      "num_tokens": 794567582.0,
      "step": 4985
    },
    {
      "epoch": 2.536113936927772,
      "grad_norm": 0.882679283618927,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8561245203018188,
      "num_tokens": 794725898.0,
      "step": 4986
    },
    {
      "epoch": 2.536622583926755,
      "grad_norm": 0.9224008321762085,
      "learning_rate": 1e-05,
      "loss": 0.4674,
      "mean_token_accuracy": 0.8492908477783203,
      "num_tokens": 794880869.0,
      "step": 4987
    },
    {
      "epoch": 2.5371312309257377,
      "grad_norm": 1.008573293685913,
      "learning_rate": 1e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.8573492765426636,
      "num_tokens": 795034685.0,
      "step": 4988
    },
    {
      "epoch": 2.53763987792472,
      "grad_norm": 0.9045929312705994,
      "learning_rate": 1e-05,
      "loss": 0.4291,
      "mean_token_accuracy": 0.8587425947189331,
      "num_tokens": 795199669.0,
      "step": 4989
    },
    {
      "epoch": 2.538148524923703,
      "grad_norm": 1.2988561391830444,
      "learning_rate": 1e-05,
      "loss": 0.44,
      "mean_token_accuracy": 0.8566228151321411,
      "num_tokens": 795359866.0,
      "step": 4990
    },
    {
      "epoch": 2.5386571719226856,
      "grad_norm": 0.8983350396156311,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8551926016807556,
      "num_tokens": 795516315.0,
      "step": 4991
    },
    {
      "epoch": 2.5391658189216684,
      "grad_norm": 0.8776088953018188,
      "learning_rate": 1e-05,
      "loss": 0.4175,
      "mean_token_accuracy": 0.8644740581512451,
      "num_tokens": 795665057.0,
      "step": 4992
    },
    {
      "epoch": 2.539674465920651,
      "grad_norm": 0.8982292413711548,
      "learning_rate": 1e-05,
      "loss": 0.4488,
      "mean_token_accuracy": 0.8547069430351257,
      "num_tokens": 795835444.0,
      "step": 4993
    },
    {
      "epoch": 2.5401831129196335,
      "grad_norm": 0.9244893789291382,
      "learning_rate": 1e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8608143329620361,
      "num_tokens": 795976783.0,
      "step": 4994
    },
    {
      "epoch": 2.5406917599186167,
      "grad_norm": 0.8885383009910583,
      "learning_rate": 1e-05,
      "loss": 0.4281,
      "mean_token_accuracy": 0.8603293895721436,
      "num_tokens": 796143149.0,
      "step": 4995
    },
    {
      "epoch": 2.541200406917599,
      "grad_norm": 1.0368354320526123,
      "learning_rate": 1e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8553367853164673,
      "num_tokens": 796287301.0,
      "step": 4996
    },
    {
      "epoch": 2.541709053916582,
      "grad_norm": 0.8786417245864868,
      "learning_rate": 1e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8627998232841492,
      "num_tokens": 796441792.0,
      "step": 4997
    },
    {
      "epoch": 2.5422177009155646,
      "grad_norm": 1.464304804801941,
      "learning_rate": 1e-05,
      "loss": 0.4416,
      "mean_token_accuracy": 0.857825756072998,
      "num_tokens": 796600291.0,
      "step": 4998
    },
    {
      "epoch": 2.5427263479145474,
      "grad_norm": 0.9433695077896118,
      "learning_rate": 1e-05,
      "loss": 0.4355,
      "mean_token_accuracy": 0.8583287596702576,
      "num_tokens": 796757003.0,
      "step": 4999
    },
    {
      "epoch": 2.5432349949135302,
      "grad_norm": 1.0070462226867676,
      "learning_rate": 1e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8625640869140625,
      "num_tokens": 796904278.0,
      "step": 5000
    },
    {
      "epoch": 2.5437436419125126,
      "grad_norm": 0.9288358092308044,
      "learning_rate": 1e-05,
      "loss": 0.4525,
      "mean_token_accuracy": 0.854245662689209,
      "num_tokens": 797061038.0,
      "step": 5001
    },
    {
      "epoch": 2.5442522889114954,
      "grad_norm": 0.8655028343200684,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.855544924736023,
      "num_tokens": 797229543.0,
      "step": 5002
    },
    {
      "epoch": 2.544760935910478,
      "grad_norm": 1.220302939414978,
      "learning_rate": 1e-05,
      "loss": 0.4801,
      "mean_token_accuracy": 0.8450947999954224,
      "num_tokens": 797390103.0,
      "step": 5003
    },
    {
      "epoch": 2.545269582909461,
      "grad_norm": 0.9486819505691528,
      "learning_rate": 1e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8622397780418396,
      "num_tokens": 797536387.0,
      "step": 5004
    },
    {
      "epoch": 2.5457782299084437,
      "grad_norm": 0.9457309246063232,
      "learning_rate": 1e-05,
      "loss": 0.4464,
      "mean_token_accuracy": 0.855649471282959,
      "num_tokens": 797694216.0,
      "step": 5005
    },
    {
      "epoch": 2.546286876907426,
      "grad_norm": 0.9005159735679626,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "mean_token_accuracy": 0.85480797290802,
      "num_tokens": 797843165.0,
      "step": 5006
    },
    {
      "epoch": 2.546795523906409,
      "grad_norm": 0.8820387721061707,
      "learning_rate": 1e-05,
      "loss": 0.4561,
      "mean_token_accuracy": 0.8531599640846252,
      "num_tokens": 798000808.0,
      "step": 5007
    },
    {
      "epoch": 2.5473041709053916,
      "grad_norm": 0.906570315361023,
      "learning_rate": 1e-05,
      "loss": 0.4202,
      "mean_token_accuracy": 0.8623149394989014,
      "num_tokens": 798149467.0,
      "step": 5008
    },
    {
      "epoch": 2.5478128179043744,
      "grad_norm": 0.8808760643005371,
      "learning_rate": 1e-05,
      "loss": 0.4578,
      "mean_token_accuracy": 0.8505076169967651,
      "num_tokens": 798307311.0,
      "step": 5009
    },
    {
      "epoch": 2.548321464903357,
      "grad_norm": 0.8601338863372803,
      "learning_rate": 1e-05,
      "loss": 0.4245,
      "mean_token_accuracy": 0.8612698316574097,
      "num_tokens": 798476152.0,
      "step": 5010
    },
    {
      "epoch": 2.5488301119023395,
      "grad_norm": 0.8980915546417236,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.8529607653617859,
      "num_tokens": 798634863.0,
      "step": 5011
    },
    {
      "epoch": 2.5493387589013223,
      "grad_norm": 0.8095453977584839,
      "learning_rate": 1e-05,
      "loss": 0.4306,
      "mean_token_accuracy": 0.8594805002212524,
      "num_tokens": 798807326.0,
      "step": 5012
    },
    {
      "epoch": 2.549847405900305,
      "grad_norm": 0.8589807748794556,
      "learning_rate": 1e-05,
      "loss": 0.4532,
      "mean_token_accuracy": 0.8550978899002075,
      "num_tokens": 798971177.0,
      "step": 5013
    },
    {
      "epoch": 2.550356052899288,
      "grad_norm": 0.8826808333396912,
      "learning_rate": 1e-05,
      "loss": 0.4989,
      "mean_token_accuracy": 0.8407763838768005,
      "num_tokens": 799134326.0,
      "step": 5014
    },
    {
      "epoch": 2.5508646998982707,
      "grad_norm": 0.8850137591362,
      "learning_rate": 1e-05,
      "loss": 0.424,
      "mean_token_accuracy": 0.8614013195037842,
      "num_tokens": 799279049.0,
      "step": 5015
    },
    {
      "epoch": 2.551373346897253,
      "grad_norm": 0.8453471064567566,
      "learning_rate": 1e-05,
      "loss": 0.4214,
      "mean_token_accuracy": 0.8633631467819214,
      "num_tokens": 799444712.0,
      "step": 5016
    },
    {
      "epoch": 2.5518819938962363,
      "grad_norm": 0.9310623407363892,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8583777546882629,
      "num_tokens": 799590600.0,
      "step": 5017
    },
    {
      "epoch": 2.5523906408952186,
      "grad_norm": 0.8986615538597107,
      "learning_rate": 1e-05,
      "loss": 0.4415,
      "mean_token_accuracy": 0.85772705078125,
      "num_tokens": 799761918.0,
      "step": 5018
    },
    {
      "epoch": 2.5528992878942014,
      "grad_norm": 0.8750797510147095,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8598487973213196,
      "num_tokens": 799928851.0,
      "step": 5019
    },
    {
      "epoch": 2.553407934893184,
      "grad_norm": 1.001911997795105,
      "learning_rate": 1e-05,
      "loss": 0.4024,
      "mean_token_accuracy": 0.8676479458808899,
      "num_tokens": 800086313.0,
      "step": 5020
    },
    {
      "epoch": 2.553916581892167,
      "grad_norm": 0.9289815425872803,
      "learning_rate": 1e-05,
      "loss": 0.4571,
      "mean_token_accuracy": 0.8532669544219971,
      "num_tokens": 800242737.0,
      "step": 5021
    },
    {
      "epoch": 2.5544252288911498,
      "grad_norm": 0.9957112669944763,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8532627820968628,
      "num_tokens": 800392700.0,
      "step": 5022
    },
    {
      "epoch": 2.554933875890132,
      "grad_norm": 0.883906364440918,
      "learning_rate": 1e-05,
      "loss": 0.4346,
      "mean_token_accuracy": 0.8591322302818298,
      "num_tokens": 800561813.0,
      "step": 5023
    },
    {
      "epoch": 2.555442522889115,
      "grad_norm": 0.9741874933242798,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8615695834159851,
      "num_tokens": 800723148.0,
      "step": 5024
    },
    {
      "epoch": 2.5559511698880977,
      "grad_norm": 0.8632175326347351,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "mean_token_accuracy": 0.8571192026138306,
      "num_tokens": 800880074.0,
      "step": 5025
    },
    {
      "epoch": 2.5564598168870805,
      "grad_norm": 0.8779349327087402,
      "learning_rate": 1e-05,
      "loss": 0.426,
      "mean_token_accuracy": 0.8629305362701416,
      "num_tokens": 801037154.0,
      "step": 5026
    },
    {
      "epoch": 2.5569684638860632,
      "grad_norm": 0.9181589484214783,
      "learning_rate": 1e-05,
      "loss": 0.4525,
      "mean_token_accuracy": 0.8532425761222839,
      "num_tokens": 801183676.0,
      "step": 5027
    },
    {
      "epoch": 2.5574771108850456,
      "grad_norm": 0.8723611235618591,
      "learning_rate": 1e-05,
      "loss": 0.4747,
      "mean_token_accuracy": 0.8468239903450012,
      "num_tokens": 801343393.0,
      "step": 5028
    },
    {
      "epoch": 2.5579857578840284,
      "grad_norm": 0.9260839223861694,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8584328293800354,
      "num_tokens": 801497266.0,
      "step": 5029
    },
    {
      "epoch": 2.558494404883011,
      "grad_norm": 0.960536003112793,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8499679565429688,
      "num_tokens": 801652284.0,
      "step": 5030
    },
    {
      "epoch": 2.559003051881994,
      "grad_norm": 0.9009841084480286,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8557119369506836,
      "num_tokens": 801811029.0,
      "step": 5031
    },
    {
      "epoch": 2.5595116988809767,
      "grad_norm": 0.8594654202461243,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8569321632385254,
      "num_tokens": 801967943.0,
      "step": 5032
    },
    {
      "epoch": 2.560020345879959,
      "grad_norm": 0.8589481115341187,
      "learning_rate": 1e-05,
      "loss": 0.4345,
      "mean_token_accuracy": 0.8608789443969727,
      "num_tokens": 802130591.0,
      "step": 5033
    },
    {
      "epoch": 2.560528992878942,
      "grad_norm": 0.8562963008880615,
      "learning_rate": 1e-05,
      "loss": 0.4188,
      "mean_token_accuracy": 0.8626354932785034,
      "num_tokens": 802277825.0,
      "step": 5034
    },
    {
      "epoch": 2.5610376398779247,
      "grad_norm": 0.816863477230072,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8625769019126892,
      "num_tokens": 802439864.0,
      "step": 5035
    },
    {
      "epoch": 2.5615462868769074,
      "grad_norm": 0.9014476537704468,
      "learning_rate": 1e-05,
      "loss": 0.4465,
      "mean_token_accuracy": 0.8575881719589233,
      "num_tokens": 802590185.0,
      "step": 5036
    },
    {
      "epoch": 2.5620549338758902,
      "grad_norm": 0.8575737476348877,
      "learning_rate": 1e-05,
      "loss": 0.3821,
      "mean_token_accuracy": 0.8743054866790771,
      "num_tokens": 802736978.0,
      "step": 5037
    },
    {
      "epoch": 2.5625635808748726,
      "grad_norm": 0.9233067631721497,
      "learning_rate": 1e-05,
      "loss": 0.4217,
      "mean_token_accuracy": 0.8624237775802612,
      "num_tokens": 802896699.0,
      "step": 5038
    },
    {
      "epoch": 2.563072227873856,
      "grad_norm": 1.0039176940917969,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.851535439491272,
      "num_tokens": 803048086.0,
      "step": 5039
    },
    {
      "epoch": 2.563580874872838,
      "grad_norm": 0.8698328733444214,
      "learning_rate": 1e-05,
      "loss": 0.4108,
      "mean_token_accuracy": 0.86533123254776,
      "num_tokens": 803217224.0,
      "step": 5040
    },
    {
      "epoch": 2.564089521871821,
      "grad_norm": 0.8888559937477112,
      "learning_rate": 1e-05,
      "loss": 0.438,
      "mean_token_accuracy": 0.8598009347915649,
      "num_tokens": 803390085.0,
      "step": 5041
    },
    {
      "epoch": 2.5645981688708037,
      "grad_norm": 0.9239501953125,
      "learning_rate": 1e-05,
      "loss": 0.4662,
      "mean_token_accuracy": 0.8478236198425293,
      "num_tokens": 803552881.0,
      "step": 5042
    },
    {
      "epoch": 2.5651068158697865,
      "grad_norm": 0.8581882119178772,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.8616005182266235,
      "num_tokens": 803715846.0,
      "step": 5043
    },
    {
      "epoch": 2.5656154628687693,
      "grad_norm": 0.881350040435791,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8613227605819702,
      "num_tokens": 803876358.0,
      "step": 5044
    },
    {
      "epoch": 2.5661241098677516,
      "grad_norm": 0.9468534588813782,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8618128299713135,
      "num_tokens": 804030876.0,
      "step": 5045
    },
    {
      "epoch": 2.5666327568667344,
      "grad_norm": 0.9433355927467346,
      "learning_rate": 1e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8575580716133118,
      "num_tokens": 804192458.0,
      "step": 5046
    },
    {
      "epoch": 2.567141403865717,
      "grad_norm": 0.8917705416679382,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.8625609874725342,
      "num_tokens": 804348187.0,
      "step": 5047
    },
    {
      "epoch": 2.5676500508647,
      "grad_norm": 0.8788217306137085,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.8561955690383911,
      "num_tokens": 804511634.0,
      "step": 5048
    },
    {
      "epoch": 2.568158697863683,
      "grad_norm": 0.8972126841545105,
      "learning_rate": 1e-05,
      "loss": 0.4605,
      "mean_token_accuracy": 0.8492066860198975,
      "num_tokens": 804669479.0,
      "step": 5049
    },
    {
      "epoch": 2.568667344862665,
      "grad_norm": 0.9535084962844849,
      "learning_rate": 1e-05,
      "loss": 0.4332,
      "mean_token_accuracy": 0.8596338033676147,
      "num_tokens": 804828982.0,
      "step": 5050
    },
    {
      "epoch": 2.569175991861648,
      "grad_norm": 0.8947502970695496,
      "learning_rate": 1e-05,
      "loss": 0.4369,
      "mean_token_accuracy": 0.8589239120483398,
      "num_tokens": 805001239.0,
      "step": 5051
    },
    {
      "epoch": 2.5696846388606307,
      "grad_norm": 0.8903855085372925,
      "learning_rate": 1e-05,
      "loss": 0.4791,
      "mean_token_accuracy": 0.8452934622764587,
      "num_tokens": 805159595.0,
      "step": 5052
    },
    {
      "epoch": 2.5701932858596135,
      "grad_norm": 1.0177345275878906,
      "learning_rate": 1e-05,
      "loss": 0.4767,
      "mean_token_accuracy": 0.847319483757019,
      "num_tokens": 805313717.0,
      "step": 5053
    },
    {
      "epoch": 2.5707019328585963,
      "grad_norm": 0.9481968879699707,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8640065789222717,
      "num_tokens": 805467391.0,
      "step": 5054
    },
    {
      "epoch": 2.5712105798575786,
      "grad_norm": 0.9168588519096375,
      "learning_rate": 1e-05,
      "loss": 0.4455,
      "mean_token_accuracy": 0.8557322025299072,
      "num_tokens": 805617795.0,
      "step": 5055
    },
    {
      "epoch": 2.5717192268565614,
      "grad_norm": 0.9425328373908997,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.8565764427185059,
      "num_tokens": 805767777.0,
      "step": 5056
    },
    {
      "epoch": 2.572227873855544,
      "grad_norm": 0.8727470636367798,
      "learning_rate": 1e-05,
      "loss": 0.4161,
      "mean_token_accuracy": 0.8638690710067749,
      "num_tokens": 805929061.0,
      "step": 5057
    },
    {
      "epoch": 2.572736520854527,
      "grad_norm": 0.9014266729354858,
      "learning_rate": 1e-05,
      "loss": 0.4523,
      "mean_token_accuracy": 0.8545477390289307,
      "num_tokens": 806083011.0,
      "step": 5058
    },
    {
      "epoch": 2.5732451678535098,
      "grad_norm": 0.8865407109260559,
      "learning_rate": 1e-05,
      "loss": 0.4416,
      "mean_token_accuracy": 0.8571542501449585,
      "num_tokens": 806240633.0,
      "step": 5059
    },
    {
      "epoch": 2.573753814852492,
      "grad_norm": 0.9044598937034607,
      "learning_rate": 1e-05,
      "loss": 0.4142,
      "mean_token_accuracy": 0.8641639351844788,
      "num_tokens": 806389989.0,
      "step": 5060
    },
    {
      "epoch": 2.5742624618514753,
      "grad_norm": 0.8423361778259277,
      "learning_rate": 1e-05,
      "loss": 0.4555,
      "mean_token_accuracy": 0.8527371883392334,
      "num_tokens": 806560090.0,
      "step": 5061
    },
    {
      "epoch": 2.5747711088504577,
      "grad_norm": 0.8408938646316528,
      "learning_rate": 1e-05,
      "loss": 0.4205,
      "mean_token_accuracy": 0.8641003370285034,
      "num_tokens": 806724467.0,
      "step": 5062
    },
    {
      "epoch": 2.5752797558494405,
      "grad_norm": 0.8934341669082642,
      "learning_rate": 1e-05,
      "loss": 0.4255,
      "mean_token_accuracy": 0.8617019653320312,
      "num_tokens": 806885928.0,
      "step": 5063
    },
    {
      "epoch": 2.5757884028484233,
      "grad_norm": 0.873944878578186,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8641436100006104,
      "num_tokens": 807048509.0,
      "step": 5064
    },
    {
      "epoch": 2.576297049847406,
      "grad_norm": 0.9249725937843323,
      "learning_rate": 1e-05,
      "loss": 0.4346,
      "mean_token_accuracy": 0.8582570552825928,
      "num_tokens": 807201053.0,
      "step": 5065
    },
    {
      "epoch": 2.576805696846389,
      "grad_norm": 0.8614180684089661,
      "learning_rate": 1e-05,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8631008863449097,
      "num_tokens": 807355916.0,
      "step": 5066
    },
    {
      "epoch": 2.577314343845371,
      "grad_norm": 0.9144124388694763,
      "learning_rate": 1e-05,
      "loss": 0.4132,
      "mean_token_accuracy": 0.865911066532135,
      "num_tokens": 807507280.0,
      "step": 5067
    },
    {
      "epoch": 2.577822990844354,
      "grad_norm": 0.8963925838470459,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8562051057815552,
      "num_tokens": 807664154.0,
      "step": 5068
    },
    {
      "epoch": 2.5783316378433367,
      "grad_norm": 0.880099356174469,
      "learning_rate": 1e-05,
      "loss": 0.4222,
      "mean_token_accuracy": 0.8610818386077881,
      "num_tokens": 807826692.0,
      "step": 5069
    },
    {
      "epoch": 2.5788402848423195,
      "grad_norm": 0.9867204427719116,
      "learning_rate": 1e-05,
      "loss": 0.3965,
      "mean_token_accuracy": 0.869681715965271,
      "num_tokens": 807982521.0,
      "step": 5070
    },
    {
      "epoch": 2.5793489318413023,
      "grad_norm": 0.8853605389595032,
      "learning_rate": 1e-05,
      "loss": 0.445,
      "mean_token_accuracy": 0.8563171625137329,
      "num_tokens": 808145213.0,
      "step": 5071
    },
    {
      "epoch": 2.5798575788402847,
      "grad_norm": 0.8659780025482178,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.8545401096343994,
      "num_tokens": 808310119.0,
      "step": 5072
    },
    {
      "epoch": 2.5803662258392674,
      "grad_norm": 0.8800836205482483,
      "learning_rate": 1e-05,
      "loss": 0.4239,
      "mean_token_accuracy": 0.8624007701873779,
      "num_tokens": 808470957.0,
      "step": 5073
    },
    {
      "epoch": 2.5808748728382502,
      "grad_norm": 0.9086866974830627,
      "learning_rate": 1e-05,
      "loss": 0.4411,
      "mean_token_accuracy": 0.8568074703216553,
      "num_tokens": 808637880.0,
      "step": 5074
    },
    {
      "epoch": 2.581383519837233,
      "grad_norm": 0.9906461834907532,
      "learning_rate": 1e-05,
      "loss": 0.4379,
      "mean_token_accuracy": 0.8569093942642212,
      "num_tokens": 808804560.0,
      "step": 5075
    },
    {
      "epoch": 2.581892166836216,
      "grad_norm": 0.8614892363548279,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.8585564494132996,
      "num_tokens": 808973018.0,
      "step": 5076
    },
    {
      "epoch": 2.582400813835198,
      "grad_norm": 0.9015856385231018,
      "learning_rate": 1e-05,
      "loss": 0.4478,
      "mean_token_accuracy": 0.8551123142242432,
      "num_tokens": 809136988.0,
      "step": 5077
    },
    {
      "epoch": 2.582909460834181,
      "grad_norm": 0.911135196685791,
      "learning_rate": 1e-05,
      "loss": 0.4221,
      "mean_token_accuracy": 0.8616445064544678,
      "num_tokens": 809288153.0,
      "step": 5078
    },
    {
      "epoch": 2.5834181078331637,
      "grad_norm": 0.9012998342514038,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "mean_token_accuracy": 0.8553784489631653,
      "num_tokens": 809443468.0,
      "step": 5079
    },
    {
      "epoch": 2.5839267548321465,
      "grad_norm": 0.9023162126541138,
      "learning_rate": 1e-05,
      "loss": 0.4107,
      "mean_token_accuracy": 0.8648260235786438,
      "num_tokens": 809594813.0,
      "step": 5080
    },
    {
      "epoch": 2.5844354018311293,
      "grad_norm": 0.8686497211456299,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "mean_token_accuracy": 0.8574556112289429,
      "num_tokens": 809760173.0,
      "step": 5081
    },
    {
      "epoch": 2.5849440488301116,
      "grad_norm": 0.9478240013122559,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8620869517326355,
      "num_tokens": 809914564.0,
      "step": 5082
    },
    {
      "epoch": 2.585452695829095,
      "grad_norm": 0.8770041465759277,
      "learning_rate": 1e-05,
      "loss": 0.435,
      "mean_token_accuracy": 0.8599441051483154,
      "num_tokens": 810071927.0,
      "step": 5083
    },
    {
      "epoch": 2.585961342828077,
      "grad_norm": 0.9543029069900513,
      "learning_rate": 1e-05,
      "loss": 0.4748,
      "mean_token_accuracy": 0.8473060131072998,
      "num_tokens": 810235007.0,
      "step": 5084
    },
    {
      "epoch": 2.58646998982706,
      "grad_norm": 0.9194167852401733,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8596623539924622,
      "num_tokens": 810398502.0,
      "step": 5085
    },
    {
      "epoch": 2.586978636826043,
      "grad_norm": 0.993401825428009,
      "learning_rate": 1e-05,
      "loss": 0.4447,
      "mean_token_accuracy": 0.855347752571106,
      "num_tokens": 810562179.0,
      "step": 5086
    },
    {
      "epoch": 2.5874872838250256,
      "grad_norm": 0.9939655065536499,
      "learning_rate": 1e-05,
      "loss": 0.4143,
      "mean_token_accuracy": 0.8665523529052734,
      "num_tokens": 810722858.0,
      "step": 5087
    },
    {
      "epoch": 2.5879959308240084,
      "grad_norm": 0.8436356782913208,
      "learning_rate": 1e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8540129661560059,
      "num_tokens": 810890208.0,
      "step": 5088
    },
    {
      "epoch": 2.5885045778229907,
      "grad_norm": 0.9794009923934937,
      "learning_rate": 1e-05,
      "loss": 0.4418,
      "mean_token_accuracy": 0.8579432368278503,
      "num_tokens": 811055142.0,
      "step": 5089
    },
    {
      "epoch": 2.5890132248219735,
      "grad_norm": 0.8651193976402283,
      "learning_rate": 1e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8586689233779907,
      "num_tokens": 811223836.0,
      "step": 5090
    },
    {
      "epoch": 2.5895218718209563,
      "grad_norm": 0.8984514474868774,
      "learning_rate": 1e-05,
      "loss": 0.4532,
      "mean_token_accuracy": 0.8534331321716309,
      "num_tokens": 811391290.0,
      "step": 5091
    },
    {
      "epoch": 2.590030518819939,
      "grad_norm": 0.8978536128997803,
      "learning_rate": 1e-05,
      "loss": 0.4433,
      "mean_token_accuracy": 0.8573331832885742,
      "num_tokens": 811552000.0,
      "step": 5092
    },
    {
      "epoch": 2.590539165818922,
      "grad_norm": 0.9069320559501648,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8586399555206299,
      "num_tokens": 811720907.0,
      "step": 5093
    },
    {
      "epoch": 2.591047812817904,
      "grad_norm": 0.9119493961334229,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "mean_token_accuracy": 0.8551663160324097,
      "num_tokens": 811879361.0,
      "step": 5094
    },
    {
      "epoch": 2.591556459816887,
      "grad_norm": 0.8818292617797852,
      "learning_rate": 1e-05,
      "loss": 0.4101,
      "mean_token_accuracy": 0.8652764558792114,
      "num_tokens": 812047878.0,
      "step": 5095
    },
    {
      "epoch": 2.5920651068158698,
      "grad_norm": 0.9392347931861877,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "mean_token_accuracy": 0.8555039763450623,
      "num_tokens": 812201258.0,
      "step": 5096
    },
    {
      "epoch": 2.5925737538148526,
      "grad_norm": 0.8664877414703369,
      "learning_rate": 1e-05,
      "loss": 0.4689,
      "mean_token_accuracy": 0.8496640920639038,
      "num_tokens": 812364086.0,
      "step": 5097
    },
    {
      "epoch": 2.5930824008138353,
      "grad_norm": 0.9408642053604126,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.8558844327926636,
      "num_tokens": 812522643.0,
      "step": 5098
    },
    {
      "epoch": 2.5935910478128177,
      "grad_norm": 0.8321124315261841,
      "learning_rate": 1e-05,
      "loss": 0.4116,
      "mean_token_accuracy": 0.866344690322876,
      "num_tokens": 812697980.0,
      "step": 5099
    },
    {
      "epoch": 2.5940996948118005,
      "grad_norm": 0.8491588830947876,
      "learning_rate": 1e-05,
      "loss": 0.437,
      "mean_token_accuracy": 0.8593347668647766,
      "num_tokens": 812857034.0,
      "step": 5100
    },
    {
      "epoch": 2.5946083418107833,
      "grad_norm": 0.8439227342605591,
      "learning_rate": 1e-05,
      "loss": 0.4388,
      "mean_token_accuracy": 0.8568952083587646,
      "num_tokens": 813014073.0,
      "step": 5101
    },
    {
      "epoch": 2.595116988809766,
      "grad_norm": 0.8874712586402893,
      "learning_rate": 1e-05,
      "loss": 0.4596,
      "mean_token_accuracy": 0.852493405342102,
      "num_tokens": 813171705.0,
      "step": 5102
    },
    {
      "epoch": 2.595625635808749,
      "grad_norm": 0.9219279885292053,
      "learning_rate": 1e-05,
      "loss": 0.4332,
      "mean_token_accuracy": 0.8584223389625549,
      "num_tokens": 813321780.0,
      "step": 5103
    },
    {
      "epoch": 2.596134282807731,
      "grad_norm": 0.9127758741378784,
      "learning_rate": 1e-05,
      "loss": 0.4599,
      "mean_token_accuracy": 0.8504177331924438,
      "num_tokens": 813478790.0,
      "step": 5104
    },
    {
      "epoch": 2.5966429298067144,
      "grad_norm": 0.8439157009124756,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8566184043884277,
      "num_tokens": 813639812.0,
      "step": 5105
    },
    {
      "epoch": 2.5971515768056967,
      "grad_norm": 0.8955456614494324,
      "learning_rate": 1e-05,
      "loss": 0.4372,
      "mean_token_accuracy": 0.8591792583465576,
      "num_tokens": 813795523.0,
      "step": 5106
    },
    {
      "epoch": 2.5976602238046795,
      "grad_norm": 0.9001061320304871,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8603255152702332,
      "num_tokens": 813955082.0,
      "step": 5107
    },
    {
      "epoch": 2.5981688708036623,
      "grad_norm": 0.8776459693908691,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "mean_token_accuracy": 0.8569793701171875,
      "num_tokens": 814113911.0,
      "step": 5108
    },
    {
      "epoch": 2.598677517802645,
      "grad_norm": 0.9070444703102112,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8551826477050781,
      "num_tokens": 814275305.0,
      "step": 5109
    },
    {
      "epoch": 2.599186164801628,
      "grad_norm": 0.8836700916290283,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8597379922866821,
      "num_tokens": 814435363.0,
      "step": 5110
    },
    {
      "epoch": 2.5996948118006102,
      "grad_norm": 0.9383623600006104,
      "learning_rate": 1e-05,
      "loss": 0.4708,
      "mean_token_accuracy": 0.850771963596344,
      "num_tokens": 814606320.0,
      "step": 5111
    },
    {
      "epoch": 2.600203458799593,
      "grad_norm": 0.9148361682891846,
      "learning_rate": 1e-05,
      "loss": 0.4284,
      "mean_token_accuracy": 0.8600580096244812,
      "num_tokens": 814760052.0,
      "step": 5112
    },
    {
      "epoch": 2.600712105798576,
      "grad_norm": 0.8902022242546082,
      "learning_rate": 1e-05,
      "loss": 0.4302,
      "mean_token_accuracy": 0.859188437461853,
      "num_tokens": 814923437.0,
      "step": 5113
    },
    {
      "epoch": 2.6012207527975586,
      "grad_norm": 0.8794373869895935,
      "learning_rate": 1e-05,
      "loss": 0.4351,
      "mean_token_accuracy": 0.8580101728439331,
      "num_tokens": 815081519.0,
      "step": 5114
    },
    {
      "epoch": 2.6017293997965414,
      "grad_norm": 0.8626607060432434,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8555400371551514,
      "num_tokens": 815246123.0,
      "step": 5115
    },
    {
      "epoch": 2.6022380467955237,
      "grad_norm": 0.87879878282547,
      "learning_rate": 1e-05,
      "loss": 0.4479,
      "mean_token_accuracy": 0.8553816080093384,
      "num_tokens": 815404054.0,
      "step": 5116
    },
    {
      "epoch": 2.6027466937945065,
      "grad_norm": 0.8513388633728027,
      "learning_rate": 1e-05,
      "loss": 0.4434,
      "mean_token_accuracy": 0.8568986058235168,
      "num_tokens": 815560199.0,
      "step": 5117
    },
    {
      "epoch": 2.6032553407934893,
      "grad_norm": 0.8843596577644348,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8611048460006714,
      "num_tokens": 815726850.0,
      "step": 5118
    },
    {
      "epoch": 2.603763987792472,
      "grad_norm": 0.8609278202056885,
      "learning_rate": 1e-05,
      "loss": 0.4906,
      "mean_token_accuracy": 0.843414306640625,
      "num_tokens": 815898444.0,
      "step": 5119
    },
    {
      "epoch": 2.604272634791455,
      "grad_norm": 0.9479242563247681,
      "learning_rate": 1e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8611266016960144,
      "num_tokens": 816039389.0,
      "step": 5120
    },
    {
      "epoch": 2.604781281790437,
      "grad_norm": 0.8942499756813049,
      "learning_rate": 1e-05,
      "loss": 0.426,
      "mean_token_accuracy": 0.8614490032196045,
      "num_tokens": 816198957.0,
      "step": 5121
    },
    {
      "epoch": 2.60528992878942,
      "grad_norm": 0.8818740248680115,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8538020849227905,
      "num_tokens": 816369450.0,
      "step": 5122
    },
    {
      "epoch": 2.605798575788403,
      "grad_norm": 0.9133926033973694,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8584133386611938,
      "num_tokens": 816527861.0,
      "step": 5123
    },
    {
      "epoch": 2.6063072227873856,
      "grad_norm": 0.8590560555458069,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8604164123535156,
      "num_tokens": 816691307.0,
      "step": 5124
    },
    {
      "epoch": 2.6068158697863684,
      "grad_norm": 0.9377419352531433,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8640687465667725,
      "num_tokens": 816850090.0,
      "step": 5125
    },
    {
      "epoch": 2.6073245167853507,
      "grad_norm": 0.8784112930297852,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8555065393447876,
      "num_tokens": 817014438.0,
      "step": 5126
    },
    {
      "epoch": 2.607833163784334,
      "grad_norm": 0.9228441119194031,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8573063015937805,
      "num_tokens": 817172384.0,
      "step": 5127
    },
    {
      "epoch": 2.6083418107833163,
      "grad_norm": 0.8088212609291077,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8604148030281067,
      "num_tokens": 817342242.0,
      "step": 5128
    },
    {
      "epoch": 2.608850457782299,
      "grad_norm": 0.9407182931900024,
      "learning_rate": 1e-05,
      "loss": 0.4391,
      "mean_token_accuracy": 0.8581157922744751,
      "num_tokens": 817492020.0,
      "step": 5129
    },
    {
      "epoch": 2.609359104781282,
      "grad_norm": 0.8717429041862488,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8543630242347717,
      "num_tokens": 817643303.0,
      "step": 5130
    },
    {
      "epoch": 2.6098677517802646,
      "grad_norm": 0.8946168422698975,
      "learning_rate": 1e-05,
      "loss": 0.3996,
      "mean_token_accuracy": 0.8689746856689453,
      "num_tokens": 817795006.0,
      "step": 5131
    },
    {
      "epoch": 2.6103763987792474,
      "grad_norm": 0.9147217273712158,
      "learning_rate": 1e-05,
      "loss": 0.4008,
      "mean_token_accuracy": 0.8693822622299194,
      "num_tokens": 817952612.0,
      "step": 5132
    },
    {
      "epoch": 2.6108850457782298,
      "grad_norm": 0.9123526811599731,
      "learning_rate": 1e-05,
      "loss": 0.4731,
      "mean_token_accuracy": 0.8474389314651489,
      "num_tokens": 818110205.0,
      "step": 5133
    },
    {
      "epoch": 2.6113936927772126,
      "grad_norm": 0.9014759659767151,
      "learning_rate": 1e-05,
      "loss": 0.4465,
      "mean_token_accuracy": 0.8557165265083313,
      "num_tokens": 818287143.0,
      "step": 5134
    },
    {
      "epoch": 2.6119023397761953,
      "grad_norm": 0.8831366896629333,
      "learning_rate": 1e-05,
      "loss": 0.4257,
      "mean_token_accuracy": 0.860625684261322,
      "num_tokens": 818445307.0,
      "step": 5135
    },
    {
      "epoch": 2.612410986775178,
      "grad_norm": 0.8970555067062378,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.860558271408081,
      "num_tokens": 818608748.0,
      "step": 5136
    },
    {
      "epoch": 2.612919633774161,
      "grad_norm": 0.9433715343475342,
      "learning_rate": 1e-05,
      "loss": 0.4256,
      "mean_token_accuracy": 0.8625460863113403,
      "num_tokens": 818761341.0,
      "step": 5137
    },
    {
      "epoch": 2.6134282807731433,
      "grad_norm": 0.9488721489906311,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.8575472831726074,
      "num_tokens": 818924573.0,
      "step": 5138
    },
    {
      "epoch": 2.613936927772126,
      "grad_norm": 0.9305161833763123,
      "learning_rate": 1e-05,
      "loss": 0.4191,
      "mean_token_accuracy": 0.8635993003845215,
      "num_tokens": 819084612.0,
      "step": 5139
    },
    {
      "epoch": 2.614445574771109,
      "grad_norm": 0.8633292317390442,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "mean_token_accuracy": 0.8654317855834961,
      "num_tokens": 819253050.0,
      "step": 5140
    },
    {
      "epoch": 2.6149542217700916,
      "grad_norm": 0.969501793384552,
      "learning_rate": 1e-05,
      "loss": 0.4224,
      "mean_token_accuracy": 0.8637001514434814,
      "num_tokens": 819417551.0,
      "step": 5141
    },
    {
      "epoch": 2.6154628687690744,
      "grad_norm": 0.8660587072372437,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.857624888420105,
      "num_tokens": 819585374.0,
      "step": 5142
    },
    {
      "epoch": 2.6159715157680568,
      "grad_norm": 0.9489619731903076,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8545424342155457,
      "num_tokens": 819750312.0,
      "step": 5143
    },
    {
      "epoch": 2.6164801627670395,
      "grad_norm": 0.8869263529777527,
      "learning_rate": 1e-05,
      "loss": 0.427,
      "mean_token_accuracy": 0.8611224293708801,
      "num_tokens": 819902977.0,
      "step": 5144
    },
    {
      "epoch": 2.6169888097660223,
      "grad_norm": 0.9305620789527893,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8522117733955383,
      "num_tokens": 820060784.0,
      "step": 5145
    },
    {
      "epoch": 2.617497456765005,
      "grad_norm": 0.9102436304092407,
      "learning_rate": 1e-05,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8633850812911987,
      "num_tokens": 820219660.0,
      "step": 5146
    },
    {
      "epoch": 2.618006103763988,
      "grad_norm": 0.8894648551940918,
      "learning_rate": 1e-05,
      "loss": 0.4781,
      "mean_token_accuracy": 0.8472837805747986,
      "num_tokens": 820390494.0,
      "step": 5147
    },
    {
      "epoch": 2.6185147507629702,
      "grad_norm": 0.9662260413169861,
      "learning_rate": 1e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.8584712743759155,
      "num_tokens": 820542132.0,
      "step": 5148
    },
    {
      "epoch": 2.6190233977619535,
      "grad_norm": 0.8438816070556641,
      "learning_rate": 1e-05,
      "loss": 0.4238,
      "mean_token_accuracy": 0.8626097440719604,
      "num_tokens": 820707454.0,
      "step": 5149
    },
    {
      "epoch": 2.619532044760936,
      "grad_norm": 0.8659455180168152,
      "learning_rate": 1e-05,
      "loss": 0.4042,
      "mean_token_accuracy": 0.8677513599395752,
      "num_tokens": 820860446.0,
      "step": 5150
    },
    {
      "epoch": 2.6200406917599186,
      "grad_norm": 0.8488020896911621,
      "learning_rate": 1e-05,
      "loss": 0.4481,
      "mean_token_accuracy": 0.8547829985618591,
      "num_tokens": 821018836.0,
      "step": 5151
    },
    {
      "epoch": 2.6205493387589014,
      "grad_norm": 1.1235007047653198,
      "learning_rate": 1e-05,
      "loss": 0.4813,
      "mean_token_accuracy": 0.8466203212738037,
      "num_tokens": 821181422.0,
      "step": 5152
    },
    {
      "epoch": 2.621057985757884,
      "grad_norm": 0.9657119512557983,
      "learning_rate": 1e-05,
      "loss": 0.4169,
      "mean_token_accuracy": 0.8650161027908325,
      "num_tokens": 821327762.0,
      "step": 5153
    },
    {
      "epoch": 2.621566632756867,
      "grad_norm": 0.8740357756614685,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "mean_token_accuracy": 0.8541836738586426,
      "num_tokens": 821489947.0,
      "step": 5154
    },
    {
      "epoch": 2.6220752797558493,
      "grad_norm": 0.8773635029792786,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "mean_token_accuracy": 0.8584322333335876,
      "num_tokens": 821657627.0,
      "step": 5155
    },
    {
      "epoch": 2.622583926754832,
      "grad_norm": 0.9291388392448425,
      "learning_rate": 1e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8563066720962524,
      "num_tokens": 821809310.0,
      "step": 5156
    },
    {
      "epoch": 2.623092573753815,
      "grad_norm": 0.869541347026825,
      "learning_rate": 1e-05,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8639887571334839,
      "num_tokens": 821966193.0,
      "step": 5157
    },
    {
      "epoch": 2.6236012207527977,
      "grad_norm": 0.8881884217262268,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8515429496765137,
      "num_tokens": 822129562.0,
      "step": 5158
    },
    {
      "epoch": 2.6241098677517805,
      "grad_norm": 0.9253519177436829,
      "learning_rate": 1e-05,
      "loss": 0.397,
      "mean_token_accuracy": 0.869998574256897,
      "num_tokens": 822287065.0,
      "step": 5159
    },
    {
      "epoch": 2.624618514750763,
      "grad_norm": 0.8654248118400574,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "mean_token_accuracy": 0.8618375062942505,
      "num_tokens": 822441494.0,
      "step": 5160
    },
    {
      "epoch": 2.6251271617497456,
      "grad_norm": 0.9040758609771729,
      "learning_rate": 1e-05,
      "loss": 0.4033,
      "mean_token_accuracy": 0.8672125339508057,
      "num_tokens": 822603565.0,
      "step": 5161
    },
    {
      "epoch": 2.6256358087487284,
      "grad_norm": 0.9012642502784729,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "mean_token_accuracy": 0.849787175655365,
      "num_tokens": 822761536.0,
      "step": 5162
    },
    {
      "epoch": 2.626144455747711,
      "grad_norm": 0.88527512550354,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8527937531471252,
      "num_tokens": 822927303.0,
      "step": 5163
    },
    {
      "epoch": 2.626653102746694,
      "grad_norm": 0.906010091304779,
      "learning_rate": 1e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8562264442443848,
      "num_tokens": 823092906.0,
      "step": 5164
    },
    {
      "epoch": 2.6271617497456763,
      "grad_norm": 0.8845453262329102,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8577530980110168,
      "num_tokens": 823249279.0,
      "step": 5165
    },
    {
      "epoch": 2.627670396744659,
      "grad_norm": 0.8868412375450134,
      "learning_rate": 1e-05,
      "loss": 0.4302,
      "mean_token_accuracy": 0.860711395740509,
      "num_tokens": 823403147.0,
      "step": 5166
    },
    {
      "epoch": 2.628179043743642,
      "grad_norm": 0.9444741010665894,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "mean_token_accuracy": 0.8536720871925354,
      "num_tokens": 823545931.0,
      "step": 5167
    },
    {
      "epoch": 2.6286876907426246,
      "grad_norm": 0.907468318939209,
      "learning_rate": 1e-05,
      "loss": 0.4335,
      "mean_token_accuracy": 0.859333872795105,
      "num_tokens": 823694330.0,
      "step": 5168
    },
    {
      "epoch": 2.6291963377416074,
      "grad_norm": 0.8744403719902039,
      "learning_rate": 1e-05,
      "loss": 0.414,
      "mean_token_accuracy": 0.863753080368042,
      "num_tokens": 823851802.0,
      "step": 5169
    },
    {
      "epoch": 2.62970498474059,
      "grad_norm": 0.8766217827796936,
      "learning_rate": 1e-05,
      "loss": 0.4386,
      "mean_token_accuracy": 0.8578379154205322,
      "num_tokens": 824009139.0,
      "step": 5170
    },
    {
      "epoch": 2.630213631739573,
      "grad_norm": 0.9645333290100098,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "mean_token_accuracy": 0.8645223379135132,
      "num_tokens": 824173599.0,
      "step": 5171
    },
    {
      "epoch": 2.6307222787385554,
      "grad_norm": 0.8777177333831787,
      "learning_rate": 1e-05,
      "loss": 0.4152,
      "mean_token_accuracy": 0.8644059896469116,
      "num_tokens": 824329297.0,
      "step": 5172
    },
    {
      "epoch": 2.631230925737538,
      "grad_norm": 0.8273940086364746,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8570989966392517,
      "num_tokens": 824511087.0,
      "step": 5173
    },
    {
      "epoch": 2.631739572736521,
      "grad_norm": 0.9432076215744019,
      "learning_rate": 1e-05,
      "loss": 0.4511,
      "mean_token_accuracy": 0.8537647724151611,
      "num_tokens": 824673484.0,
      "step": 5174
    },
    {
      "epoch": 2.6322482197355037,
      "grad_norm": 0.848598837852478,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.8540287017822266,
      "num_tokens": 824843437.0,
      "step": 5175
    },
    {
      "epoch": 2.6327568667344865,
      "grad_norm": 0.963729739189148,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.8569854497909546,
      "num_tokens": 825002632.0,
      "step": 5176
    },
    {
      "epoch": 2.633265513733469,
      "grad_norm": 0.8712653517723083,
      "learning_rate": 1e-05,
      "loss": 0.4404,
      "mean_token_accuracy": 0.8580498695373535,
      "num_tokens": 825163042.0,
      "step": 5177
    },
    {
      "epoch": 2.6337741607324516,
      "grad_norm": 0.8973901867866516,
      "learning_rate": 1e-05,
      "loss": 0.442,
      "mean_token_accuracy": 0.8566519021987915,
      "num_tokens": 825320835.0,
      "step": 5178
    },
    {
      "epoch": 2.6342828077314344,
      "grad_norm": 0.8860601782798767,
      "learning_rate": 1e-05,
      "loss": 0.4357,
      "mean_token_accuracy": 0.8590834736824036,
      "num_tokens": 825488497.0,
      "step": 5179
    },
    {
      "epoch": 2.634791454730417,
      "grad_norm": 0.8676003217697144,
      "learning_rate": 1e-05,
      "loss": 0.4228,
      "mean_token_accuracy": 0.8617715835571289,
      "num_tokens": 825642437.0,
      "step": 5180
    },
    {
      "epoch": 2.6353001017294,
      "grad_norm": 0.9539165496826172,
      "learning_rate": 1e-05,
      "loss": 0.4409,
      "mean_token_accuracy": 0.8583078384399414,
      "num_tokens": 825798090.0,
      "step": 5181
    },
    {
      "epoch": 2.6358087487283823,
      "grad_norm": 0.8959965705871582,
      "learning_rate": 1e-05,
      "loss": 0.4679,
      "mean_token_accuracy": 0.8503422737121582,
      "num_tokens": 825949889.0,
      "step": 5182
    },
    {
      "epoch": 2.636317395727365,
      "grad_norm": 0.9380295872688293,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.859856903553009,
      "num_tokens": 826089977.0,
      "step": 5183
    },
    {
      "epoch": 2.636826042726348,
      "grad_norm": 0.924473762512207,
      "learning_rate": 1e-05,
      "loss": 0.4188,
      "mean_token_accuracy": 0.8627535104751587,
      "num_tokens": 826249106.0,
      "step": 5184
    },
    {
      "epoch": 2.6373346897253307,
      "grad_norm": 0.9031071066856384,
      "learning_rate": 1e-05,
      "loss": 0.4491,
      "mean_token_accuracy": 0.8555418252944946,
      "num_tokens": 826417254.0,
      "step": 5185
    },
    {
      "epoch": 2.6378433367243135,
      "grad_norm": 0.8903203010559082,
      "learning_rate": 1e-05,
      "loss": 0.4263,
      "mean_token_accuracy": 0.8602005839347839,
      "num_tokens": 826566435.0,
      "step": 5186
    },
    {
      "epoch": 2.638351983723296,
      "grad_norm": 0.941855788230896,
      "learning_rate": 1e-05,
      "loss": 0.4503,
      "mean_token_accuracy": 0.8543573617935181,
      "num_tokens": 826722136.0,
      "step": 5187
    },
    {
      "epoch": 2.6388606307222786,
      "grad_norm": 1.0012236833572388,
      "learning_rate": 1e-05,
      "loss": 0.4267,
      "mean_token_accuracy": 0.8614889979362488,
      "num_tokens": 826880946.0,
      "step": 5188
    },
    {
      "epoch": 2.6393692777212614,
      "grad_norm": 0.9274325370788574,
      "learning_rate": 1e-05,
      "loss": 0.4573,
      "mean_token_accuracy": 0.8530974984169006,
      "num_tokens": 827041777.0,
      "step": 5189
    },
    {
      "epoch": 2.639877924720244,
      "grad_norm": 0.8896151185035706,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8657070994377136,
      "num_tokens": 827204054.0,
      "step": 5190
    },
    {
      "epoch": 2.640386571719227,
      "grad_norm": 0.8736908435821533,
      "learning_rate": 1e-05,
      "loss": 0.4105,
      "mean_token_accuracy": 0.8652385473251343,
      "num_tokens": 827369516.0,
      "step": 5191
    },
    {
      "epoch": 2.6408952187182093,
      "grad_norm": 0.8775960206985474,
      "learning_rate": 1e-05,
      "loss": 0.4167,
      "mean_token_accuracy": 0.863031268119812,
      "num_tokens": 827526209.0,
      "step": 5192
    },
    {
      "epoch": 2.6414038657171925,
      "grad_norm": 0.9726622104644775,
      "learning_rate": 1e-05,
      "loss": 0.4312,
      "mean_token_accuracy": 0.8588466048240662,
      "num_tokens": 827678096.0,
      "step": 5193
    },
    {
      "epoch": 2.641912512716175,
      "grad_norm": 0.8615018725395203,
      "learning_rate": 1e-05,
      "loss": 0.4055,
      "mean_token_accuracy": 0.8670532703399658,
      "num_tokens": 827841027.0,
      "step": 5194
    },
    {
      "epoch": 2.6424211597151577,
      "grad_norm": 0.9759061336517334,
      "learning_rate": 1e-05,
      "loss": 0.4488,
      "mean_token_accuracy": 0.8542364835739136,
      "num_tokens": 828000994.0,
      "step": 5195
    },
    {
      "epoch": 2.6429298067141405,
      "grad_norm": 0.9128785729408264,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8505659699440002,
      "num_tokens": 828167926.0,
      "step": 5196
    },
    {
      "epoch": 2.6434384537131232,
      "grad_norm": 0.9594423770904541,
      "learning_rate": 1e-05,
      "loss": 0.409,
      "mean_token_accuracy": 0.8666039705276489,
      "num_tokens": 828333229.0,
      "step": 5197
    },
    {
      "epoch": 2.643947100712106,
      "grad_norm": 1.0733044147491455,
      "learning_rate": 1e-05,
      "loss": 0.4618,
      "mean_token_accuracy": 0.8506227731704712,
      "num_tokens": 828492310.0,
      "step": 5198
    },
    {
      "epoch": 2.6444557477110884,
      "grad_norm": 0.9024893045425415,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8601677417755127,
      "num_tokens": 828659357.0,
      "step": 5199
    },
    {
      "epoch": 2.644964394710071,
      "grad_norm": 1.140589952468872,
      "learning_rate": 1e-05,
      "loss": 0.44,
      "mean_token_accuracy": 0.8560845851898193,
      "num_tokens": 828811219.0,
      "step": 5200
    },
    {
      "epoch": 2.645473041709054,
      "grad_norm": 0.9783995747566223,
      "learning_rate": 1e-05,
      "loss": 0.4132,
      "mean_token_accuracy": 0.8665125370025635,
      "num_tokens": 828964453.0,
      "step": 5201
    },
    {
      "epoch": 2.6459816887080367,
      "grad_norm": 1.0958985090255737,
      "learning_rate": 1e-05,
      "loss": 0.4041,
      "mean_token_accuracy": 0.8671479821205139,
      "num_tokens": 829121022.0,
      "step": 5202
    },
    {
      "epoch": 2.6464903357070195,
      "grad_norm": 1.1051671504974365,
      "learning_rate": 1e-05,
      "loss": 0.4448,
      "mean_token_accuracy": 0.8555877804756165,
      "num_tokens": 829285283.0,
      "step": 5203
    },
    {
      "epoch": 2.646998982706002,
      "grad_norm": 0.8696182370185852,
      "learning_rate": 1e-05,
      "loss": 0.4355,
      "mean_token_accuracy": 0.8576204776763916,
      "num_tokens": 829442232.0,
      "step": 5204
    },
    {
      "epoch": 2.6475076297049847,
      "grad_norm": 1.0274218320846558,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "mean_token_accuracy": 0.8589468598365784,
      "num_tokens": 829597449.0,
      "step": 5205
    },
    {
      "epoch": 2.6480162767039674,
      "grad_norm": 0.9053404331207275,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8558415174484253,
      "num_tokens": 829766930.0,
      "step": 5206
    },
    {
      "epoch": 2.6485249237029502,
      "grad_norm": 0.900059163570404,
      "learning_rate": 1e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8552563190460205,
      "num_tokens": 829930119.0,
      "step": 5207
    },
    {
      "epoch": 2.649033570701933,
      "grad_norm": 1.030701994895935,
      "learning_rate": 1e-05,
      "loss": 0.4379,
      "mean_token_accuracy": 0.8582024574279785,
      "num_tokens": 830084634.0,
      "step": 5208
    },
    {
      "epoch": 2.6495422177009154,
      "grad_norm": 0.8864624500274658,
      "learning_rate": 1e-05,
      "loss": 0.4398,
      "mean_token_accuracy": 0.8589258790016174,
      "num_tokens": 830249760.0,
      "step": 5209
    },
    {
      "epoch": 2.650050864699898,
      "grad_norm": 0.9129674434661865,
      "learning_rate": 1e-05,
      "loss": 0.395,
      "mean_token_accuracy": 0.8705196380615234,
      "num_tokens": 830411922.0,
      "step": 5210
    },
    {
      "epoch": 2.650559511698881,
      "grad_norm": 0.9383035898208618,
      "learning_rate": 1e-05,
      "loss": 0.455,
      "mean_token_accuracy": 0.8523730635643005,
      "num_tokens": 830575618.0,
      "step": 5211
    },
    {
      "epoch": 2.6510681586978637,
      "grad_norm": 1.006765604019165,
      "learning_rate": 1e-05,
      "loss": 0.4102,
      "mean_token_accuracy": 0.866306722164154,
      "num_tokens": 830725899.0,
      "step": 5212
    },
    {
      "epoch": 2.6515768056968465,
      "grad_norm": 0.8999183773994446,
      "learning_rate": 1e-05,
      "loss": 0.4255,
      "mean_token_accuracy": 0.8611314296722412,
      "num_tokens": 830890593.0,
      "step": 5213
    },
    {
      "epoch": 2.652085452695829,
      "grad_norm": 0.8834280371665955,
      "learning_rate": 1e-05,
      "loss": 0.4258,
      "mean_token_accuracy": 0.8608800768852234,
      "num_tokens": 831048409.0,
      "step": 5214
    },
    {
      "epoch": 2.6525940996948116,
      "grad_norm": 0.9180814623832703,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.8565956354141235,
      "num_tokens": 831204167.0,
      "step": 5215
    },
    {
      "epoch": 2.6531027466937944,
      "grad_norm": 0.9318739175796509,
      "learning_rate": 1e-05,
      "loss": 0.4682,
      "mean_token_accuracy": 0.8499984741210938,
      "num_tokens": 831363323.0,
      "step": 5216
    },
    {
      "epoch": 2.653611393692777,
      "grad_norm": 0.8888118863105774,
      "learning_rate": 1e-05,
      "loss": 0.4623,
      "mean_token_accuracy": 0.8512741327285767,
      "num_tokens": 831519336.0,
      "step": 5217
    },
    {
      "epoch": 2.65412004069176,
      "grad_norm": 0.9057247638702393,
      "learning_rate": 1e-05,
      "loss": 0.434,
      "mean_token_accuracy": 0.8602588176727295,
      "num_tokens": 831680359.0,
      "step": 5218
    },
    {
      "epoch": 2.654628687690743,
      "grad_norm": 0.9005915522575378,
      "learning_rate": 1e-05,
      "loss": 0.4847,
      "mean_token_accuracy": 0.8457971811294556,
      "num_tokens": 831848068.0,
      "step": 5219
    },
    {
      "epoch": 2.6551373346897256,
      "grad_norm": 0.9270437359809875,
      "learning_rate": 1e-05,
      "loss": 0.4351,
      "mean_token_accuracy": 0.8590574264526367,
      "num_tokens": 831993691.0,
      "step": 5220
    },
    {
      "epoch": 2.655645981688708,
      "grad_norm": 0.8919311761856079,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "mean_token_accuracy": 0.8572813272476196,
      "num_tokens": 832163882.0,
      "step": 5221
    },
    {
      "epoch": 2.6561546286876907,
      "grad_norm": 0.833203136920929,
      "learning_rate": 1e-05,
      "loss": 0.4304,
      "mean_token_accuracy": 0.8595755100250244,
      "num_tokens": 832317882.0,
      "step": 5222
    },
    {
      "epoch": 2.6566632756866735,
      "grad_norm": 0.8920441269874573,
      "learning_rate": 1e-05,
      "loss": 0.4583,
      "mean_token_accuracy": 0.8540435433387756,
      "num_tokens": 832484402.0,
      "step": 5223
    },
    {
      "epoch": 2.6571719226856563,
      "grad_norm": 0.8928776979446411,
      "learning_rate": 1e-05,
      "loss": 0.4392,
      "mean_token_accuracy": 0.857858419418335,
      "num_tokens": 832638879.0,
      "step": 5224
    },
    {
      "epoch": 2.657680569684639,
      "grad_norm": 0.8863247036933899,
      "learning_rate": 1e-05,
      "loss": 0.4161,
      "mean_token_accuracy": 0.8645328879356384,
      "num_tokens": 832789130.0,
      "step": 5225
    },
    {
      "epoch": 2.6581892166836214,
      "grad_norm": 1.0162090063095093,
      "learning_rate": 1e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8517112731933594,
      "num_tokens": 832949967.0,
      "step": 5226
    },
    {
      "epoch": 2.658697863682604,
      "grad_norm": 0.9088463187217712,
      "learning_rate": 1e-05,
      "loss": 0.4341,
      "mean_token_accuracy": 0.8591833710670471,
      "num_tokens": 833119872.0,
      "step": 5227
    },
    {
      "epoch": 2.659206510681587,
      "grad_norm": 0.8535668253898621,
      "learning_rate": 1e-05,
      "loss": 0.4602,
      "mean_token_accuracy": 0.8516002893447876,
      "num_tokens": 833279512.0,
      "step": 5228
    },
    {
      "epoch": 2.6597151576805698,
      "grad_norm": 1.0515375137329102,
      "learning_rate": 1e-05,
      "loss": 0.4635,
      "mean_token_accuracy": 0.8504531383514404,
      "num_tokens": 833433563.0,
      "step": 5229
    },
    {
      "epoch": 2.6602238046795526,
      "grad_norm": 0.910163938999176,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8567389249801636,
      "num_tokens": 833592791.0,
      "step": 5230
    },
    {
      "epoch": 2.660732451678535,
      "grad_norm": 0.9681553840637207,
      "learning_rate": 1e-05,
      "loss": 0.4037,
      "mean_token_accuracy": 0.8662945628166199,
      "num_tokens": 833763220.0,
      "step": 5231
    },
    {
      "epoch": 2.6612410986775177,
      "grad_norm": 0.8657732009887695,
      "learning_rate": 1e-05,
      "loss": 0.4031,
      "mean_token_accuracy": 0.8685414791107178,
      "num_tokens": 833906060.0,
      "step": 5232
    },
    {
      "epoch": 2.6617497456765005,
      "grad_norm": 0.9288004040718079,
      "learning_rate": 1e-05,
      "loss": 0.4358,
      "mean_token_accuracy": 0.8585642576217651,
      "num_tokens": 834064692.0,
      "step": 5233
    },
    {
      "epoch": 2.6622583926754833,
      "grad_norm": 0.8908243179321289,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.8539721965789795,
      "num_tokens": 834238761.0,
      "step": 5234
    },
    {
      "epoch": 2.662767039674466,
      "grad_norm": 0.8896021842956543,
      "learning_rate": 1e-05,
      "loss": 0.4372,
      "mean_token_accuracy": 0.8590289354324341,
      "num_tokens": 834394611.0,
      "step": 5235
    },
    {
      "epoch": 2.6632756866734484,
      "grad_norm": 0.9499205350875854,
      "learning_rate": 1e-05,
      "loss": 0.436,
      "mean_token_accuracy": 0.8576411008834839,
      "num_tokens": 834556371.0,
      "step": 5236
    },
    {
      "epoch": 2.663784333672431,
      "grad_norm": 0.9268885254859924,
      "learning_rate": 1e-05,
      "loss": 0.4321,
      "mean_token_accuracy": 0.859523355960846,
      "num_tokens": 834720648.0,
      "step": 5237
    },
    {
      "epoch": 2.664292980671414,
      "grad_norm": 0.973332941532135,
      "learning_rate": 1e-05,
      "loss": 0.4416,
      "mean_token_accuracy": 0.8574525117874146,
      "num_tokens": 834886463.0,
      "step": 5238
    },
    {
      "epoch": 2.6648016276703967,
      "grad_norm": 0.8768401741981506,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8543452620506287,
      "num_tokens": 835045036.0,
      "step": 5239
    },
    {
      "epoch": 2.6653102746693795,
      "grad_norm": 0.9415298700332642,
      "learning_rate": 1e-05,
      "loss": 0.4741,
      "mean_token_accuracy": 0.8475942611694336,
      "num_tokens": 835210506.0,
      "step": 5240
    },
    {
      "epoch": 2.6658189216683623,
      "grad_norm": 0.8788790106773376,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8509476184844971,
      "num_tokens": 835378216.0,
      "step": 5241
    },
    {
      "epoch": 2.666327568667345,
      "grad_norm": 0.9069360494613647,
      "learning_rate": 1e-05,
      "loss": 0.4477,
      "mean_token_accuracy": 0.8549486994743347,
      "num_tokens": 835541684.0,
      "step": 5242
    },
    {
      "epoch": 2.6668362156663274,
      "grad_norm": 0.8734854459762573,
      "learning_rate": 1e-05,
      "loss": 0.4218,
      "mean_token_accuracy": 0.862857460975647,
      "num_tokens": 835698814.0,
      "step": 5243
    },
    {
      "epoch": 2.6673448626653102,
      "grad_norm": 0.8554462790489197,
      "learning_rate": 1e-05,
      "loss": 0.4067,
      "mean_token_accuracy": 0.8660014867782593,
      "num_tokens": 835845878.0,
      "step": 5244
    },
    {
      "epoch": 2.667853509664293,
      "grad_norm": 0.9750162959098816,
      "learning_rate": 1e-05,
      "loss": 0.4122,
      "mean_token_accuracy": 0.865595281124115,
      "num_tokens": 836006330.0,
      "step": 5245
    },
    {
      "epoch": 2.668362156663276,
      "grad_norm": 0.8666068315505981,
      "learning_rate": 1e-05,
      "loss": 0.4482,
      "mean_token_accuracy": 0.8550605773925781,
      "num_tokens": 836171735.0,
      "step": 5246
    },
    {
      "epoch": 2.6688708036622586,
      "grad_norm": 1.0751583576202393,
      "learning_rate": 1e-05,
      "loss": 0.4291,
      "mean_token_accuracy": 0.8609598875045776,
      "num_tokens": 836330652.0,
      "step": 5247
    },
    {
      "epoch": 2.669379450661241,
      "grad_norm": 0.9035525321960449,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8590593338012695,
      "num_tokens": 836490407.0,
      "step": 5248
    },
    {
      "epoch": 2.6698880976602237,
      "grad_norm": 0.8673198223114014,
      "learning_rate": 1e-05,
      "loss": 0.4581,
      "mean_token_accuracy": 0.853705644607544,
      "num_tokens": 836657627.0,
      "step": 5249
    },
    {
      "epoch": 2.6703967446592065,
      "grad_norm": 0.8926166296005249,
      "learning_rate": 1e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8502712845802307,
      "num_tokens": 836806889.0,
      "step": 5250
    },
    {
      "epoch": 2.6709053916581893,
      "grad_norm": 0.8546720147132874,
      "learning_rate": 1e-05,
      "loss": 0.4045,
      "mean_token_accuracy": 0.8660810589790344,
      "num_tokens": 836957002.0,
      "step": 5251
    },
    {
      "epoch": 2.671414038657172,
      "grad_norm": 0.8993261456489563,
      "learning_rate": 1e-05,
      "loss": 0.4193,
      "mean_token_accuracy": 0.8635987043380737,
      "num_tokens": 837114999.0,
      "step": 5252
    },
    {
      "epoch": 2.6719226856561544,
      "grad_norm": 0.835578203201294,
      "learning_rate": 1e-05,
      "loss": 0.4183,
      "mean_token_accuracy": 0.8642175197601318,
      "num_tokens": 837278032.0,
      "step": 5253
    },
    {
      "epoch": 2.672431332655137,
      "grad_norm": 0.8844635486602783,
      "learning_rate": 1e-05,
      "loss": 0.4582,
      "mean_token_accuracy": 0.8512252569198608,
      "num_tokens": 837446232.0,
      "step": 5254
    },
    {
      "epoch": 2.67293997965412,
      "grad_norm": 0.9081737399101257,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8528986573219299,
      "num_tokens": 837606045.0,
      "step": 5255
    },
    {
      "epoch": 2.673448626653103,
      "grad_norm": 0.8913959860801697,
      "learning_rate": 1e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8601134419441223,
      "num_tokens": 837755582.0,
      "step": 5256
    },
    {
      "epoch": 2.6739572736520856,
      "grad_norm": 0.8811616897583008,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8632052540779114,
      "num_tokens": 837919220.0,
      "step": 5257
    },
    {
      "epoch": 2.674465920651068,
      "grad_norm": 0.9206438660621643,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "mean_token_accuracy": 0.8559415340423584,
      "num_tokens": 838095626.0,
      "step": 5258
    },
    {
      "epoch": 2.6749745676500507,
      "grad_norm": 0.8314093947410583,
      "learning_rate": 1e-05,
      "loss": 0.4218,
      "mean_token_accuracy": 0.8621360659599304,
      "num_tokens": 838261047.0,
      "step": 5259
    },
    {
      "epoch": 2.6754832146490335,
      "grad_norm": 0.8358904123306274,
      "learning_rate": 1e-05,
      "loss": 0.4264,
      "mean_token_accuracy": 0.8617807030677795,
      "num_tokens": 838424701.0,
      "step": 5260
    },
    {
      "epoch": 2.6759918616480163,
      "grad_norm": 0.9278913140296936,
      "learning_rate": 1e-05,
      "loss": 0.4137,
      "mean_token_accuracy": 0.8653403520584106,
      "num_tokens": 838578163.0,
      "step": 5261
    },
    {
      "epoch": 2.676500508646999,
      "grad_norm": 0.8403149843215942,
      "learning_rate": 1e-05,
      "loss": 0.389,
      "mean_token_accuracy": 0.8716791868209839,
      "num_tokens": 838749312.0,
      "step": 5262
    },
    {
      "epoch": 2.677009155645982,
      "grad_norm": 0.8926715850830078,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8627158403396606,
      "num_tokens": 838907950.0,
      "step": 5263
    },
    {
      "epoch": 2.6775178026449646,
      "grad_norm": 0.9633036851882935,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8537728786468506,
      "num_tokens": 839056583.0,
      "step": 5264
    },
    {
      "epoch": 2.678026449643947,
      "grad_norm": 0.8729433417320251,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8604938983917236,
      "num_tokens": 839210164.0,
      "step": 5265
    },
    {
      "epoch": 2.6785350966429298,
      "grad_norm": 0.8906210660934448,
      "learning_rate": 1e-05,
      "loss": 0.4153,
      "mean_token_accuracy": 0.8655133247375488,
      "num_tokens": 839375666.0,
      "step": 5266
    },
    {
      "epoch": 2.6790437436419126,
      "grad_norm": 0.9205353856086731,
      "learning_rate": 1e-05,
      "loss": 0.4236,
      "mean_token_accuracy": 0.8611639738082886,
      "num_tokens": 839530690.0,
      "step": 5267
    },
    {
      "epoch": 2.6795523906408953,
      "grad_norm": 0.8499439358711243,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8617724776268005,
      "num_tokens": 839686915.0,
      "step": 5268
    },
    {
      "epoch": 2.680061037639878,
      "grad_norm": 0.9117785692214966,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.856441855430603,
      "num_tokens": 839849116.0,
      "step": 5269
    },
    {
      "epoch": 2.6805696846388605,
      "grad_norm": 0.9000932574272156,
      "learning_rate": 1e-05,
      "loss": 0.4369,
      "mean_token_accuracy": 0.8586184978485107,
      "num_tokens": 840014745.0,
      "step": 5270
    },
    {
      "epoch": 2.6810783316378433,
      "grad_norm": 0.9353579878807068,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8558241724967957,
      "num_tokens": 840164253.0,
      "step": 5271
    },
    {
      "epoch": 2.681586978636826,
      "grad_norm": 0.9166135191917419,
      "learning_rate": 1e-05,
      "loss": 0.4654,
      "mean_token_accuracy": 0.8499311804771423,
      "num_tokens": 840328107.0,
      "step": 5272
    },
    {
      "epoch": 2.682095625635809,
      "grad_norm": 0.8736609816551208,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "mean_token_accuracy": 0.8634559512138367,
      "num_tokens": 840486789.0,
      "step": 5273
    },
    {
      "epoch": 2.6826042726347916,
      "grad_norm": 0.907143771648407,
      "learning_rate": 1e-05,
      "loss": 0.4289,
      "mean_token_accuracy": 0.8602548837661743,
      "num_tokens": 840645654.0,
      "step": 5274
    },
    {
      "epoch": 2.683112919633774,
      "grad_norm": 0.9404701590538025,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.8499409556388855,
      "num_tokens": 840794418.0,
      "step": 5275
    },
    {
      "epoch": 2.6836215666327567,
      "grad_norm": 0.9110710024833679,
      "learning_rate": 1e-05,
      "loss": 0.4306,
      "mean_token_accuracy": 0.858879566192627,
      "num_tokens": 840940420.0,
      "step": 5276
    },
    {
      "epoch": 2.6841302136317395,
      "grad_norm": 0.8523752689361572,
      "learning_rate": 1e-05,
      "loss": 0.442,
      "mean_token_accuracy": 0.855581521987915,
      "num_tokens": 841097573.0,
      "step": 5277
    },
    {
      "epoch": 2.6846388606307223,
      "grad_norm": 0.9715843796730042,
      "learning_rate": 1e-05,
      "loss": 0.4193,
      "mean_token_accuracy": 0.8620656728744507,
      "num_tokens": 841254500.0,
      "step": 5278
    },
    {
      "epoch": 2.685147507629705,
      "grad_norm": 0.9293642044067383,
      "learning_rate": 1e-05,
      "loss": 0.4394,
      "mean_token_accuracy": 0.8580410480499268,
      "num_tokens": 841405456.0,
      "step": 5279
    },
    {
      "epoch": 2.6856561546286875,
      "grad_norm": 0.8933702111244202,
      "learning_rate": 1e-05,
      "loss": 0.4288,
      "mean_token_accuracy": 0.8603155016899109,
      "num_tokens": 841560795.0,
      "step": 5280
    },
    {
      "epoch": 2.6861648016276702,
      "grad_norm": 0.9727333188056946,
      "learning_rate": 1e-05,
      "loss": 0.4187,
      "mean_token_accuracy": 0.8624151945114136,
      "num_tokens": 841717082.0,
      "step": 5281
    },
    {
      "epoch": 2.686673448626653,
      "grad_norm": 0.9016296863555908,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8495412468910217,
      "num_tokens": 841878106.0,
      "step": 5282
    },
    {
      "epoch": 2.687182095625636,
      "grad_norm": 0.8928479552268982,
      "learning_rate": 1e-05,
      "loss": 0.4379,
      "mean_token_accuracy": 0.8579577207565308,
      "num_tokens": 842040480.0,
      "step": 5283
    },
    {
      "epoch": 2.6876907426246186,
      "grad_norm": 1.0036600828170776,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8550857305526733,
      "num_tokens": 842189540.0,
      "step": 5284
    },
    {
      "epoch": 2.688199389623601,
      "grad_norm": 0.8805148005485535,
      "learning_rate": 1e-05,
      "loss": 0.445,
      "mean_token_accuracy": 0.8565136194229126,
      "num_tokens": 842345555.0,
      "step": 5285
    },
    {
      "epoch": 2.688708036622584,
      "grad_norm": 0.905439019203186,
      "learning_rate": 1e-05,
      "loss": 0.408,
      "mean_token_accuracy": 0.8662591576576233,
      "num_tokens": 842497226.0,
      "step": 5286
    },
    {
      "epoch": 2.6892166836215665,
      "grad_norm": 0.9065870046615601,
      "learning_rate": 1e-05,
      "loss": 0.4245,
      "mean_token_accuracy": 0.8618921637535095,
      "num_tokens": 842654020.0,
      "step": 5287
    },
    {
      "epoch": 2.6897253306205493,
      "grad_norm": 0.8972324132919312,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8648066520690918,
      "num_tokens": 842814612.0,
      "step": 5288
    },
    {
      "epoch": 2.690233977619532,
      "grad_norm": 0.9083245992660522,
      "learning_rate": 1e-05,
      "loss": 0.4448,
      "mean_token_accuracy": 0.855293869972229,
      "num_tokens": 842969822.0,
      "step": 5289
    },
    {
      "epoch": 2.690742624618515,
      "grad_norm": 1.0097336769104004,
      "learning_rate": 1e-05,
      "loss": 0.4653,
      "mean_token_accuracy": 0.8518034219741821,
      "num_tokens": 843117888.0,
      "step": 5290
    },
    {
      "epoch": 2.6912512716174977,
      "grad_norm": 0.8871720433235168,
      "learning_rate": 1e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.858801007270813,
      "num_tokens": 843276391.0,
      "step": 5291
    },
    {
      "epoch": 2.69175991861648,
      "grad_norm": 0.8812194466590881,
      "learning_rate": 1e-05,
      "loss": 0.4198,
      "mean_token_accuracy": 0.8638508319854736,
      "num_tokens": 843448573.0,
      "step": 5292
    },
    {
      "epoch": 2.692268565615463,
      "grad_norm": 0.9194967150688171,
      "learning_rate": 1e-05,
      "loss": 0.4313,
      "mean_token_accuracy": 0.8597676157951355,
      "num_tokens": 843610730.0,
      "step": 5293
    },
    {
      "epoch": 2.6927772126144456,
      "grad_norm": 0.963604211807251,
      "learning_rate": 1e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.8583006858825684,
      "num_tokens": 843753683.0,
      "step": 5294
    },
    {
      "epoch": 2.6932858596134284,
      "grad_norm": 0.9230378866195679,
      "learning_rate": 1e-05,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8627330660820007,
      "num_tokens": 843921987.0,
      "step": 5295
    },
    {
      "epoch": 2.693794506612411,
      "grad_norm": 0.8835469484329224,
      "learning_rate": 1e-05,
      "loss": 0.4258,
      "mean_token_accuracy": 0.8617545962333679,
      "num_tokens": 844084227.0,
      "step": 5296
    },
    {
      "epoch": 2.6943031536113935,
      "grad_norm": 1.0125384330749512,
      "learning_rate": 1e-05,
      "loss": 0.4426,
      "mean_token_accuracy": 0.8542388677597046,
      "num_tokens": 844241768.0,
      "step": 5297
    },
    {
      "epoch": 2.6948118006103763,
      "grad_norm": 0.9577308893203735,
      "learning_rate": 1e-05,
      "loss": 0.4608,
      "mean_token_accuracy": 0.8525644540786743,
      "num_tokens": 844401669.0,
      "step": 5298
    },
    {
      "epoch": 2.695320447609359,
      "grad_norm": 0.9190992116928101,
      "learning_rate": 1e-05,
      "loss": 0.3843,
      "mean_token_accuracy": 0.8727595806121826,
      "num_tokens": 844543490.0,
      "step": 5299
    },
    {
      "epoch": 2.695829094608342,
      "grad_norm": 1.065569281578064,
      "learning_rate": 1e-05,
      "loss": 0.4478,
      "mean_token_accuracy": 0.8555886745452881,
      "num_tokens": 844700667.0,
      "step": 5300
    },
    {
      "epoch": 2.6963377416073246,
      "grad_norm": 1.1497288942337036,
      "learning_rate": 1e-05,
      "loss": 0.4772,
      "mean_token_accuracy": 0.8471383452415466,
      "num_tokens": 844859397.0,
      "step": 5301
    },
    {
      "epoch": 2.696846388606307,
      "grad_norm": 1.0540322065353394,
      "learning_rate": 1e-05,
      "loss": 0.4164,
      "mean_token_accuracy": 0.8639854192733765,
      "num_tokens": 845028014.0,
      "step": 5302
    },
    {
      "epoch": 2.6973550356052898,
      "grad_norm": 0.9754948019981384,
      "learning_rate": 1e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8623305559158325,
      "num_tokens": 845188251.0,
      "step": 5303
    },
    {
      "epoch": 2.6978636826042726,
      "grad_norm": 0.9905082583427429,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8608218431472778,
      "num_tokens": 845341860.0,
      "step": 5304
    },
    {
      "epoch": 2.6983723296032553,
      "grad_norm": 0.9921637177467346,
      "learning_rate": 1e-05,
      "loss": 0.413,
      "mean_token_accuracy": 0.8662896156311035,
      "num_tokens": 845496873.0,
      "step": 5305
    },
    {
      "epoch": 2.698880976602238,
      "grad_norm": 1.0702582597732544,
      "learning_rate": 1e-05,
      "loss": 0.453,
      "mean_token_accuracy": 0.8552283048629761,
      "num_tokens": 845655021.0,
      "step": 5306
    },
    {
      "epoch": 2.6993896236012205,
      "grad_norm": 1.0450761318206787,
      "learning_rate": 1e-05,
      "loss": 0.4211,
      "mean_token_accuracy": 0.8628575801849365,
      "num_tokens": 845820482.0,
      "step": 5307
    },
    {
      "epoch": 2.6998982706002037,
      "grad_norm": 0.9499820470809937,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8609029650688171,
      "num_tokens": 845981043.0,
      "step": 5308
    },
    {
      "epoch": 2.700406917599186,
      "grad_norm": 1.038619875907898,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "mean_token_accuracy": 0.8579244017601013,
      "num_tokens": 846137074.0,
      "step": 5309
    },
    {
      "epoch": 2.700915564598169,
      "grad_norm": 0.97670978307724,
      "learning_rate": 1e-05,
      "loss": 0.4531,
      "mean_token_accuracy": 0.8525657653808594,
      "num_tokens": 846289014.0,
      "step": 5310
    },
    {
      "epoch": 2.7014242115971516,
      "grad_norm": 0.8953006267547607,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8571546077728271,
      "num_tokens": 846463658.0,
      "step": 5311
    },
    {
      "epoch": 2.7019328585961344,
      "grad_norm": 1.0701743364334106,
      "learning_rate": 1e-05,
      "loss": 0.4394,
      "mean_token_accuracy": 0.8594745397567749,
      "num_tokens": 846620497.0,
      "step": 5312
    },
    {
      "epoch": 2.702441505595117,
      "grad_norm": 0.9389851689338684,
      "learning_rate": 1e-05,
      "loss": 0.3982,
      "mean_token_accuracy": 0.8693704605102539,
      "num_tokens": 846782407.0,
      "step": 5313
    },
    {
      "epoch": 2.7029501525940995,
      "grad_norm": 1.068837285041809,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "mean_token_accuracy": 0.8585376143455505,
      "num_tokens": 846939843.0,
      "step": 5314
    },
    {
      "epoch": 2.7034587995930823,
      "grad_norm": 0.9822969436645508,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "mean_token_accuracy": 0.8587103486061096,
      "num_tokens": 847100525.0,
      "step": 5315
    },
    {
      "epoch": 2.703967446592065,
      "grad_norm": 0.973683774471283,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8573872447013855,
      "num_tokens": 847258615.0,
      "step": 5316
    },
    {
      "epoch": 2.704476093591048,
      "grad_norm": 1.0660804510116577,
      "learning_rate": 1e-05,
      "loss": 0.4266,
      "mean_token_accuracy": 0.8625261783599854,
      "num_tokens": 847412916.0,
      "step": 5317
    },
    {
      "epoch": 2.7049847405900307,
      "grad_norm": 0.8933954238891602,
      "learning_rate": 1e-05,
      "loss": 0.4272,
      "mean_token_accuracy": 0.8619201183319092,
      "num_tokens": 847566136.0,
      "step": 5318
    },
    {
      "epoch": 2.705493387589013,
      "grad_norm": 1.0075711011886597,
      "learning_rate": 1e-05,
      "loss": 0.46,
      "mean_token_accuracy": 0.8495711088180542,
      "num_tokens": 847728434.0,
      "step": 5319
    },
    {
      "epoch": 2.706002034587996,
      "grad_norm": 0.9256721138954163,
      "learning_rate": 1e-05,
      "loss": 0.4518,
      "mean_token_accuracy": 0.8546531200408936,
      "num_tokens": 847880060.0,
      "step": 5320
    },
    {
      "epoch": 2.7065106815869786,
      "grad_norm": 0.9893863797187805,
      "learning_rate": 1e-05,
      "loss": 0.3943,
      "mean_token_accuracy": 0.8700924515724182,
      "num_tokens": 848022342.0,
      "step": 5321
    },
    {
      "epoch": 2.7070193285859614,
      "grad_norm": 0.9653140306472778,
      "learning_rate": 1e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8531556129455566,
      "num_tokens": 848174723.0,
      "step": 5322
    },
    {
      "epoch": 2.707527975584944,
      "grad_norm": 0.9758864641189575,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8584966659545898,
      "num_tokens": 848325984.0,
      "step": 5323
    },
    {
      "epoch": 2.7080366225839265,
      "grad_norm": 0.8586357235908508,
      "learning_rate": 1e-05,
      "loss": 0.4152,
      "mean_token_accuracy": 0.8653855323791504,
      "num_tokens": 848487841.0,
      "step": 5324
    },
    {
      "epoch": 2.7085452695829093,
      "grad_norm": 0.9456252455711365,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.861905574798584,
      "num_tokens": 848636882.0,
      "step": 5325
    },
    {
      "epoch": 2.709053916581892,
      "grad_norm": 0.8794358372688293,
      "learning_rate": 1e-05,
      "loss": 0.4379,
      "mean_token_accuracy": 0.8578913807868958,
      "num_tokens": 848802975.0,
      "step": 5326
    },
    {
      "epoch": 2.709562563580875,
      "grad_norm": 0.8466426730155945,
      "learning_rate": 1e-05,
      "loss": 0.4398,
      "mean_token_accuracy": 0.8574482798576355,
      "num_tokens": 848967586.0,
      "step": 5327
    },
    {
      "epoch": 2.7100712105798577,
      "grad_norm": 0.9271215796470642,
      "learning_rate": 1e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8618929386138916,
      "num_tokens": 849119947.0,
      "step": 5328
    },
    {
      "epoch": 2.71057985757884,
      "grad_norm": 0.9016660451889038,
      "learning_rate": 1e-05,
      "loss": 0.4078,
      "mean_token_accuracy": 0.8671656847000122,
      "num_tokens": 849278651.0,
      "step": 5329
    },
    {
      "epoch": 2.7110885045778232,
      "grad_norm": 0.9124810695648193,
      "learning_rate": 1e-05,
      "loss": 0.4176,
      "mean_token_accuracy": 0.8643842935562134,
      "num_tokens": 849449730.0,
      "step": 5330
    },
    {
      "epoch": 2.7115971515768056,
      "grad_norm": 0.878895103931427,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "mean_token_accuracy": 0.8634142875671387,
      "num_tokens": 849606858.0,
      "step": 5331
    },
    {
      "epoch": 2.7121057985757884,
      "grad_norm": 0.9118165373802185,
      "learning_rate": 1e-05,
      "loss": 0.4594,
      "mean_token_accuracy": 0.8515703678131104,
      "num_tokens": 849776273.0,
      "step": 5332
    },
    {
      "epoch": 2.712614445574771,
      "grad_norm": 0.9616394639015198,
      "learning_rate": 1e-05,
      "loss": 0.4341,
      "mean_token_accuracy": 0.8593379259109497,
      "num_tokens": 849934124.0,
      "step": 5333
    },
    {
      "epoch": 2.713123092573754,
      "grad_norm": 0.9179736375808716,
      "learning_rate": 1e-05,
      "loss": 0.4465,
      "mean_token_accuracy": 0.855577826499939,
      "num_tokens": 850089497.0,
      "step": 5334
    },
    {
      "epoch": 2.7136317395727367,
      "grad_norm": 0.9623274207115173,
      "learning_rate": 1e-05,
      "loss": 0.4419,
      "mean_token_accuracy": 0.8587888479232788,
      "num_tokens": 850252982.0,
      "step": 5335
    },
    {
      "epoch": 2.714140386571719,
      "grad_norm": 0.9166063070297241,
      "learning_rate": 1e-05,
      "loss": 0.4202,
      "mean_token_accuracy": 0.8629820346832275,
      "num_tokens": 850402164.0,
      "step": 5336
    },
    {
      "epoch": 2.714649033570702,
      "grad_norm": 0.8992436528205872,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.859114944934845,
      "num_tokens": 850563953.0,
      "step": 5337
    },
    {
      "epoch": 2.7151576805696847,
      "grad_norm": 0.8603736162185669,
      "learning_rate": 1e-05,
      "loss": 0.4544,
      "mean_token_accuracy": 0.854015588760376,
      "num_tokens": 850732466.0,
      "step": 5338
    },
    {
      "epoch": 2.7156663275686674,
      "grad_norm": 0.9173347353935242,
      "learning_rate": 1e-05,
      "loss": 0.4126,
      "mean_token_accuracy": 0.8645404577255249,
      "num_tokens": 850881803.0,
      "step": 5339
    },
    {
      "epoch": 2.7161749745676502,
      "grad_norm": 0.9201670289039612,
      "learning_rate": 1e-05,
      "loss": 0.4526,
      "mean_token_accuracy": 0.8551905751228333,
      "num_tokens": 851032068.0,
      "step": 5340
    },
    {
      "epoch": 2.7166836215666326,
      "grad_norm": 0.8854498267173767,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "mean_token_accuracy": 0.8620210886001587,
      "num_tokens": 851184436.0,
      "step": 5341
    },
    {
      "epoch": 2.7171922685656154,
      "grad_norm": 0.8970783948898315,
      "learning_rate": 1e-05,
      "loss": 0.4309,
      "mean_token_accuracy": 0.8604885339736938,
      "num_tokens": 851344640.0,
      "step": 5342
    },
    {
      "epoch": 2.717700915564598,
      "grad_norm": 0.8435708284378052,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "mean_token_accuracy": 0.8662682771682739,
      "num_tokens": 851496643.0,
      "step": 5343
    },
    {
      "epoch": 2.718209562563581,
      "grad_norm": 0.9770410060882568,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "mean_token_accuracy": 0.8620365858078003,
      "num_tokens": 851650774.0,
      "step": 5344
    },
    {
      "epoch": 2.7187182095625637,
      "grad_norm": 0.941686749458313,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8531053066253662,
      "num_tokens": 851807168.0,
      "step": 5345
    },
    {
      "epoch": 2.719226856561546,
      "grad_norm": 0.9545437097549438,
      "learning_rate": 1e-05,
      "loss": 0.4636,
      "mean_token_accuracy": 0.8520953059196472,
      "num_tokens": 851963425.0,
      "step": 5346
    },
    {
      "epoch": 2.719735503560529,
      "grad_norm": 0.9199371933937073,
      "learning_rate": 1e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.8604966998100281,
      "num_tokens": 852127672.0,
      "step": 5347
    },
    {
      "epoch": 2.7202441505595116,
      "grad_norm": 0.8954357504844666,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "mean_token_accuracy": 0.8590489625930786,
      "num_tokens": 852286142.0,
      "step": 5348
    },
    {
      "epoch": 2.7207527975584944,
      "grad_norm": 0.9719069004058838,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.8575436472892761,
      "num_tokens": 852423657.0,
      "step": 5349
    },
    {
      "epoch": 2.721261444557477,
      "grad_norm": 0.8555158376693726,
      "learning_rate": 1e-05,
      "loss": 0.4039,
      "mean_token_accuracy": 0.8683286905288696,
      "num_tokens": 852582032.0,
      "step": 5350
    },
    {
      "epoch": 2.7217700915564595,
      "grad_norm": 0.9108577966690063,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.8620162010192871,
      "num_tokens": 852731358.0,
      "step": 5351
    },
    {
      "epoch": 2.722278738555443,
      "grad_norm": 0.9062488079071045,
      "learning_rate": 1e-05,
      "loss": 0.4061,
      "mean_token_accuracy": 0.8677446246147156,
      "num_tokens": 852884998.0,
      "step": 5352
    },
    {
      "epoch": 2.722787385554425,
      "grad_norm": 0.926082193851471,
      "learning_rate": 1e-05,
      "loss": 0.4133,
      "mean_token_accuracy": 0.8656351566314697,
      "num_tokens": 853052940.0,
      "step": 5353
    },
    {
      "epoch": 2.723296032553408,
      "grad_norm": 0.9328706860542297,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8513386845588684,
      "num_tokens": 853215576.0,
      "step": 5354
    },
    {
      "epoch": 2.7238046795523907,
      "grad_norm": 0.9114854335784912,
      "learning_rate": 1e-05,
      "loss": 0.4695,
      "mean_token_accuracy": 0.8492288589477539,
      "num_tokens": 853379664.0,
      "step": 5355
    },
    {
      "epoch": 2.7243133265513735,
      "grad_norm": 0.8919292092323303,
      "learning_rate": 1e-05,
      "loss": 0.4308,
      "mean_token_accuracy": 0.8609844446182251,
      "num_tokens": 853539145.0,
      "step": 5356
    },
    {
      "epoch": 2.7248219735503563,
      "grad_norm": 0.9118651151657104,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8546961545944214,
      "num_tokens": 853695632.0,
      "step": 5357
    },
    {
      "epoch": 2.7253306205493386,
      "grad_norm": 0.9264186024665833,
      "learning_rate": 1e-05,
      "loss": 0.3998,
      "mean_token_accuracy": 0.8693535923957825,
      "num_tokens": 853850299.0,
      "step": 5358
    },
    {
      "epoch": 2.7258392675483214,
      "grad_norm": 0.9042052626609802,
      "learning_rate": 1e-05,
      "loss": 0.4388,
      "mean_token_accuracy": 0.8559480309486389,
      "num_tokens": 854002137.0,
      "step": 5359
    },
    {
      "epoch": 2.726347914547304,
      "grad_norm": 0.918724536895752,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.8582726120948792,
      "num_tokens": 854162459.0,
      "step": 5360
    },
    {
      "epoch": 2.726856561546287,
      "grad_norm": 0.841949999332428,
      "learning_rate": 1e-05,
      "loss": 0.3968,
      "mean_token_accuracy": 0.8698896765708923,
      "num_tokens": 854320980.0,
      "step": 5361
    },
    {
      "epoch": 2.7273652085452698,
      "grad_norm": 0.8705538511276245,
      "learning_rate": 1e-05,
      "loss": 0.4274,
      "mean_token_accuracy": 0.8598239421844482,
      "num_tokens": 854480464.0,
      "step": 5362
    },
    {
      "epoch": 2.727873855544252,
      "grad_norm": 0.8634495139122009,
      "learning_rate": 1e-05,
      "loss": 0.4234,
      "mean_token_accuracy": 0.862626314163208,
      "num_tokens": 854639187.0,
      "step": 5363
    },
    {
      "epoch": 2.728382502543235,
      "grad_norm": 0.9334564208984375,
      "learning_rate": 1e-05,
      "loss": 0.4344,
      "mean_token_accuracy": 0.8592795133590698,
      "num_tokens": 854791740.0,
      "step": 5364
    },
    {
      "epoch": 2.7288911495422177,
      "grad_norm": 0.8462790846824646,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.8633471727371216,
      "num_tokens": 854959936.0,
      "step": 5365
    },
    {
      "epoch": 2.7293997965412005,
      "grad_norm": 0.8552737832069397,
      "learning_rate": 1e-05,
      "loss": 0.4143,
      "mean_token_accuracy": 0.86388099193573,
      "num_tokens": 855114343.0,
      "step": 5366
    },
    {
      "epoch": 2.7299084435401832,
      "grad_norm": 0.896813154220581,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "mean_token_accuracy": 0.8606491088867188,
      "num_tokens": 855274165.0,
      "step": 5367
    },
    {
      "epoch": 2.7304170905391656,
      "grad_norm": 0.8741781115531921,
      "learning_rate": 1e-05,
      "loss": 0.4101,
      "mean_token_accuracy": 0.8644659519195557,
      "num_tokens": 855437674.0,
      "step": 5368
    },
    {
      "epoch": 2.7309257375381484,
      "grad_norm": 0.915451169013977,
      "learning_rate": 1e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.8615007996559143,
      "num_tokens": 855605378.0,
      "step": 5369
    },
    {
      "epoch": 2.731434384537131,
      "grad_norm": 0.9084851741790771,
      "learning_rate": 1e-05,
      "loss": 0.4309,
      "mean_token_accuracy": 0.8610782623291016,
      "num_tokens": 855766653.0,
      "step": 5370
    },
    {
      "epoch": 2.731943031536114,
      "grad_norm": 0.8731338977813721,
      "learning_rate": 1e-05,
      "loss": 0.4212,
      "mean_token_accuracy": 0.8621797561645508,
      "num_tokens": 855925570.0,
      "step": 5371
    },
    {
      "epoch": 2.7324516785350967,
      "grad_norm": 0.8459338545799255,
      "learning_rate": 1e-05,
      "loss": 0.4635,
      "mean_token_accuracy": 0.8504737615585327,
      "num_tokens": 856096377.0,
      "step": 5372
    },
    {
      "epoch": 2.732960325534079,
      "grad_norm": 0.8906593918800354,
      "learning_rate": 1e-05,
      "loss": 0.4381,
      "mean_token_accuracy": 0.857587456703186,
      "num_tokens": 856248152.0,
      "step": 5373
    },
    {
      "epoch": 2.7334689725330623,
      "grad_norm": 0.9160838723182678,
      "learning_rate": 1e-05,
      "loss": 0.446,
      "mean_token_accuracy": 0.855549156665802,
      "num_tokens": 856408236.0,
      "step": 5374
    },
    {
      "epoch": 2.7339776195320447,
      "grad_norm": 0.8952018618583679,
      "learning_rate": 1e-05,
      "loss": 0.4576,
      "mean_token_accuracy": 0.8525691628456116,
      "num_tokens": 856563894.0,
      "step": 5375
    },
    {
      "epoch": 2.7344862665310274,
      "grad_norm": 0.8565658926963806,
      "learning_rate": 1e-05,
      "loss": 0.4328,
      "mean_token_accuracy": 0.8586984276771545,
      "num_tokens": 856726568.0,
      "step": 5376
    },
    {
      "epoch": 2.7349949135300102,
      "grad_norm": 0.8847385048866272,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "mean_token_accuracy": 0.8605058193206787,
      "num_tokens": 856885734.0,
      "step": 5377
    },
    {
      "epoch": 2.735503560528993,
      "grad_norm": 0.9560387134552002,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8562148809432983,
      "num_tokens": 857033785.0,
      "step": 5378
    },
    {
      "epoch": 2.736012207527976,
      "grad_norm": 0.9595639109611511,
      "learning_rate": 1e-05,
      "loss": 0.4265,
      "mean_token_accuracy": 0.86170893907547,
      "num_tokens": 857194080.0,
      "step": 5379
    },
    {
      "epoch": 2.736520854526958,
      "grad_norm": 0.8554273247718811,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "mean_token_accuracy": 0.8590626120567322,
      "num_tokens": 857354595.0,
      "step": 5380
    },
    {
      "epoch": 2.737029501525941,
      "grad_norm": 0.9065148830413818,
      "learning_rate": 1e-05,
      "loss": 0.4217,
      "mean_token_accuracy": 0.863584041595459,
      "num_tokens": 857506990.0,
      "step": 5381
    },
    {
      "epoch": 2.7375381485249237,
      "grad_norm": 0.8713171482086182,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8588833212852478,
      "num_tokens": 857678702.0,
      "step": 5382
    },
    {
      "epoch": 2.7380467955239065,
      "grad_norm": 0.845331072807312,
      "learning_rate": 1e-05,
      "loss": 0.4405,
      "mean_token_accuracy": 0.8578164577484131,
      "num_tokens": 857837239.0,
      "step": 5383
    },
    {
      "epoch": 2.7385554425228893,
      "grad_norm": 0.8636237978935242,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.8616138696670532,
      "num_tokens": 858003534.0,
      "step": 5384
    },
    {
      "epoch": 2.7390640895218716,
      "grad_norm": 0.829346776008606,
      "learning_rate": 1e-05,
      "loss": 0.4485,
      "mean_token_accuracy": 0.8571034073829651,
      "num_tokens": 858184527.0,
      "step": 5385
    },
    {
      "epoch": 2.7395727365208544,
      "grad_norm": 0.8747344017028809,
      "learning_rate": 1e-05,
      "loss": 0.458,
      "mean_token_accuracy": 0.8544100522994995,
      "num_tokens": 858340027.0,
      "step": 5386
    },
    {
      "epoch": 2.740081383519837,
      "grad_norm": 0.8807426691055298,
      "learning_rate": 1e-05,
      "loss": 0.4507,
      "mean_token_accuracy": 0.8541029691696167,
      "num_tokens": 858501686.0,
      "step": 5387
    },
    {
      "epoch": 2.74059003051882,
      "grad_norm": 0.8712559938430786,
      "learning_rate": 1e-05,
      "loss": 0.4528,
      "mean_token_accuracy": 0.8543274402618408,
      "num_tokens": 858667008.0,
      "step": 5388
    },
    {
      "epoch": 2.741098677517803,
      "grad_norm": 0.9187637567520142,
      "learning_rate": 1e-05,
      "loss": 0.4588,
      "mean_token_accuracy": 0.8529598712921143,
      "num_tokens": 858829541.0,
      "step": 5389
    },
    {
      "epoch": 2.741607324516785,
      "grad_norm": 0.8646373748779297,
      "learning_rate": 1e-05,
      "loss": 0.4231,
      "mean_token_accuracy": 0.8623921871185303,
      "num_tokens": 858997524.0,
      "step": 5390
    },
    {
      "epoch": 2.742115971515768,
      "grad_norm": 0.9268781542778015,
      "learning_rate": 1e-05,
      "loss": 0.4373,
      "mean_token_accuracy": 0.8578705787658691,
      "num_tokens": 859151303.0,
      "step": 5391
    },
    {
      "epoch": 2.7426246185147507,
      "grad_norm": 0.8708943128585815,
      "learning_rate": 1e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.8591492772102356,
      "num_tokens": 859308138.0,
      "step": 5392
    },
    {
      "epoch": 2.7431332655137335,
      "grad_norm": 0.8756990432739258,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8604567646980286,
      "num_tokens": 859468758.0,
      "step": 5393
    },
    {
      "epoch": 2.7436419125127163,
      "grad_norm": 0.9198269248008728,
      "learning_rate": 1e-05,
      "loss": 0.4562,
      "mean_token_accuracy": 0.8523392677307129,
      "num_tokens": 859624523.0,
      "step": 5394
    },
    {
      "epoch": 2.7441505595116986,
      "grad_norm": 0.8417802453041077,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8584886789321899,
      "num_tokens": 859793482.0,
      "step": 5395
    },
    {
      "epoch": 2.744659206510682,
      "grad_norm": 0.846386730670929,
      "learning_rate": 1e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8598151803016663,
      "num_tokens": 859964920.0,
      "step": 5396
    },
    {
      "epoch": 2.745167853509664,
      "grad_norm": 0.9259663820266724,
      "learning_rate": 1e-05,
      "loss": 0.4666,
      "mean_token_accuracy": 0.8496070504188538,
      "num_tokens": 860126092.0,
      "step": 5397
    },
    {
      "epoch": 2.745676500508647,
      "grad_norm": 0.8447114825248718,
      "learning_rate": 1e-05,
      "loss": 0.4115,
      "mean_token_accuracy": 0.865730881690979,
      "num_tokens": 860276802.0,
      "step": 5398
    },
    {
      "epoch": 2.7461851475076298,
      "grad_norm": 0.838446855545044,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "mean_token_accuracy": 0.8559407591819763,
      "num_tokens": 860439424.0,
      "step": 5399
    },
    {
      "epoch": 2.7466937945066126,
      "grad_norm": 0.8424673080444336,
      "learning_rate": 1e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.860883891582489,
      "num_tokens": 860603738.0,
      "step": 5400
    },
    {
      "epoch": 2.7472024415055953,
      "grad_norm": 0.8394173383712769,
      "learning_rate": 1e-05,
      "loss": 0.4237,
      "mean_token_accuracy": 0.8619527816772461,
      "num_tokens": 860762795.0,
      "step": 5401
    },
    {
      "epoch": 2.7477110885045777,
      "grad_norm": 0.8642099499702454,
      "learning_rate": 1e-05,
      "loss": 0.4355,
      "mean_token_accuracy": 0.8594051003456116,
      "num_tokens": 860922428.0,
      "step": 5402
    },
    {
      "epoch": 2.7482197355035605,
      "grad_norm": 0.8128042221069336,
      "learning_rate": 1e-05,
      "loss": 0.3997,
      "mean_token_accuracy": 0.8694398403167725,
      "num_tokens": 861090451.0,
      "step": 5403
    },
    {
      "epoch": 2.7487283825025433,
      "grad_norm": 0.8231059908866882,
      "learning_rate": 1e-05,
      "loss": 0.4087,
      "mean_token_accuracy": 0.8662785887718201,
      "num_tokens": 861269373.0,
      "step": 5404
    },
    {
      "epoch": 2.749237029501526,
      "grad_norm": 0.8843866586685181,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8592203855514526,
      "num_tokens": 861434807.0,
      "step": 5405
    },
    {
      "epoch": 2.749745676500509,
      "grad_norm": 0.83979731798172,
      "learning_rate": 1e-05,
      "loss": 0.4496,
      "mean_token_accuracy": 0.8535430431365967,
      "num_tokens": 861599524.0,
      "step": 5406
    },
    {
      "epoch": 2.750254323499491,
      "grad_norm": 0.9112589359283447,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.8580020070075989,
      "num_tokens": 861753733.0,
      "step": 5407
    },
    {
      "epoch": 2.750762970498474,
      "grad_norm": 0.9031829237937927,
      "learning_rate": 1e-05,
      "loss": 0.4541,
      "mean_token_accuracy": 0.8531202673912048,
      "num_tokens": 861913031.0,
      "step": 5408
    },
    {
      "epoch": 2.7512716174974567,
      "grad_norm": 0.8952576518058777,
      "learning_rate": 1e-05,
      "loss": 0.424,
      "mean_token_accuracy": 0.8609805107116699,
      "num_tokens": 862064227.0,
      "step": 5409
    },
    {
      "epoch": 2.7517802644964395,
      "grad_norm": 0.9540919661521912,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "mean_token_accuracy": 0.8651838302612305,
      "num_tokens": 862222025.0,
      "step": 5410
    },
    {
      "epoch": 2.7522889114954223,
      "grad_norm": 0.8260163068771362,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.8638889789581299,
      "num_tokens": 862386155.0,
      "step": 5411
    },
    {
      "epoch": 2.7527975584944047,
      "grad_norm": 0.8747431039810181,
      "learning_rate": 1e-05,
      "loss": 0.4409,
      "mean_token_accuracy": 0.8570093512535095,
      "num_tokens": 862540675.0,
      "step": 5412
    },
    {
      "epoch": 2.7533062054933874,
      "grad_norm": 0.8887156844139099,
      "learning_rate": 1e-05,
      "loss": 0.4518,
      "mean_token_accuracy": 0.8555654287338257,
      "num_tokens": 862704936.0,
      "step": 5413
    },
    {
      "epoch": 2.7538148524923702,
      "grad_norm": 0.872695803642273,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8563168048858643,
      "num_tokens": 862862721.0,
      "step": 5414
    },
    {
      "epoch": 2.754323499491353,
      "grad_norm": 0.8667381405830383,
      "learning_rate": 1e-05,
      "loss": 0.4279,
      "mean_token_accuracy": 0.8610851764678955,
      "num_tokens": 863023723.0,
      "step": 5415
    },
    {
      "epoch": 2.754832146490336,
      "grad_norm": 0.8873748183250427,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8571150302886963,
      "num_tokens": 863189315.0,
      "step": 5416
    },
    {
      "epoch": 2.755340793489318,
      "grad_norm": 0.8880362510681152,
      "learning_rate": 1e-05,
      "loss": 0.4232,
      "mean_token_accuracy": 0.8624112606048584,
      "num_tokens": 863350818.0,
      "step": 5417
    },
    {
      "epoch": 2.7558494404883014,
      "grad_norm": 0.8671992421150208,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "mean_token_accuracy": 0.8548204898834229,
      "num_tokens": 863516158.0,
      "step": 5418
    },
    {
      "epoch": 2.7563580874872837,
      "grad_norm": 0.9104628562927246,
      "learning_rate": 1e-05,
      "loss": 0.4661,
      "mean_token_accuracy": 0.8503981232643127,
      "num_tokens": 863684595.0,
      "step": 5419
    },
    {
      "epoch": 2.7568667344862665,
      "grad_norm": 0.922897458076477,
      "learning_rate": 1e-05,
      "loss": 0.4607,
      "mean_token_accuracy": 0.8522496223449707,
      "num_tokens": 863840254.0,
      "step": 5420
    },
    {
      "epoch": 2.7573753814852493,
      "grad_norm": 0.8758529424667358,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.856182873249054,
      "num_tokens": 863991638.0,
      "step": 5421
    },
    {
      "epoch": 2.757884028484232,
      "grad_norm": 0.9188312292098999,
      "learning_rate": 1e-05,
      "loss": 0.4234,
      "mean_token_accuracy": 0.8620766401290894,
      "num_tokens": 864145411.0,
      "step": 5422
    },
    {
      "epoch": 2.758392675483215,
      "grad_norm": 0.8687071204185486,
      "learning_rate": 1e-05,
      "loss": 0.3912,
      "mean_token_accuracy": 0.8708814382553101,
      "num_tokens": 864310881.0,
      "step": 5423
    },
    {
      "epoch": 2.758901322482197,
      "grad_norm": 0.8740066289901733,
      "learning_rate": 1e-05,
      "loss": 0.406,
      "mean_token_accuracy": 0.8678805232048035,
      "num_tokens": 864463131.0,
      "step": 5424
    },
    {
      "epoch": 2.75940996948118,
      "grad_norm": 0.9356034398078918,
      "learning_rate": 1e-05,
      "loss": 0.4575,
      "mean_token_accuracy": 0.8515549898147583,
      "num_tokens": 864630170.0,
      "step": 5425
    },
    {
      "epoch": 2.759918616480163,
      "grad_norm": 0.8386475443840027,
      "learning_rate": 1e-05,
      "loss": 0.4128,
      "mean_token_accuracy": 0.864818274974823,
      "num_tokens": 864802104.0,
      "step": 5426
    },
    {
      "epoch": 2.7604272634791456,
      "grad_norm": 0.914970874786377,
      "learning_rate": 1e-05,
      "loss": 0.4729,
      "mean_token_accuracy": 0.8484303951263428,
      "num_tokens": 864970249.0,
      "step": 5427
    },
    {
      "epoch": 2.7609359104781284,
      "grad_norm": 0.8303055167198181,
      "learning_rate": 1e-05,
      "loss": 0.4338,
      "mean_token_accuracy": 0.8578987121582031,
      "num_tokens": 865146102.0,
      "step": 5428
    },
    {
      "epoch": 2.7614445574771107,
      "grad_norm": 0.8983274698257446,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8555811643600464,
      "num_tokens": 865293014.0,
      "step": 5429
    },
    {
      "epoch": 2.7619532044760935,
      "grad_norm": 0.8957774639129639,
      "learning_rate": 1e-05,
      "loss": 0.3939,
      "mean_token_accuracy": 0.870738685131073,
      "num_tokens": 865438180.0,
      "step": 5430
    },
    {
      "epoch": 2.7624618514750763,
      "grad_norm": 0.9148992896080017,
      "learning_rate": 1e-05,
      "loss": 0.4501,
      "mean_token_accuracy": 0.8549180030822754,
      "num_tokens": 865605006.0,
      "step": 5431
    },
    {
      "epoch": 2.762970498474059,
      "grad_norm": 0.9894981384277344,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8587237596511841,
      "num_tokens": 865761672.0,
      "step": 5432
    },
    {
      "epoch": 2.763479145473042,
      "grad_norm": 0.9346439242362976,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "mean_token_accuracy": 0.8564201593399048,
      "num_tokens": 865913316.0,
      "step": 5433
    },
    {
      "epoch": 2.763987792472024,
      "grad_norm": 0.9156526923179626,
      "learning_rate": 1e-05,
      "loss": 0.4479,
      "mean_token_accuracy": 0.8553595542907715,
      "num_tokens": 866080300.0,
      "step": 5434
    },
    {
      "epoch": 2.764496439471007,
      "grad_norm": 0.9552758932113647,
      "learning_rate": 1e-05,
      "loss": 0.4049,
      "mean_token_accuracy": 0.8684354424476624,
      "num_tokens": 866237570.0,
      "step": 5435
    },
    {
      "epoch": 2.7650050864699898,
      "grad_norm": 0.9023818969726562,
      "learning_rate": 1e-05,
      "loss": 0.4581,
      "mean_token_accuracy": 0.8515496253967285,
      "num_tokens": 866401658.0,
      "step": 5436
    },
    {
      "epoch": 2.7655137334689726,
      "grad_norm": 0.9802160859107971,
      "learning_rate": 1e-05,
      "loss": 0.435,
      "mean_token_accuracy": 0.8586635589599609,
      "num_tokens": 866565329.0,
      "step": 5437
    },
    {
      "epoch": 2.7660223804679553,
      "grad_norm": 0.893657386302948,
      "learning_rate": 1e-05,
      "loss": 0.4258,
      "mean_token_accuracy": 0.8619270920753479,
      "num_tokens": 866735792.0,
      "step": 5438
    },
    {
      "epoch": 2.7665310274669377,
      "grad_norm": 0.9348874688148499,
      "learning_rate": 1e-05,
      "loss": 0.4513,
      "mean_token_accuracy": 0.8542196154594421,
      "num_tokens": 866903239.0,
      "step": 5439
    },
    {
      "epoch": 2.767039674465921,
      "grad_norm": 0.8428106904029846,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8591272234916687,
      "num_tokens": 867069334.0,
      "step": 5440
    },
    {
      "epoch": 2.7675483214649033,
      "grad_norm": 0.8445457220077515,
      "learning_rate": 1e-05,
      "loss": 0.426,
      "mean_token_accuracy": 0.8615500330924988,
      "num_tokens": 867232722.0,
      "step": 5441
    },
    {
      "epoch": 2.768056968463886,
      "grad_norm": 0.9008581042289734,
      "learning_rate": 1e-05,
      "loss": 0.4579,
      "mean_token_accuracy": 0.8527448177337646,
      "num_tokens": 867389752.0,
      "step": 5442
    },
    {
      "epoch": 2.768565615462869,
      "grad_norm": 0.8951466083526611,
      "learning_rate": 1e-05,
      "loss": 0.4637,
      "mean_token_accuracy": 0.8508716821670532,
      "num_tokens": 867550520.0,
      "step": 5443
    },
    {
      "epoch": 2.7690742624618516,
      "grad_norm": 0.8982107043266296,
      "learning_rate": 1e-05,
      "loss": 0.4459,
      "mean_token_accuracy": 0.8546098470687866,
      "num_tokens": 867692997.0,
      "step": 5444
    },
    {
      "epoch": 2.7695829094608344,
      "grad_norm": 0.8363794088363647,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8605350255966187,
      "num_tokens": 867864658.0,
      "step": 5445
    },
    {
      "epoch": 2.7700915564598168,
      "grad_norm": 0.9093189835548401,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8598928451538086,
      "num_tokens": 868020290.0,
      "step": 5446
    },
    {
      "epoch": 2.7706002034587995,
      "grad_norm": 0.8726778626441956,
      "learning_rate": 1e-05,
      "loss": 0.4449,
      "mean_token_accuracy": 0.8547472953796387,
      "num_tokens": 868176893.0,
      "step": 5447
    },
    {
      "epoch": 2.7711088504577823,
      "grad_norm": 0.9353107213973999,
      "learning_rate": 1e-05,
      "loss": 0.4366,
      "mean_token_accuracy": 0.8588535785675049,
      "num_tokens": 868330610.0,
      "step": 5448
    },
    {
      "epoch": 2.771617497456765,
      "grad_norm": 0.8884314298629761,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8559328317642212,
      "num_tokens": 868486653.0,
      "step": 5449
    },
    {
      "epoch": 2.772126144455748,
      "grad_norm": 0.8668583631515503,
      "learning_rate": 1e-05,
      "loss": 0.4361,
      "mean_token_accuracy": 0.8602858185768127,
      "num_tokens": 868649173.0,
      "step": 5450
    },
    {
      "epoch": 2.7726347914547302,
      "grad_norm": 0.8272715210914612,
      "learning_rate": 1e-05,
      "loss": 0.4305,
      "mean_token_accuracy": 0.8603529930114746,
      "num_tokens": 868815495.0,
      "step": 5451
    },
    {
      "epoch": 2.773143438453713,
      "grad_norm": 0.8570767045021057,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "mean_token_accuracy": 0.8625004291534424,
      "num_tokens": 868970827.0,
      "step": 5452
    },
    {
      "epoch": 2.773652085452696,
      "grad_norm": 0.9126412272453308,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "mean_token_accuracy": 0.8636242151260376,
      "num_tokens": 869120830.0,
      "step": 5453
    },
    {
      "epoch": 2.7741607324516786,
      "grad_norm": 0.9080553650856018,
      "learning_rate": 1e-05,
      "loss": 0.4183,
      "mean_token_accuracy": 0.8638372421264648,
      "num_tokens": 869274987.0,
      "step": 5454
    },
    {
      "epoch": 2.7746693794506614,
      "grad_norm": 0.8964677453041077,
      "learning_rate": 1e-05,
      "loss": 0.4248,
      "mean_token_accuracy": 0.863158643245697,
      "num_tokens": 869423755.0,
      "step": 5455
    },
    {
      "epoch": 2.7751780264496437,
      "grad_norm": 0.9210769534111023,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8563529849052429,
      "num_tokens": 869586556.0,
      "step": 5456
    },
    {
      "epoch": 2.7756866734486265,
      "grad_norm": 0.881309449672699,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8571156859397888,
      "num_tokens": 869750823.0,
      "step": 5457
    },
    {
      "epoch": 2.7761953204476093,
      "grad_norm": 0.811819314956665,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "mean_token_accuracy": 0.85698401927948,
      "num_tokens": 869923056.0,
      "step": 5458
    },
    {
      "epoch": 2.776703967446592,
      "grad_norm": 0.8605573773384094,
      "learning_rate": 1e-05,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8466289043426514,
      "num_tokens": 870090234.0,
      "step": 5459
    },
    {
      "epoch": 2.777212614445575,
      "grad_norm": 0.8861764669418335,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8470626473426819,
      "num_tokens": 870245934.0,
      "step": 5460
    },
    {
      "epoch": 2.777721261444557,
      "grad_norm": 0.9186680316925049,
      "learning_rate": 1e-05,
      "loss": 0.407,
      "mean_token_accuracy": 0.8667311668395996,
      "num_tokens": 870397537.0,
      "step": 5461
    },
    {
      "epoch": 2.7782299084435405,
      "grad_norm": 0.8719220161437988,
      "learning_rate": 1e-05,
      "loss": 0.4368,
      "mean_token_accuracy": 0.8578290939331055,
      "num_tokens": 870553892.0,
      "step": 5462
    },
    {
      "epoch": 2.778738555442523,
      "grad_norm": 1.9770593643188477,
      "learning_rate": 1e-05,
      "loss": 0.4188,
      "mean_token_accuracy": 0.8625924587249756,
      "num_tokens": 870700935.0,
      "step": 5463
    },
    {
      "epoch": 2.7792472024415056,
      "grad_norm": 0.9255325794219971,
      "learning_rate": 1e-05,
      "loss": 0.4317,
      "mean_token_accuracy": 0.8586496114730835,
      "num_tokens": 870854822.0,
      "step": 5464
    },
    {
      "epoch": 2.7797558494404884,
      "grad_norm": 0.8715844750404358,
      "learning_rate": 1e-05,
      "loss": 0.429,
      "mean_token_accuracy": 0.8607480525970459,
      "num_tokens": 871014660.0,
      "step": 5465
    },
    {
      "epoch": 2.780264496439471,
      "grad_norm": 0.9867140054702759,
      "learning_rate": 1e-05,
      "loss": 0.4819,
      "mean_token_accuracy": 0.8451007604598999,
      "num_tokens": 871168666.0,
      "step": 5466
    },
    {
      "epoch": 2.780773143438454,
      "grad_norm": 0.9128692150115967,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8578115105628967,
      "num_tokens": 871331772.0,
      "step": 5467
    },
    {
      "epoch": 2.7812817904374363,
      "grad_norm": 0.8914899230003357,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8648374676704407,
      "num_tokens": 871494942.0,
      "step": 5468
    },
    {
      "epoch": 2.781790437436419,
      "grad_norm": 0.933683454990387,
      "learning_rate": 1e-05,
      "loss": 0.4725,
      "mean_token_accuracy": 0.8490318059921265,
      "num_tokens": 871654840.0,
      "step": 5469
    },
    {
      "epoch": 2.782299084435402,
      "grad_norm": 0.8720911145210266,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8578515648841858,
      "num_tokens": 871816010.0,
      "step": 5470
    },
    {
      "epoch": 2.7828077314343846,
      "grad_norm": 0.8903201222419739,
      "learning_rate": 1e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.8601359128952026,
      "num_tokens": 871986772.0,
      "step": 5471
    },
    {
      "epoch": 2.7833163784333674,
      "grad_norm": 0.9022795557975769,
      "learning_rate": 1e-05,
      "loss": 0.4237,
      "mean_token_accuracy": 0.8618631362915039,
      "num_tokens": 872141817.0,
      "step": 5472
    },
    {
      "epoch": 2.7838250254323498,
      "grad_norm": 0.8795859813690186,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8556370139122009,
      "num_tokens": 872295335.0,
      "step": 5473
    },
    {
      "epoch": 2.7843336724313326,
      "grad_norm": 1.0180273056030273,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "mean_token_accuracy": 0.8556839227676392,
      "num_tokens": 872455047.0,
      "step": 5474
    },
    {
      "epoch": 2.7848423194303153,
      "grad_norm": 0.8734360933303833,
      "learning_rate": 1e-05,
      "loss": 0.4585,
      "mean_token_accuracy": 0.8520325422286987,
      "num_tokens": 872622737.0,
      "step": 5475
    },
    {
      "epoch": 2.785350966429298,
      "grad_norm": 0.9091038107872009,
      "learning_rate": 1e-05,
      "loss": 0.4484,
      "mean_token_accuracy": 0.8580871820449829,
      "num_tokens": 872792380.0,
      "step": 5476
    },
    {
      "epoch": 2.785859613428281,
      "grad_norm": 0.9413979649543762,
      "learning_rate": 1e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8612833023071289,
      "num_tokens": 872943238.0,
      "step": 5477
    },
    {
      "epoch": 2.7863682604272633,
      "grad_norm": 0.9505724310874939,
      "learning_rate": 1e-05,
      "loss": 0.4652,
      "mean_token_accuracy": 0.850635290145874,
      "num_tokens": 873101516.0,
      "step": 5478
    },
    {
      "epoch": 2.786876907426246,
      "grad_norm": 0.9366916418075562,
      "learning_rate": 1e-05,
      "loss": 0.4306,
      "mean_token_accuracy": 0.8588675260543823,
      "num_tokens": 873250345.0,
      "step": 5479
    },
    {
      "epoch": 2.787385554425229,
      "grad_norm": 1.0134268999099731,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8575441837310791,
      "num_tokens": 873396170.0,
      "step": 5480
    },
    {
      "epoch": 2.7878942014242116,
      "grad_norm": 0.9557996392250061,
      "learning_rate": 1e-05,
      "loss": 0.4344,
      "mean_token_accuracy": 0.8593090772628784,
      "num_tokens": 873553305.0,
      "step": 5481
    },
    {
      "epoch": 2.7884028484231944,
      "grad_norm": 0.8980011343955994,
      "learning_rate": 1e-05,
      "loss": 0.4501,
      "mean_token_accuracy": 0.8541994094848633,
      "num_tokens": 873721507.0,
      "step": 5482
    },
    {
      "epoch": 2.7889114954221768,
      "grad_norm": 0.9143933653831482,
      "learning_rate": 1e-05,
      "loss": 0.4173,
      "mean_token_accuracy": 0.8620104789733887,
      "num_tokens": 873888032.0,
      "step": 5483
    },
    {
      "epoch": 2.78942014242116,
      "grad_norm": 0.8855127096176147,
      "learning_rate": 1e-05,
      "loss": 0.4035,
      "mean_token_accuracy": 0.8683799505233765,
      "num_tokens": 874047437.0,
      "step": 5484
    },
    {
      "epoch": 2.7899287894201423,
      "grad_norm": 0.8832699060440063,
      "learning_rate": 1e-05,
      "loss": 0.4408,
      "mean_token_accuracy": 0.8589814901351929,
      "num_tokens": 874204022.0,
      "step": 5485
    },
    {
      "epoch": 2.790437436419125,
      "grad_norm": 0.924237847328186,
      "learning_rate": 1e-05,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8617229461669922,
      "num_tokens": 874355294.0,
      "step": 5486
    },
    {
      "epoch": 2.790946083418108,
      "grad_norm": 0.9835050106048584,
      "learning_rate": 1e-05,
      "loss": 0.4451,
      "mean_token_accuracy": 0.8557791709899902,
      "num_tokens": 874515758.0,
      "step": 5487
    },
    {
      "epoch": 2.7914547304170907,
      "grad_norm": 0.8630616664886475,
      "learning_rate": 1e-05,
      "loss": 0.4172,
      "mean_token_accuracy": 0.8630630373954773,
      "num_tokens": 874686080.0,
      "step": 5488
    },
    {
      "epoch": 2.7919633774160735,
      "grad_norm": 0.9042604565620422,
      "learning_rate": 1e-05,
      "loss": 0.4324,
      "mean_token_accuracy": 0.8611968755722046,
      "num_tokens": 874840681.0,
      "step": 5489
    },
    {
      "epoch": 2.792472024415056,
      "grad_norm": 0.9088775515556335,
      "learning_rate": 1e-05,
      "loss": 0.4251,
      "mean_token_accuracy": 0.8623006343841553,
      "num_tokens": 875008971.0,
      "step": 5490
    },
    {
      "epoch": 2.7929806714140386,
      "grad_norm": 0.8844015598297119,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8596288561820984,
      "num_tokens": 875167442.0,
      "step": 5491
    },
    {
      "epoch": 2.7934893184130214,
      "grad_norm": 0.8631628751754761,
      "learning_rate": 1e-05,
      "loss": 0.4089,
      "mean_token_accuracy": 0.8667330145835876,
      "num_tokens": 875321096.0,
      "step": 5492
    },
    {
      "epoch": 2.793997965412004,
      "grad_norm": 0.9279149174690247,
      "learning_rate": 1e-05,
      "loss": 0.4317,
      "mean_token_accuracy": 0.8598275780677795,
      "num_tokens": 875481221.0,
      "step": 5493
    },
    {
      "epoch": 2.794506612410987,
      "grad_norm": 0.8831683397293091,
      "learning_rate": 1e-05,
      "loss": 0.4163,
      "mean_token_accuracy": 0.8644309043884277,
      "num_tokens": 875635412.0,
      "step": 5494
    },
    {
      "epoch": 2.7950152594099693,
      "grad_norm": 0.877562940120697,
      "learning_rate": 1e-05,
      "loss": 0.4227,
      "mean_token_accuracy": 0.8641342520713806,
      "num_tokens": 875803130.0,
      "step": 5495
    },
    {
      "epoch": 2.795523906408952,
      "grad_norm": 0.9235041737556458,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8580608367919922,
      "num_tokens": 875976489.0,
      "step": 5496
    },
    {
      "epoch": 2.796032553407935,
      "grad_norm": 0.9082411527633667,
      "learning_rate": 1e-05,
      "loss": 0.4874,
      "mean_token_accuracy": 0.8443915247917175,
      "num_tokens": 876135385.0,
      "step": 5497
    },
    {
      "epoch": 2.7965412004069177,
      "grad_norm": 0.8878147006034851,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8584445118904114,
      "num_tokens": 876298846.0,
      "step": 5498
    },
    {
      "epoch": 2.7970498474059005,
      "grad_norm": 0.8781887888908386,
      "learning_rate": 1e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8640130758285522,
      "num_tokens": 876461343.0,
      "step": 5499
    },
    {
      "epoch": 2.797558494404883,
      "grad_norm": 0.924721360206604,
      "learning_rate": 1e-05,
      "loss": 0.4667,
      "mean_token_accuracy": 0.849367082118988,
      "num_tokens": 876621222.0,
      "step": 5500
    },
    {
      "epoch": 2.7980671414038656,
      "grad_norm": 0.9024066925048828,
      "learning_rate": 1e-05,
      "loss": 0.4244,
      "mean_token_accuracy": 0.8612882494926453,
      "num_tokens": 876767008.0,
      "step": 5501
    },
    {
      "epoch": 2.7985757884028484,
      "grad_norm": 0.9149547815322876,
      "learning_rate": 1e-05,
      "loss": 0.4455,
      "mean_token_accuracy": 0.8569803237915039,
      "num_tokens": 876924869.0,
      "step": 5502
    },
    {
      "epoch": 2.799084435401831,
      "grad_norm": 0.8639458417892456,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8605494499206543,
      "num_tokens": 877084246.0,
      "step": 5503
    },
    {
      "epoch": 2.799593082400814,
      "grad_norm": 0.9297685027122498,
      "learning_rate": 1e-05,
      "loss": 0.447,
      "mean_token_accuracy": 0.8527074456214905,
      "num_tokens": 877247182.0,
      "step": 5504
    },
    {
      "epoch": 2.8001017293997963,
      "grad_norm": 0.854729413986206,
      "learning_rate": 1e-05,
      "loss": 0.4216,
      "mean_token_accuracy": 0.8616640567779541,
      "num_tokens": 877401770.0,
      "step": 5505
    },
    {
      "epoch": 2.8006103763987795,
      "grad_norm": 0.886106014251709,
      "learning_rate": 1e-05,
      "loss": 0.4153,
      "mean_token_accuracy": 0.863771915435791,
      "num_tokens": 877562351.0,
      "step": 5506
    },
    {
      "epoch": 2.801119023397762,
      "grad_norm": 1.0278208255767822,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.86051344871521,
      "num_tokens": 877717791.0,
      "step": 5507
    },
    {
      "epoch": 2.8016276703967447,
      "grad_norm": 0.8804529309272766,
      "learning_rate": 1e-05,
      "loss": 0.4322,
      "mean_token_accuracy": 0.8584067821502686,
      "num_tokens": 877878025.0,
      "step": 5508
    },
    {
      "epoch": 2.8021363173957274,
      "grad_norm": 0.9551355838775635,
      "learning_rate": 1e-05,
      "loss": 0.457,
      "mean_token_accuracy": 0.8528353571891785,
      "num_tokens": 878036046.0,
      "step": 5509
    },
    {
      "epoch": 2.8026449643947102,
      "grad_norm": 0.919446587562561,
      "learning_rate": 1e-05,
      "loss": 0.4611,
      "mean_token_accuracy": 0.8518680930137634,
      "num_tokens": 878193570.0,
      "step": 5510
    },
    {
      "epoch": 2.803153611393693,
      "grad_norm": 0.9995241165161133,
      "learning_rate": 1e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.8584878444671631,
      "num_tokens": 878356399.0,
      "step": 5511
    },
    {
      "epoch": 2.8036622583926754,
      "grad_norm": 0.9080137610435486,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8556913137435913,
      "num_tokens": 878511125.0,
      "step": 5512
    },
    {
      "epoch": 2.804170905391658,
      "grad_norm": 0.9686445593833923,
      "learning_rate": 1e-05,
      "loss": 0.3943,
      "mean_token_accuracy": 0.8702806234359741,
      "num_tokens": 878667723.0,
      "step": 5513
    },
    {
      "epoch": 2.804679552390641,
      "grad_norm": 0.9286717176437378,
      "learning_rate": 1e-05,
      "loss": 0.4105,
      "mean_token_accuracy": 0.8658961057662964,
      "num_tokens": 878816670.0,
      "step": 5514
    },
    {
      "epoch": 2.8051881993896237,
      "grad_norm": 0.9878254532814026,
      "learning_rate": 1e-05,
      "loss": 0.4449,
      "mean_token_accuracy": 0.855570375919342,
      "num_tokens": 878967803.0,
      "step": 5515
    },
    {
      "epoch": 2.8056968463886065,
      "grad_norm": 1.0075044631958008,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.8624618053436279,
      "num_tokens": 879137968.0,
      "step": 5516
    },
    {
      "epoch": 2.806205493387589,
      "grad_norm": 0.9074199199676514,
      "learning_rate": 1e-05,
      "loss": 0.4135,
      "mean_token_accuracy": 0.8656484484672546,
      "num_tokens": 879297687.0,
      "step": 5517
    },
    {
      "epoch": 2.8067141403865716,
      "grad_norm": 0.9003438353538513,
      "learning_rate": 1e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8533244132995605,
      "num_tokens": 879459289.0,
      "step": 5518
    },
    {
      "epoch": 2.8072227873855544,
      "grad_norm": 0.9956253170967102,
      "learning_rate": 1e-05,
      "loss": 0.435,
      "mean_token_accuracy": 0.8599238395690918,
      "num_tokens": 879627794.0,
      "step": 5519
    },
    {
      "epoch": 2.807731434384537,
      "grad_norm": 0.9911926984786987,
      "learning_rate": 1e-05,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8532723188400269,
      "num_tokens": 879793433.0,
      "step": 5520
    },
    {
      "epoch": 2.80824008138352,
      "grad_norm": 0.9630784392356873,
      "learning_rate": 1e-05,
      "loss": 0.4679,
      "mean_token_accuracy": 0.8500961065292358,
      "num_tokens": 879947230.0,
      "step": 5521
    },
    {
      "epoch": 2.8087487283825023,
      "grad_norm": 1.0050899982452393,
      "learning_rate": 1e-05,
      "loss": 0.4583,
      "mean_token_accuracy": 0.8508563041687012,
      "num_tokens": 880110903.0,
      "step": 5522
    },
    {
      "epoch": 2.809257375381485,
      "grad_norm": 0.9409875869750977,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8560415506362915,
      "num_tokens": 880269275.0,
      "step": 5523
    },
    {
      "epoch": 2.809766022380468,
      "grad_norm": 0.8914468288421631,
      "learning_rate": 1e-05,
      "loss": 0.3964,
      "mean_token_accuracy": 0.8705465793609619,
      "num_tokens": 880423872.0,
      "step": 5524
    },
    {
      "epoch": 2.8102746693794507,
      "grad_norm": 0.9682357311248779,
      "learning_rate": 1e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8604957461357117,
      "num_tokens": 880573676.0,
      "step": 5525
    },
    {
      "epoch": 2.8107833163784335,
      "grad_norm": 0.8917509913444519,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.8595092296600342,
      "num_tokens": 880730161.0,
      "step": 5526
    },
    {
      "epoch": 2.811291963377416,
      "grad_norm": 0.8812057375907898,
      "learning_rate": 1e-05,
      "loss": 0.4241,
      "mean_token_accuracy": 0.8612045049667358,
      "num_tokens": 880886957.0,
      "step": 5527
    },
    {
      "epoch": 2.811800610376399,
      "grad_norm": 1.008043646812439,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "mean_token_accuracy": 0.8575454354286194,
      "num_tokens": 881055356.0,
      "step": 5528
    },
    {
      "epoch": 2.8123092573753814,
      "grad_norm": 0.8801813125610352,
      "learning_rate": 1e-05,
      "loss": 0.4229,
      "mean_token_accuracy": 0.8622205257415771,
      "num_tokens": 881206094.0,
      "step": 5529
    },
    {
      "epoch": 2.812817904374364,
      "grad_norm": 0.9803150296211243,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "mean_token_accuracy": 0.8542446494102478,
      "num_tokens": 881381086.0,
      "step": 5530
    },
    {
      "epoch": 2.813326551373347,
      "grad_norm": 0.9275711178779602,
      "learning_rate": 1e-05,
      "loss": 0.4295,
      "mean_token_accuracy": 0.860144317150116,
      "num_tokens": 881543453.0,
      "step": 5531
    },
    {
      "epoch": 2.8138351983723298,
      "grad_norm": 0.9270220398902893,
      "learning_rate": 1e-05,
      "loss": 0.4209,
      "mean_token_accuracy": 0.8627274036407471,
      "num_tokens": 881697429.0,
      "step": 5532
    },
    {
      "epoch": 2.8143438453713125,
      "grad_norm": 0.9209438562393188,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8603096008300781,
      "num_tokens": 881842639.0,
      "step": 5533
    },
    {
      "epoch": 2.814852492370295,
      "grad_norm": 0.8858495354652405,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8545648455619812,
      "num_tokens": 882012327.0,
      "step": 5534
    },
    {
      "epoch": 2.8153611393692777,
      "grad_norm": 0.9024306535720825,
      "learning_rate": 1e-05,
      "loss": 0.4189,
      "mean_token_accuracy": 0.862636387348175,
      "num_tokens": 882179920.0,
      "step": 5535
    },
    {
      "epoch": 2.8158697863682605,
      "grad_norm": 0.9070698618888855,
      "learning_rate": 1e-05,
      "loss": 0.4272,
      "mean_token_accuracy": 0.8598899841308594,
      "num_tokens": 882333209.0,
      "step": 5536
    },
    {
      "epoch": 2.8163784333672433,
      "grad_norm": 0.8896937370300293,
      "learning_rate": 1e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.8578742742538452,
      "num_tokens": 882491328.0,
      "step": 5537
    },
    {
      "epoch": 2.816887080366226,
      "grad_norm": 0.9919584393501282,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8596024513244629,
      "num_tokens": 882643243.0,
      "step": 5538
    },
    {
      "epoch": 2.8173957273652084,
      "grad_norm": 0.8577784299850464,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8593538999557495,
      "num_tokens": 882803061.0,
      "step": 5539
    },
    {
      "epoch": 2.817904374364191,
      "grad_norm": 0.8803346157073975,
      "learning_rate": 1e-05,
      "loss": 0.4224,
      "mean_token_accuracy": 0.8638498783111572,
      "num_tokens": 882969232.0,
      "step": 5540
    },
    {
      "epoch": 2.818413021363174,
      "grad_norm": 0.9800869822502136,
      "learning_rate": 1e-05,
      "loss": 0.4647,
      "mean_token_accuracy": 0.8502121567726135,
      "num_tokens": 883124158.0,
      "step": 5541
    },
    {
      "epoch": 2.8189216683621567,
      "grad_norm": 0.8235204219818115,
      "learning_rate": 1e-05,
      "loss": 0.3985,
      "mean_token_accuracy": 0.8684201240539551,
      "num_tokens": 883279013.0,
      "step": 5542
    },
    {
      "epoch": 2.8194303153611395,
      "grad_norm": 0.8743343353271484,
      "learning_rate": 1e-05,
      "loss": 0.4594,
      "mean_token_accuracy": 0.8527327179908752,
      "num_tokens": 883438475.0,
      "step": 5543
    },
    {
      "epoch": 2.819938962360122,
      "grad_norm": 0.8948822021484375,
      "learning_rate": 1e-05,
      "loss": 0.4324,
      "mean_token_accuracy": 0.8579428791999817,
      "num_tokens": 883592025.0,
      "step": 5544
    },
    {
      "epoch": 2.8204476093591047,
      "grad_norm": 0.9139871597290039,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8574783802032471,
      "num_tokens": 883747762.0,
      "step": 5545
    },
    {
      "epoch": 2.8209562563580874,
      "grad_norm": 0.8890607953071594,
      "learning_rate": 1e-05,
      "loss": 0.3932,
      "mean_token_accuracy": 0.8691313862800598,
      "num_tokens": 883908818.0,
      "step": 5546
    },
    {
      "epoch": 2.8214649033570702,
      "grad_norm": 0.8738477230072021,
      "learning_rate": 1e-05,
      "loss": 0.4637,
      "mean_token_accuracy": 0.8499952554702759,
      "num_tokens": 884072795.0,
      "step": 5547
    },
    {
      "epoch": 2.821973550356053,
      "grad_norm": 0.9552207589149475,
      "learning_rate": 1e-05,
      "loss": 0.452,
      "mean_token_accuracy": 0.8557916283607483,
      "num_tokens": 884225754.0,
      "step": 5548
    },
    {
      "epoch": 2.8224821973550354,
      "grad_norm": 0.8908520340919495,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "mean_token_accuracy": 0.8545007705688477,
      "num_tokens": 884389141.0,
      "step": 5549
    },
    {
      "epoch": 2.822990844354018,
      "grad_norm": 0.8425365686416626,
      "learning_rate": 1e-05,
      "loss": 0.455,
      "mean_token_accuracy": 0.8545691967010498,
      "num_tokens": 884553353.0,
      "step": 5550
    },
    {
      "epoch": 2.823499491353001,
      "grad_norm": 0.888638436794281,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8544846773147583,
      "num_tokens": 884717764.0,
      "step": 5551
    },
    {
      "epoch": 2.8240081383519837,
      "grad_norm": 0.8891406655311584,
      "learning_rate": 1e-05,
      "loss": 0.4578,
      "mean_token_accuracy": 0.8521091938018799,
      "num_tokens": 884885565.0,
      "step": 5552
    },
    {
      "epoch": 2.8245167853509665,
      "grad_norm": 0.8493015170097351,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8548928499221802,
      "num_tokens": 885042881.0,
      "step": 5553
    },
    {
      "epoch": 2.8250254323499493,
      "grad_norm": 0.8271024823188782,
      "learning_rate": 1e-05,
      "loss": 0.4313,
      "mean_token_accuracy": 0.8589669466018677,
      "num_tokens": 885208221.0,
      "step": 5554
    },
    {
      "epoch": 2.825534079348932,
      "grad_norm": 0.896240770816803,
      "learning_rate": 1e-05,
      "loss": 0.4434,
      "mean_token_accuracy": 0.8566911816596985,
      "num_tokens": 885368977.0,
      "step": 5555
    },
    {
      "epoch": 2.8260427263479144,
      "grad_norm": 0.8497312068939209,
      "learning_rate": 1e-05,
      "loss": 0.394,
      "mean_token_accuracy": 0.8694432973861694,
      "num_tokens": 885522863.0,
      "step": 5556
    },
    {
      "epoch": 2.826551373346897,
      "grad_norm": 0.8087031841278076,
      "learning_rate": 1e-05,
      "loss": 0.4243,
      "mean_token_accuracy": 0.8616116046905518,
      "num_tokens": 885682654.0,
      "step": 5557
    },
    {
      "epoch": 2.82706002034588,
      "grad_norm": 0.8608993291854858,
      "learning_rate": 1e-05,
      "loss": 0.4482,
      "mean_token_accuracy": 0.8543695211410522,
      "num_tokens": 885850081.0,
      "step": 5558
    },
    {
      "epoch": 2.827568667344863,
      "grad_norm": 0.9493992328643799,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8574357032775879,
      "num_tokens": 886026553.0,
      "step": 5559
    },
    {
      "epoch": 2.8280773143438456,
      "grad_norm": 0.8679805994033813,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8610323667526245,
      "num_tokens": 886186789.0,
      "step": 5560
    },
    {
      "epoch": 2.828585961342828,
      "grad_norm": 0.8890089988708496,
      "learning_rate": 1e-05,
      "loss": 0.45,
      "mean_token_accuracy": 0.8530704975128174,
      "num_tokens": 886337087.0,
      "step": 5561
    },
    {
      "epoch": 2.8290946083418107,
      "grad_norm": 0.8177929520606995,
      "learning_rate": 1e-05,
      "loss": 0.43,
      "mean_token_accuracy": 0.8591690063476562,
      "num_tokens": 886496034.0,
      "step": 5562
    },
    {
      "epoch": 2.8296032553407935,
      "grad_norm": 0.8867534399032593,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "mean_token_accuracy": 0.8648470640182495,
      "num_tokens": 886666027.0,
      "step": 5563
    },
    {
      "epoch": 2.8301119023397763,
      "grad_norm": 0.9145981073379517,
      "learning_rate": 1e-05,
      "loss": 0.4471,
      "mean_token_accuracy": 0.8576685786247253,
      "num_tokens": 886817995.0,
      "step": 5564
    },
    {
      "epoch": 2.830620549338759,
      "grad_norm": 0.8816673755645752,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "mean_token_accuracy": 0.8556328415870667,
      "num_tokens": 886966338.0,
      "step": 5565
    },
    {
      "epoch": 2.8311291963377414,
      "grad_norm": 0.8810669779777527,
      "learning_rate": 1e-05,
      "loss": 0.4099,
      "mean_token_accuracy": 0.8664361834526062,
      "num_tokens": 887120555.0,
      "step": 5566
    },
    {
      "epoch": 2.831637843336724,
      "grad_norm": 0.8689042329788208,
      "learning_rate": 1e-05,
      "loss": 0.418,
      "mean_token_accuracy": 0.8618471026420593,
      "num_tokens": 887278500.0,
      "step": 5567
    },
    {
      "epoch": 2.832146490335707,
      "grad_norm": 0.8702934980392456,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "mean_token_accuracy": 0.8622481822967529,
      "num_tokens": 887434554.0,
      "step": 5568
    },
    {
      "epoch": 2.8326551373346898,
      "grad_norm": 0.83709716796875,
      "learning_rate": 1e-05,
      "loss": 0.4021,
      "mean_token_accuracy": 0.869196891784668,
      "num_tokens": 887590296.0,
      "step": 5569
    },
    {
      "epoch": 2.8331637843336726,
      "grad_norm": 0.9024631977081299,
      "learning_rate": 1e-05,
      "loss": 0.4312,
      "mean_token_accuracy": 0.8614013195037842,
      "num_tokens": 887749092.0,
      "step": 5570
    },
    {
      "epoch": 2.833672431332655,
      "grad_norm": 0.9151177406311035,
      "learning_rate": 1e-05,
      "loss": 0.4592,
      "mean_token_accuracy": 0.8518662452697754,
      "num_tokens": 887909082.0,
      "step": 5571
    },
    {
      "epoch": 2.8341810783316377,
      "grad_norm": 0.8768104910850525,
      "learning_rate": 1e-05,
      "loss": 0.4528,
      "mean_token_accuracy": 0.8537473678588867,
      "num_tokens": 888072082.0,
      "step": 5572
    },
    {
      "epoch": 2.8346897253306205,
      "grad_norm": 0.8094744682312012,
      "learning_rate": 1e-05,
      "loss": 0.4193,
      "mean_token_accuracy": 0.8633731007575989,
      "num_tokens": 888229339.0,
      "step": 5573
    },
    {
      "epoch": 2.8351983723296033,
      "grad_norm": 0.9014713764190674,
      "learning_rate": 1e-05,
      "loss": 0.4608,
      "mean_token_accuracy": 0.8511308431625366,
      "num_tokens": 888384734.0,
      "step": 5574
    },
    {
      "epoch": 2.835707019328586,
      "grad_norm": 0.8186359405517578,
      "learning_rate": 1e-05,
      "loss": 0.395,
      "mean_token_accuracy": 0.8703606128692627,
      "num_tokens": 888546458.0,
      "step": 5575
    },
    {
      "epoch": 2.836215666327569,
      "grad_norm": 0.8721616268157959,
      "learning_rate": 1e-05,
      "loss": 0.4399,
      "mean_token_accuracy": 0.8587380647659302,
      "num_tokens": 888705436.0,
      "step": 5576
    },
    {
      "epoch": 2.8367243133265516,
      "grad_norm": 0.8719232082366943,
      "learning_rate": 1e-05,
      "loss": 0.4441,
      "mean_token_accuracy": 0.8571681976318359,
      "num_tokens": 888863764.0,
      "step": 5577
    },
    {
      "epoch": 2.837232960325534,
      "grad_norm": 0.8821418881416321,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8637368083000183,
      "num_tokens": 889025194.0,
      "step": 5578
    },
    {
      "epoch": 2.8377416073245167,
      "grad_norm": 0.9229016304016113,
      "learning_rate": 1e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8599576950073242,
      "num_tokens": 889182518.0,
      "step": 5579
    },
    {
      "epoch": 2.8382502543234995,
      "grad_norm": 0.9832534790039062,
      "learning_rate": 1e-05,
      "loss": 0.4583,
      "mean_token_accuracy": 0.8509209156036377,
      "num_tokens": 889348596.0,
      "step": 5580
    },
    {
      "epoch": 2.8387589013224823,
      "grad_norm": 0.8540241122245789,
      "learning_rate": 1e-05,
      "loss": 0.404,
      "mean_token_accuracy": 0.8679183125495911,
      "num_tokens": 889511461.0,
      "step": 5581
    },
    {
      "epoch": 2.839267548321465,
      "grad_norm": 0.912282407283783,
      "learning_rate": 1e-05,
      "loss": 0.4606,
      "mean_token_accuracy": 0.852907657623291,
      "num_tokens": 889676197.0,
      "step": 5582
    },
    {
      "epoch": 2.8397761953204474,
      "grad_norm": 0.8290195465087891,
      "learning_rate": 1e-05,
      "loss": 0.439,
      "mean_token_accuracy": 0.8586235046386719,
      "num_tokens": 889843485.0,
      "step": 5583
    },
    {
      "epoch": 2.8402848423194302,
      "grad_norm": 1.198011040687561,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8564248085021973,
      "num_tokens": 890005117.0,
      "step": 5584
    },
    {
      "epoch": 2.840793489318413,
      "grad_norm": 0.8554531335830688,
      "learning_rate": 1e-05,
      "loss": 0.4494,
      "mean_token_accuracy": 0.8548786044120789,
      "num_tokens": 890158166.0,
      "step": 5585
    },
    {
      "epoch": 2.841302136317396,
      "grad_norm": 0.8591538667678833,
      "learning_rate": 1e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8624992370605469,
      "num_tokens": 890316760.0,
      "step": 5586
    },
    {
      "epoch": 2.8418107833163786,
      "grad_norm": 0.9349908232688904,
      "learning_rate": 1e-05,
      "loss": 0.4285,
      "mean_token_accuracy": 0.8602296113967896,
      "num_tokens": 890465667.0,
      "step": 5587
    },
    {
      "epoch": 2.842319430315361,
      "grad_norm": 0.8721975088119507,
      "learning_rate": 1e-05,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8565325736999512,
      "num_tokens": 890620871.0,
      "step": 5588
    },
    {
      "epoch": 2.8428280773143437,
      "grad_norm": 0.940648078918457,
      "learning_rate": 1e-05,
      "loss": 0.4604,
      "mean_token_accuracy": 0.8533288836479187,
      "num_tokens": 890780390.0,
      "step": 5589
    },
    {
      "epoch": 2.8433367243133265,
      "grad_norm": 0.8890752792358398,
      "learning_rate": 1e-05,
      "loss": 0.3947,
      "mean_token_accuracy": 0.8674182295799255,
      "num_tokens": 890936449.0,
      "step": 5590
    },
    {
      "epoch": 2.8438453713123093,
      "grad_norm": 0.8123888373374939,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8599176406860352,
      "num_tokens": 891102174.0,
      "step": 5591
    },
    {
      "epoch": 2.844354018311292,
      "grad_norm": 0.954985499382019,
      "learning_rate": 1e-05,
      "loss": 0.438,
      "mean_token_accuracy": 0.8574029207229614,
      "num_tokens": 891269034.0,
      "step": 5592
    },
    {
      "epoch": 2.8448626653102744,
      "grad_norm": 0.8957969546318054,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "mean_token_accuracy": 0.8660489320755005,
      "num_tokens": 891427523.0,
      "step": 5593
    },
    {
      "epoch": 2.845371312309257,
      "grad_norm": 1.1233398914337158,
      "learning_rate": 1e-05,
      "loss": 0.4481,
      "mean_token_accuracy": 0.8538949489593506,
      "num_tokens": 891601141.0,
      "step": 5594
    },
    {
      "epoch": 2.84587995930824,
      "grad_norm": 0.8899928331375122,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "mean_token_accuracy": 0.8564096689224243,
      "num_tokens": 891758193.0,
      "step": 5595
    },
    {
      "epoch": 2.846388606307223,
      "grad_norm": 0.9822993874549866,
      "learning_rate": 1e-05,
      "loss": 0.4715,
      "mean_token_accuracy": 0.8498365879058838,
      "num_tokens": 891911699.0,
      "step": 5596
    },
    {
      "epoch": 2.8468972533062056,
      "grad_norm": 0.9196913242340088,
      "learning_rate": 1e-05,
      "loss": 0.4416,
      "mean_token_accuracy": 0.8551076650619507,
      "num_tokens": 892079968.0,
      "step": 5597
    },
    {
      "epoch": 2.8474059003051884,
      "grad_norm": 0.8975830078125,
      "learning_rate": 1e-05,
      "loss": 0.4346,
      "mean_token_accuracy": 0.8593940734863281,
      "num_tokens": 892236979.0,
      "step": 5598
    },
    {
      "epoch": 2.847914547304171,
      "grad_norm": 0.9877001643180847,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "mean_token_accuracy": 0.8644639849662781,
      "num_tokens": 892387456.0,
      "step": 5599
    },
    {
      "epoch": 2.8484231943031535,
      "grad_norm": 0.9261271357536316,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8556768298149109,
      "num_tokens": 892536633.0,
      "step": 5600
    },
    {
      "epoch": 2.8489318413021363,
      "grad_norm": 0.9050893187522888,
      "learning_rate": 1e-05,
      "loss": 0.4141,
      "mean_token_accuracy": 0.864107072353363,
      "num_tokens": 892699765.0,
      "step": 5601
    },
    {
      "epoch": 2.849440488301119,
      "grad_norm": 0.899746835231781,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8585728406906128,
      "num_tokens": 892862170.0,
      "step": 5602
    },
    {
      "epoch": 2.849949135300102,
      "grad_norm": 0.9267527461051941,
      "learning_rate": 1e-05,
      "loss": 0.4537,
      "mean_token_accuracy": 0.850830614566803,
      "num_tokens": 893017234.0,
      "step": 5603
    },
    {
      "epoch": 2.8504577822990846,
      "grad_norm": 1.0719654560089111,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8597113490104675,
      "num_tokens": 893170520.0,
      "step": 5604
    },
    {
      "epoch": 2.850966429298067,
      "grad_norm": 0.912548840045929,
      "learning_rate": 1e-05,
      "loss": 0.4171,
      "mean_token_accuracy": 0.8634018301963806,
      "num_tokens": 893319754.0,
      "step": 5605
    },
    {
      "epoch": 2.8514750762970498,
      "grad_norm": 0.9132776260375977,
      "learning_rate": 1e-05,
      "loss": 0.4662,
      "mean_token_accuracy": 0.8516316413879395,
      "num_tokens": 893480701.0,
      "step": 5606
    },
    {
      "epoch": 2.8519837232960326,
      "grad_norm": 0.9201282262802124,
      "learning_rate": 1e-05,
      "loss": 0.4529,
      "mean_token_accuracy": 0.8552930355072021,
      "num_tokens": 893634155.0,
      "step": 5607
    },
    {
      "epoch": 2.8524923702950153,
      "grad_norm": 0.8679859042167664,
      "learning_rate": 1e-05,
      "loss": 0.469,
      "mean_token_accuracy": 0.8502346277236938,
      "num_tokens": 893795202.0,
      "step": 5608
    },
    {
      "epoch": 2.853001017293998,
      "grad_norm": 0.9941920638084412,
      "learning_rate": 1e-05,
      "loss": 0.4047,
      "mean_token_accuracy": 0.867050290107727,
      "num_tokens": 893943577.0,
      "step": 5609
    },
    {
      "epoch": 2.8535096642929805,
      "grad_norm": 0.8621466755867004,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "mean_token_accuracy": 0.8550496697425842,
      "num_tokens": 894108299.0,
      "step": 5610
    },
    {
      "epoch": 2.8540183112919633,
      "grad_norm": 0.9322388768196106,
      "learning_rate": 1e-05,
      "loss": 0.4418,
      "mean_token_accuracy": 0.8577138185501099,
      "num_tokens": 894259102.0,
      "step": 5611
    },
    {
      "epoch": 2.854526958290946,
      "grad_norm": 0.909308135509491,
      "learning_rate": 1e-05,
      "loss": 0.457,
      "mean_token_accuracy": 0.8523279428482056,
      "num_tokens": 894419933.0,
      "step": 5612
    },
    {
      "epoch": 2.855035605289929,
      "grad_norm": 0.8565107583999634,
      "learning_rate": 1e-05,
      "loss": 0.417,
      "mean_token_accuracy": 0.8655107021331787,
      "num_tokens": 894576241.0,
      "step": 5613
    },
    {
      "epoch": 2.8555442522889116,
      "grad_norm": 0.9509232640266418,
      "learning_rate": 1e-05,
      "loss": 0.4517,
      "mean_token_accuracy": 0.8545820713043213,
      "num_tokens": 894741124.0,
      "step": 5614
    },
    {
      "epoch": 2.856052899287894,
      "grad_norm": 1.0113739967346191,
      "learning_rate": 1e-05,
      "loss": 0.4472,
      "mean_token_accuracy": 0.857478141784668,
      "num_tokens": 894904122.0,
      "step": 5615
    },
    {
      "epoch": 2.8565615462868768,
      "grad_norm": 1.0040079355239868,
      "learning_rate": 1e-05,
      "loss": 0.473,
      "mean_token_accuracy": 0.8473595380783081,
      "num_tokens": 895056647.0,
      "step": 5616
    },
    {
      "epoch": 2.8570701932858595,
      "grad_norm": 0.8770785331726074,
      "learning_rate": 1e-05,
      "loss": 0.4513,
      "mean_token_accuracy": 0.8533390760421753,
      "num_tokens": 895213822.0,
      "step": 5617
    },
    {
      "epoch": 2.8575788402848423,
      "grad_norm": 0.8685998320579529,
      "learning_rate": 1e-05,
      "loss": 0.4817,
      "mean_token_accuracy": 0.8471949100494385,
      "num_tokens": 895380518.0,
      "step": 5618
    },
    {
      "epoch": 2.858087487283825,
      "grad_norm": 0.94172203540802,
      "learning_rate": 1e-05,
      "loss": 0.4364,
      "mean_token_accuracy": 0.8586889505386353,
      "num_tokens": 895544268.0,
      "step": 5619
    },
    {
      "epoch": 2.8585961342828075,
      "grad_norm": 0.8902906775474548,
      "learning_rate": 1e-05,
      "loss": 0.4298,
      "mean_token_accuracy": 0.8607203960418701,
      "num_tokens": 895702008.0,
      "step": 5620
    },
    {
      "epoch": 2.8591047812817907,
      "grad_norm": 0.850885272026062,
      "learning_rate": 1e-05,
      "loss": 0.4,
      "mean_token_accuracy": 0.8698835372924805,
      "num_tokens": 895851412.0,
      "step": 5621
    },
    {
      "epoch": 2.859613428280773,
      "grad_norm": 1.0017894506454468,
      "learning_rate": 1e-05,
      "loss": 0.4209,
      "mean_token_accuracy": 0.8614801168441772,
      "num_tokens": 896011546.0,
      "step": 5622
    },
    {
      "epoch": 2.860122075279756,
      "grad_norm": 1.015999674797058,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.8582688570022583,
      "num_tokens": 896162494.0,
      "step": 5623
    },
    {
      "epoch": 2.8606307222787386,
      "grad_norm": 0.923900306224823,
      "learning_rate": 1e-05,
      "loss": 0.4187,
      "mean_token_accuracy": 0.864201545715332,
      "num_tokens": 896329246.0,
      "step": 5624
    },
    {
      "epoch": 2.8611393692777214,
      "grad_norm": 0.9373534321784973,
      "learning_rate": 1e-05,
      "loss": 0.4025,
      "mean_token_accuracy": 0.868912935256958,
      "num_tokens": 896485456.0,
      "step": 5625
    },
    {
      "epoch": 2.861648016276704,
      "grad_norm": 0.8971603512763977,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8586189150810242,
      "num_tokens": 896652517.0,
      "step": 5626
    },
    {
      "epoch": 2.8621566632756865,
      "grad_norm": 0.9007195234298706,
      "learning_rate": 1e-05,
      "loss": 0.4288,
      "mean_token_accuracy": 0.859435498714447,
      "num_tokens": 896821016.0,
      "step": 5627
    },
    {
      "epoch": 2.8626653102746693,
      "grad_norm": 0.9142847657203674,
      "learning_rate": 1e-05,
      "loss": 0.4184,
      "mean_token_accuracy": 0.8641673922538757,
      "num_tokens": 896970551.0,
      "step": 5628
    },
    {
      "epoch": 2.863173957273652,
      "grad_norm": 0.8780483603477478,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8598891496658325,
      "num_tokens": 897136957.0,
      "step": 5629
    },
    {
      "epoch": 2.863682604272635,
      "grad_norm": 0.8631423711776733,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8595150709152222,
      "num_tokens": 897291892.0,
      "step": 5630
    },
    {
      "epoch": 2.8641912512716177,
      "grad_norm": 0.8913904428482056,
      "learning_rate": 1e-05,
      "loss": 0.4177,
      "mean_token_accuracy": 0.8650398850440979,
      "num_tokens": 897452981.0,
      "step": 5631
    },
    {
      "epoch": 2.8646998982706,
      "grad_norm": 0.9493880867958069,
      "learning_rate": 1e-05,
      "loss": 0.4601,
      "mean_token_accuracy": 0.8521661162376404,
      "num_tokens": 897605769.0,
      "step": 5632
    },
    {
      "epoch": 2.865208545269583,
      "grad_norm": 0.8981471657752991,
      "learning_rate": 1e-05,
      "loss": 0.414,
      "mean_token_accuracy": 0.8634394407272339,
      "num_tokens": 897760281.0,
      "step": 5633
    },
    {
      "epoch": 2.8657171922685656,
      "grad_norm": 0.8748188018798828,
      "learning_rate": 1e-05,
      "loss": 0.4279,
      "mean_token_accuracy": 0.8613673448562622,
      "num_tokens": 897914249.0,
      "step": 5634
    },
    {
      "epoch": 2.8662258392675484,
      "grad_norm": 0.8937290906906128,
      "learning_rate": 1e-05,
      "loss": 0.4369,
      "mean_token_accuracy": 0.8586488962173462,
      "num_tokens": 898064141.0,
      "step": 5635
    },
    {
      "epoch": 2.866734486266531,
      "grad_norm": 0.8642660975456238,
      "learning_rate": 1e-05,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8620635271072388,
      "num_tokens": 898228863.0,
      "step": 5636
    },
    {
      "epoch": 2.8672431332655135,
      "grad_norm": 0.9388706088066101,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.8560210466384888,
      "num_tokens": 898373794.0,
      "step": 5637
    },
    {
      "epoch": 2.8677517802644963,
      "grad_norm": 0.8486452698707581,
      "learning_rate": 1e-05,
      "loss": 0.4155,
      "mean_token_accuracy": 0.8632423281669617,
      "num_tokens": 898530951.0,
      "step": 5638
    },
    {
      "epoch": 2.868260427263479,
      "grad_norm": 0.8945029377937317,
      "learning_rate": 1e-05,
      "loss": 0.4494,
      "mean_token_accuracy": 0.8540161848068237,
      "num_tokens": 898691706.0,
      "step": 5639
    },
    {
      "epoch": 2.868769074262462,
      "grad_norm": 0.8432234525680542,
      "learning_rate": 1e-05,
      "loss": 0.414,
      "mean_token_accuracy": 0.8628842830657959,
      "num_tokens": 898851712.0,
      "step": 5640
    },
    {
      "epoch": 2.8692777212614446,
      "grad_norm": 0.9244741201400757,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "mean_token_accuracy": 0.8574188351631165,
      "num_tokens": 899015049.0,
      "step": 5641
    },
    {
      "epoch": 2.869786368260427,
      "grad_norm": 0.8768820762634277,
      "learning_rate": 1e-05,
      "loss": 0.4311,
      "mean_token_accuracy": 0.8579252362251282,
      "num_tokens": 899173646.0,
      "step": 5642
    },
    {
      "epoch": 2.87029501525941,
      "grad_norm": 0.8740848898887634,
      "learning_rate": 1e-05,
      "loss": 0.4339,
      "mean_token_accuracy": 0.8589560389518738,
      "num_tokens": 899340664.0,
      "step": 5643
    },
    {
      "epoch": 2.8708036622583926,
      "grad_norm": 0.8888913989067078,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8611010313034058,
      "num_tokens": 899486743.0,
      "step": 5644
    },
    {
      "epoch": 2.8713123092573754,
      "grad_norm": 0.8741505146026611,
      "learning_rate": 1e-05,
      "loss": 0.4353,
      "mean_token_accuracy": 0.8584800958633423,
      "num_tokens": 899653132.0,
      "step": 5645
    },
    {
      "epoch": 2.871820956256358,
      "grad_norm": 0.9070521593093872,
      "learning_rate": 1e-05,
      "loss": 0.4266,
      "mean_token_accuracy": 0.8619548678398132,
      "num_tokens": 899809627.0,
      "step": 5646
    },
    {
      "epoch": 2.872329603255341,
      "grad_norm": 0.8843440413475037,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8559424877166748,
      "num_tokens": 899975130.0,
      "step": 5647
    },
    {
      "epoch": 2.8728382502543237,
      "grad_norm": 0.8765766620635986,
      "learning_rate": 1e-05,
      "loss": 0.432,
      "mean_token_accuracy": 0.8621829152107239,
      "num_tokens": 900137262.0,
      "step": 5648
    },
    {
      "epoch": 2.873346897253306,
      "grad_norm": 0.9020034074783325,
      "learning_rate": 1e-05,
      "loss": 0.4698,
      "mean_token_accuracy": 0.8482083082199097,
      "num_tokens": 900291517.0,
      "step": 5649
    },
    {
      "epoch": 2.873855544252289,
      "grad_norm": 0.9273054003715515,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.8615149855613708,
      "num_tokens": 900459708.0,
      "step": 5650
    },
    {
      "epoch": 2.8743641912512716,
      "grad_norm": 0.8844437599182129,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8597338199615479,
      "num_tokens": 900612066.0,
      "step": 5651
    },
    {
      "epoch": 2.8748728382502544,
      "grad_norm": 0.8729273676872253,
      "learning_rate": 1e-05,
      "loss": 0.4494,
      "mean_token_accuracy": 0.8557407855987549,
      "num_tokens": 900777165.0,
      "step": 5652
    },
    {
      "epoch": 2.875381485249237,
      "grad_norm": 0.8966065049171448,
      "learning_rate": 1e-05,
      "loss": 0.4195,
      "mean_token_accuracy": 0.8638300895690918,
      "num_tokens": 900940416.0,
      "step": 5653
    },
    {
      "epoch": 2.8758901322482195,
      "grad_norm": 0.8397489786148071,
      "learning_rate": 1e-05,
      "loss": 0.4217,
      "mean_token_accuracy": 0.8620539903640747,
      "num_tokens": 901100446.0,
      "step": 5654
    },
    {
      "epoch": 2.8763987792472023,
      "grad_norm": 0.9159259796142578,
      "learning_rate": 1e-05,
      "loss": 0.4203,
      "mean_token_accuracy": 0.8628263473510742,
      "num_tokens": 901251879.0,
      "step": 5655
    },
    {
      "epoch": 2.876907426246185,
      "grad_norm": 0.8920116424560547,
      "learning_rate": 1e-05,
      "loss": 0.4114,
      "mean_token_accuracy": 0.8661978840827942,
      "num_tokens": 901413081.0,
      "step": 5656
    },
    {
      "epoch": 2.877416073245168,
      "grad_norm": 0.8771112561225891,
      "learning_rate": 1e-05,
      "loss": 0.4109,
      "mean_token_accuracy": 0.865467369556427,
      "num_tokens": 901575779.0,
      "step": 5657
    },
    {
      "epoch": 2.8779247202441507,
      "grad_norm": 0.9231264591217041,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "mean_token_accuracy": 0.857517421245575,
      "num_tokens": 901726168.0,
      "step": 5658
    },
    {
      "epoch": 2.878433367243133,
      "grad_norm": 0.9495973587036133,
      "learning_rate": 1e-05,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8618420362472534,
      "num_tokens": 901879638.0,
      "step": 5659
    },
    {
      "epoch": 2.878942014242116,
      "grad_norm": 0.9396060109138489,
      "learning_rate": 1e-05,
      "loss": 0.4537,
      "mean_token_accuracy": 0.8520980477333069,
      "num_tokens": 902035368.0,
      "step": 5660
    },
    {
      "epoch": 2.8794506612410986,
      "grad_norm": 0.9062393307685852,
      "learning_rate": 1e-05,
      "loss": 0.4219,
      "mean_token_accuracy": 0.8610861301422119,
      "num_tokens": 902198864.0,
      "step": 5661
    },
    {
      "epoch": 2.8799593082400814,
      "grad_norm": 0.8732383847236633,
      "learning_rate": 1e-05,
      "loss": 0.4285,
      "mean_token_accuracy": 0.860880970954895,
      "num_tokens": 902361599.0,
      "step": 5662
    },
    {
      "epoch": 2.880467955239064,
      "grad_norm": 0.9061823487281799,
      "learning_rate": 1e-05,
      "loss": 0.4451,
      "mean_token_accuracy": 0.8560501933097839,
      "num_tokens": 902520237.0,
      "step": 5663
    },
    {
      "epoch": 2.8809766022380465,
      "grad_norm": 0.9862648844718933,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "mean_token_accuracy": 0.8644874095916748,
      "num_tokens": 902671212.0,
      "step": 5664
    },
    {
      "epoch": 2.8814852492370298,
      "grad_norm": 0.8495729565620422,
      "learning_rate": 1e-05,
      "loss": 0.456,
      "mean_token_accuracy": 0.8526976108551025,
      "num_tokens": 902828552.0,
      "step": 5665
    },
    {
      "epoch": 2.881993896236012,
      "grad_norm": 0.8538773655891418,
      "learning_rate": 1e-05,
      "loss": 0.4511,
      "mean_token_accuracy": 0.8544628024101257,
      "num_tokens": 902987411.0,
      "step": 5666
    },
    {
      "epoch": 2.882502543234995,
      "grad_norm": 0.9478979706764221,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "mean_token_accuracy": 0.8565550446510315,
      "num_tokens": 903143437.0,
      "step": 5667
    },
    {
      "epoch": 2.8830111902339777,
      "grad_norm": 0.8719910979270935,
      "learning_rate": 1e-05,
      "loss": 0.4433,
      "mean_token_accuracy": 0.8571025729179382,
      "num_tokens": 903308282.0,
      "step": 5668
    },
    {
      "epoch": 2.8835198372329605,
      "grad_norm": 0.9409666657447815,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "mean_token_accuracy": 0.8566873669624329,
      "num_tokens": 903468922.0,
      "step": 5669
    },
    {
      "epoch": 2.8840284842319432,
      "grad_norm": 0.8633664846420288,
      "learning_rate": 1e-05,
      "loss": 0.431,
      "mean_token_accuracy": 0.8594985008239746,
      "num_tokens": 903628314.0,
      "step": 5670
    },
    {
      "epoch": 2.8845371312309256,
      "grad_norm": 0.8893129229545593,
      "learning_rate": 1e-05,
      "loss": 0.4282,
      "mean_token_accuracy": 0.8600053787231445,
      "num_tokens": 903785061.0,
      "step": 5671
    },
    {
      "epoch": 2.8850457782299084,
      "grad_norm": 0.9648041725158691,
      "learning_rate": 1e-05,
      "loss": 0.4296,
      "mean_token_accuracy": 0.8607710599899292,
      "num_tokens": 903947269.0,
      "step": 5672
    },
    {
      "epoch": 2.885554425228891,
      "grad_norm": 0.9494853615760803,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8642484545707703,
      "num_tokens": 904095166.0,
      "step": 5673
    },
    {
      "epoch": 2.886063072227874,
      "grad_norm": 0.8843656778335571,
      "learning_rate": 1e-05,
      "loss": 0.4657,
      "mean_token_accuracy": 0.851182222366333,
      "num_tokens": 904246591.0,
      "step": 5674
    },
    {
      "epoch": 2.8865717192268567,
      "grad_norm": 0.9313526153564453,
      "learning_rate": 1e-05,
      "loss": 0.433,
      "mean_token_accuracy": 0.8591406345367432,
      "num_tokens": 904404768.0,
      "step": 5675
    },
    {
      "epoch": 2.887080366225839,
      "grad_norm": 0.9152849316596985,
      "learning_rate": 1e-05,
      "loss": 0.442,
      "mean_token_accuracy": 0.8567684888839722,
      "num_tokens": 904565888.0,
      "step": 5676
    },
    {
      "epoch": 2.887589013224822,
      "grad_norm": 0.923521876335144,
      "learning_rate": 1e-05,
      "loss": 0.4595,
      "mean_token_accuracy": 0.8506315350532532,
      "num_tokens": 904722127.0,
      "step": 5677
    },
    {
      "epoch": 2.8880976602238047,
      "grad_norm": 0.9015587568283081,
      "learning_rate": 1e-05,
      "loss": 0.4542,
      "mean_token_accuracy": 0.8540414571762085,
      "num_tokens": 904888197.0,
      "step": 5678
    },
    {
      "epoch": 2.8886063072227874,
      "grad_norm": 0.8459432721138,
      "learning_rate": 1e-05,
      "loss": 0.4505,
      "mean_token_accuracy": 0.8547311425209045,
      "num_tokens": 905057713.0,
      "step": 5679
    },
    {
      "epoch": 2.8891149542217702,
      "grad_norm": 0.8662464618682861,
      "learning_rate": 1e-05,
      "loss": 0.4375,
      "mean_token_accuracy": 0.8573547601699829,
      "num_tokens": 905218458.0,
      "step": 5680
    },
    {
      "epoch": 2.8896236012207526,
      "grad_norm": 0.9608275294303894,
      "learning_rate": 1e-05,
      "loss": 0.4451,
      "mean_token_accuracy": 0.8553639650344849,
      "num_tokens": 905368564.0,
      "step": 5681
    },
    {
      "epoch": 2.8901322482197354,
      "grad_norm": 0.8665888905525208,
      "learning_rate": 1e-05,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8467645049095154,
      "num_tokens": 905543712.0,
      "step": 5682
    },
    {
      "epoch": 2.890640895218718,
      "grad_norm": 0.9251812696456909,
      "learning_rate": 1e-05,
      "loss": 0.4439,
      "mean_token_accuracy": 0.8561020493507385,
      "num_tokens": 905691310.0,
      "step": 5683
    },
    {
      "epoch": 2.891149542217701,
      "grad_norm": 0.9309903979301453,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.861987829208374,
      "num_tokens": 905848988.0,
      "step": 5684
    },
    {
      "epoch": 2.8916581892166837,
      "grad_norm": 0.8957622647285461,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8643568754196167,
      "num_tokens": 906001421.0,
      "step": 5685
    },
    {
      "epoch": 2.892166836215666,
      "grad_norm": 0.960602343082428,
      "learning_rate": 1e-05,
      "loss": 0.4615,
      "mean_token_accuracy": 0.8526604175567627,
      "num_tokens": 906155224.0,
      "step": 5686
    },
    {
      "epoch": 2.8926754832146493,
      "grad_norm": 0.9190289974212646,
      "learning_rate": 1e-05,
      "loss": 0.4316,
      "mean_token_accuracy": 0.8602994680404663,
      "num_tokens": 906310194.0,
      "step": 5687
    },
    {
      "epoch": 2.8931841302136316,
      "grad_norm": 0.9160748720169067,
      "learning_rate": 1e-05,
      "loss": 0.4681,
      "mean_token_accuracy": 0.8495457768440247,
      "num_tokens": 906474923.0,
      "step": 5688
    },
    {
      "epoch": 2.8936927772126144,
      "grad_norm": 0.9127035140991211,
      "learning_rate": 1e-05,
      "loss": 0.3988,
      "mean_token_accuracy": 0.8672407269477844,
      "num_tokens": 906624664.0,
      "step": 5689
    },
    {
      "epoch": 2.894201424211597,
      "grad_norm": 0.8942268490791321,
      "learning_rate": 1e-05,
      "loss": 0.4532,
      "mean_token_accuracy": 0.854296088218689,
      "num_tokens": 906788218.0,
      "step": 5690
    },
    {
      "epoch": 2.89471007121058,
      "grad_norm": 0.9423340559005737,
      "learning_rate": 1e-05,
      "loss": 0.4359,
      "mean_token_accuracy": 0.8581220507621765,
      "num_tokens": 906934206.0,
      "step": 5691
    },
    {
      "epoch": 2.895218718209563,
      "grad_norm": 0.8928824663162231,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.854724109172821,
      "num_tokens": 907091206.0,
      "step": 5692
    },
    {
      "epoch": 2.895727365208545,
      "grad_norm": 0.8230746388435364,
      "learning_rate": 1e-05,
      "loss": 0.4251,
      "mean_token_accuracy": 0.8614833354949951,
      "num_tokens": 907264336.0,
      "step": 5693
    },
    {
      "epoch": 2.896236012207528,
      "grad_norm": 0.9321528077125549,
      "learning_rate": 1e-05,
      "loss": 0.446,
      "mean_token_accuracy": 0.8563940525054932,
      "num_tokens": 907426047.0,
      "step": 5694
    },
    {
      "epoch": 2.8967446592065107,
      "grad_norm": 0.8536028861999512,
      "learning_rate": 1e-05,
      "loss": 0.4308,
      "mean_token_accuracy": 0.8602784872055054,
      "num_tokens": 907587211.0,
      "step": 5695
    },
    {
      "epoch": 2.8972533062054935,
      "grad_norm": 0.8194431066513062,
      "learning_rate": 1e-05,
      "loss": 0.4351,
      "mean_token_accuracy": 0.859073281288147,
      "num_tokens": 907753455.0,
      "step": 5696
    },
    {
      "epoch": 2.8977619532044763,
      "grad_norm": 0.8605436682701111,
      "learning_rate": 1e-05,
      "loss": 0.3989,
      "mean_token_accuracy": 0.8686566948890686,
      "num_tokens": 907908915.0,
      "step": 5697
    },
    {
      "epoch": 2.8982706002034586,
      "grad_norm": 1.587167739868164,
      "learning_rate": 1e-05,
      "loss": 0.4312,
      "mean_token_accuracy": 0.8598580360412598,
      "num_tokens": 908071238.0,
      "step": 5698
    },
    {
      "epoch": 2.8987792472024414,
      "grad_norm": 0.918057918548584,
      "learning_rate": 1e-05,
      "loss": 0.4259,
      "mean_token_accuracy": 0.8622524738311768,
      "num_tokens": 908221402.0,
      "step": 5699
    },
    {
      "epoch": 2.899287894201424,
      "grad_norm": 0.9031367897987366,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8568249344825745,
      "num_tokens": 908384365.0,
      "step": 5700
    },
    {
      "epoch": 2.899796541200407,
      "grad_norm": 0.8905377984046936,
      "learning_rate": 1e-05,
      "loss": 0.4137,
      "mean_token_accuracy": 0.8642496466636658,
      "num_tokens": 908531089.0,
      "step": 5701
    },
    {
      "epoch": 2.9003051881993898,
      "grad_norm": 0.9216523766517639,
      "learning_rate": 1e-05,
      "loss": 0.4417,
      "mean_token_accuracy": 0.8555152416229248,
      "num_tokens": 908692929.0,
      "step": 5702
    },
    {
      "epoch": 2.900813835198372,
      "grad_norm": 0.8346421718597412,
      "learning_rate": 1e-05,
      "loss": 0.4149,
      "mean_token_accuracy": 0.8649652004241943,
      "num_tokens": 908848433.0,
      "step": 5703
    },
    {
      "epoch": 2.901322482197355,
      "grad_norm": 0.8686506748199463,
      "learning_rate": 1e-05,
      "loss": 0.437,
      "mean_token_accuracy": 0.858033299446106,
      "num_tokens": 909009416.0,
      "step": 5704
    },
    {
      "epoch": 2.9018311291963377,
      "grad_norm": 0.8234359622001648,
      "learning_rate": 1e-05,
      "loss": 0.4011,
      "mean_token_accuracy": 0.8687787055969238,
      "num_tokens": 909166803.0,
      "step": 5705
    },
    {
      "epoch": 2.9023397761953205,
      "grad_norm": 0.8834549784660339,
      "learning_rate": 1e-05,
      "loss": 0.4207,
      "mean_token_accuracy": 0.8631112575531006,
      "num_tokens": 909326555.0,
      "step": 5706
    },
    {
      "epoch": 2.9028484231943033,
      "grad_norm": 0.8309847712516785,
      "learning_rate": 1e-05,
      "loss": 0.4092,
      "mean_token_accuracy": 0.8655686974525452,
      "num_tokens": 909484045.0,
      "step": 5707
    },
    {
      "epoch": 2.9033570701932856,
      "grad_norm": 0.8650888800621033,
      "learning_rate": 1e-05,
      "loss": 0.4233,
      "mean_token_accuracy": 0.8623444437980652,
      "num_tokens": 909639003.0,
      "step": 5708
    },
    {
      "epoch": 2.903865717192269,
      "grad_norm": 0.891726016998291,
      "learning_rate": 1e-05,
      "loss": 0.4534,
      "mean_token_accuracy": 0.8554078936576843,
      "num_tokens": 909799087.0,
      "step": 5709
    },
    {
      "epoch": 2.904374364191251,
      "grad_norm": 0.8960121870040894,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8592454791069031,
      "num_tokens": 909943894.0,
      "step": 5710
    },
    {
      "epoch": 2.904883011190234,
      "grad_norm": 0.9292611479759216,
      "learning_rate": 1e-05,
      "loss": 0.4558,
      "mean_token_accuracy": 0.8520259261131287,
      "num_tokens": 910111511.0,
      "step": 5711
    },
    {
      "epoch": 2.9053916581892167,
      "grad_norm": 0.7885063886642456,
      "learning_rate": 1e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8613859415054321,
      "num_tokens": 910282610.0,
      "step": 5712
    },
    {
      "epoch": 2.9059003051881995,
      "grad_norm": 0.8445349931716919,
      "learning_rate": 1e-05,
      "loss": 0.4386,
      "mean_token_accuracy": 0.8575445413589478,
      "num_tokens": 910445177.0,
      "step": 5713
    },
    {
      "epoch": 2.9064089521871823,
      "grad_norm": 0.8545286059379578,
      "learning_rate": 1e-05,
      "loss": 0.4319,
      "mean_token_accuracy": 0.8597594499588013,
      "num_tokens": 910603096.0,
      "step": 5714
    },
    {
      "epoch": 2.9069175991861647,
      "grad_norm": 0.8807753324508667,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8599470853805542,
      "num_tokens": 910760377.0,
      "step": 5715
    },
    {
      "epoch": 2.9074262461851474,
      "grad_norm": 0.8718137145042419,
      "learning_rate": 1e-05,
      "loss": 0.4324,
      "mean_token_accuracy": 0.8583258390426636,
      "num_tokens": 910907889.0,
      "step": 5716
    },
    {
      "epoch": 2.9079348931841302,
      "grad_norm": 0.9387333989143372,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "mean_token_accuracy": 0.8550516366958618,
      "num_tokens": 911055378.0,
      "step": 5717
    },
    {
      "epoch": 2.908443540183113,
      "grad_norm": 0.8214811086654663,
      "learning_rate": 1e-05,
      "loss": 0.4314,
      "mean_token_accuracy": 0.8608577847480774,
      "num_tokens": 911219506.0,
      "step": 5718
    },
    {
      "epoch": 2.908952187182096,
      "grad_norm": 0.8933064937591553,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "mean_token_accuracy": 0.8551285266876221,
      "num_tokens": 911381527.0,
      "step": 5719
    },
    {
      "epoch": 2.909460834181078,
      "grad_norm": 0.8504346609115601,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8552653789520264,
      "num_tokens": 911543675.0,
      "step": 5720
    },
    {
      "epoch": 2.909969481180061,
      "grad_norm": 0.8746544122695923,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "mean_token_accuracy": 0.855337381362915,
      "num_tokens": 911704808.0,
      "step": 5721
    },
    {
      "epoch": 2.9104781281790437,
      "grad_norm": 0.8669593930244446,
      "learning_rate": 1e-05,
      "loss": 0.4189,
      "mean_token_accuracy": 0.864909291267395,
      "num_tokens": 911875094.0,
      "step": 5722
    },
    {
      "epoch": 2.9109867751780265,
      "grad_norm": 0.9222279787063599,
      "learning_rate": 1e-05,
      "loss": 0.4572,
      "mean_token_accuracy": 0.8513796925544739,
      "num_tokens": 912035787.0,
      "step": 5723
    },
    {
      "epoch": 2.9114954221770093,
      "grad_norm": 0.8894267082214355,
      "learning_rate": 1e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.8590693473815918,
      "num_tokens": 912185571.0,
      "step": 5724
    },
    {
      "epoch": 2.9120040691759916,
      "grad_norm": 0.8288862705230713,
      "learning_rate": 1e-05,
      "loss": 0.398,
      "mean_token_accuracy": 0.8673139810562134,
      "num_tokens": 912347614.0,
      "step": 5725
    },
    {
      "epoch": 2.9125127161749744,
      "grad_norm": 0.8889040946960449,
      "learning_rate": 1e-05,
      "loss": 0.4656,
      "mean_token_accuracy": 0.8502926826477051,
      "num_tokens": 912513109.0,
      "step": 5726
    },
    {
      "epoch": 2.913021363173957,
      "grad_norm": 0.8634620308876038,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "mean_token_accuracy": 0.8603816032409668,
      "num_tokens": 912679521.0,
      "step": 5727
    },
    {
      "epoch": 2.91353001017294,
      "grad_norm": 0.809637725353241,
      "learning_rate": 1e-05,
      "loss": 0.4429,
      "mean_token_accuracy": 0.8566316962242126,
      "num_tokens": 912850119.0,
      "step": 5728
    },
    {
      "epoch": 2.914038657171923,
      "grad_norm": 0.857983410358429,
      "learning_rate": 1e-05,
      "loss": 0.4201,
      "mean_token_accuracy": 0.863347053527832,
      "num_tokens": 913006967.0,
      "step": 5729
    },
    {
      "epoch": 2.914547304170905,
      "grad_norm": 0.8073707818984985,
      "learning_rate": 1e-05,
      "loss": 0.4353,
      "mean_token_accuracy": 0.8588767647743225,
      "num_tokens": 913178881.0,
      "step": 5730
    },
    {
      "epoch": 2.9150559511698884,
      "grad_norm": 0.8749431371688843,
      "learning_rate": 1e-05,
      "loss": 0.4402,
      "mean_token_accuracy": 0.8577772974967957,
      "num_tokens": 913345431.0,
      "step": 5731
    },
    {
      "epoch": 2.9155645981688707,
      "grad_norm": 0.8460366725921631,
      "learning_rate": 1e-05,
      "loss": 0.4291,
      "mean_token_accuracy": 0.8599922060966492,
      "num_tokens": 913518334.0,
      "step": 5732
    },
    {
      "epoch": 2.9160732451678535,
      "grad_norm": 0.8206888437271118,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.86368727684021,
      "num_tokens": 913677797.0,
      "step": 5733
    },
    {
      "epoch": 2.9165818921668363,
      "grad_norm": 0.9418649077415466,
      "learning_rate": 1e-05,
      "loss": 0.4434,
      "mean_token_accuracy": 0.8554619550704956,
      "num_tokens": 913821428.0,
      "step": 5734
    },
    {
      "epoch": 2.917090539165819,
      "grad_norm": 0.8547963500022888,
      "learning_rate": 1e-05,
      "loss": 0.4176,
      "mean_token_accuracy": 0.8628682494163513,
      "num_tokens": 913969348.0,
      "step": 5735
    },
    {
      "epoch": 2.917599186164802,
      "grad_norm": 0.8919240236282349,
      "learning_rate": 1e-05,
      "loss": 0.4215,
      "mean_token_accuracy": 0.8628110885620117,
      "num_tokens": 914126038.0,
      "step": 5736
    },
    {
      "epoch": 2.918107833163784,
      "grad_norm": 0.914557933807373,
      "learning_rate": 1e-05,
      "loss": 0.4351,
      "mean_token_accuracy": 0.8588264584541321,
      "num_tokens": 914289459.0,
      "step": 5737
    },
    {
      "epoch": 2.918616480162767,
      "grad_norm": 0.817688524723053,
      "learning_rate": 1e-05,
      "loss": 0.4534,
      "mean_token_accuracy": 0.8541843891143799,
      "num_tokens": 914459291.0,
      "step": 5738
    },
    {
      "epoch": 2.9191251271617498,
      "grad_norm": 0.9456674456596375,
      "learning_rate": 1e-05,
      "loss": 0.4091,
      "mean_token_accuracy": 0.8657861948013306,
      "num_tokens": 914608971.0,
      "step": 5739
    },
    {
      "epoch": 2.9196337741607326,
      "grad_norm": 0.855796754360199,
      "learning_rate": 1e-05,
      "loss": 0.4178,
      "mean_token_accuracy": 0.8632485270500183,
      "num_tokens": 914765805.0,
      "step": 5740
    },
    {
      "epoch": 2.9201424211597153,
      "grad_norm": 0.9075614213943481,
      "learning_rate": 1e-05,
      "loss": 0.4277,
      "mean_token_accuracy": 0.8599561452865601,
      "num_tokens": 914923608.0,
      "step": 5741
    },
    {
      "epoch": 2.9206510681586977,
      "grad_norm": 1.9019993543624878,
      "learning_rate": 1e-05,
      "loss": 0.4904,
      "mean_token_accuracy": 0.844566822052002,
      "num_tokens": 915083509.0,
      "step": 5742
    },
    {
      "epoch": 2.9211597151576805,
      "grad_norm": 0.9290248155593872,
      "learning_rate": 1e-05,
      "loss": 0.4099,
      "mean_token_accuracy": 0.865955650806427,
      "num_tokens": 915245487.0,
      "step": 5743
    },
    {
      "epoch": 2.9216683621566633,
      "grad_norm": 1.0105957984924316,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "mean_token_accuracy": 0.8545387983322144,
      "num_tokens": 915391939.0,
      "step": 5744
    },
    {
      "epoch": 2.922177009155646,
      "grad_norm": 0.8672761917114258,
      "learning_rate": 1e-05,
      "loss": 0.4149,
      "mean_token_accuracy": 0.8643245697021484,
      "num_tokens": 915543628.0,
      "step": 5745
    },
    {
      "epoch": 2.922685656154629,
      "grad_norm": 0.8365616798400879,
      "learning_rate": 1e-05,
      "loss": 0.4474,
      "mean_token_accuracy": 0.8555518388748169,
      "num_tokens": 915708594.0,
      "step": 5746
    },
    {
      "epoch": 2.923194303153611,
      "grad_norm": 0.9645183682441711,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "mean_token_accuracy": 0.8628501892089844,
      "num_tokens": 915850846.0,
      "step": 5747
    },
    {
      "epoch": 2.923702950152594,
      "grad_norm": 0.9386368989944458,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "mean_token_accuracy": 0.8545777201652527,
      "num_tokens": 915997116.0,
      "step": 5748
    },
    {
      "epoch": 2.9242115971515767,
      "grad_norm": 0.8516026735305786,
      "learning_rate": 1e-05,
      "loss": 0.4543,
      "mean_token_accuracy": 0.8535363078117371,
      "num_tokens": 916164228.0,
      "step": 5749
    },
    {
      "epoch": 2.9247202441505595,
      "grad_norm": 1.0035674571990967,
      "learning_rate": 1e-05,
      "loss": 0.4681,
      "mean_token_accuracy": 0.8500915765762329,
      "num_tokens": 916313182.0,
      "step": 5750
    },
    {
      "epoch": 2.9252288911495423,
      "grad_norm": 0.8694740533828735,
      "learning_rate": 1e-05,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8602680563926697,
      "num_tokens": 916468393.0,
      "step": 5751
    },
    {
      "epoch": 2.9257375381485247,
      "grad_norm": 0.9328964352607727,
      "learning_rate": 1e-05,
      "loss": 0.4351,
      "mean_token_accuracy": 0.8590227365493774,
      "num_tokens": 916617077.0,
      "step": 5752
    },
    {
      "epoch": 2.926246185147508,
      "grad_norm": 0.9030390977859497,
      "learning_rate": 1e-05,
      "loss": 0.4144,
      "mean_token_accuracy": 0.8632201552391052,
      "num_tokens": 916764508.0,
      "step": 5753
    },
    {
      "epoch": 2.9267548321464902,
      "grad_norm": 0.918763279914856,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8566667437553406,
      "num_tokens": 916935162.0,
      "step": 5754
    },
    {
      "epoch": 2.927263479145473,
      "grad_norm": 0.8663561344146729,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "mean_token_accuracy": 0.853164553642273,
      "num_tokens": 917106077.0,
      "step": 5755
    },
    {
      "epoch": 2.927772126144456,
      "grad_norm": 0.9716460704803467,
      "learning_rate": 1e-05,
      "loss": 0.4704,
      "mean_token_accuracy": 0.8501912355422974,
      "num_tokens": 917266918.0,
      "step": 5756
    },
    {
      "epoch": 2.9282807731434386,
      "grad_norm": 0.9873749017715454,
      "learning_rate": 1e-05,
      "loss": 0.4009,
      "mean_token_accuracy": 0.868045449256897,
      "num_tokens": 917420760.0,
      "step": 5757
    },
    {
      "epoch": 2.9287894201424214,
      "grad_norm": 0.9092333912849426,
      "learning_rate": 1e-05,
      "loss": 0.4254,
      "mean_token_accuracy": 0.8617450594902039,
      "num_tokens": 917586569.0,
      "step": 5758
    },
    {
      "epoch": 2.9292980671414037,
      "grad_norm": 0.8969524502754211,
      "learning_rate": 1e-05,
      "loss": 0.458,
      "mean_token_accuracy": 0.8509863615036011,
      "num_tokens": 917750067.0,
      "step": 5759
    },
    {
      "epoch": 2.9298067141403865,
      "grad_norm": 0.9118671417236328,
      "learning_rate": 1e-05,
      "loss": 0.4474,
      "mean_token_accuracy": 0.8554458618164062,
      "num_tokens": 917913379.0,
      "step": 5760
    },
    {
      "epoch": 2.9303153611393693,
      "grad_norm": 0.8653302788734436,
      "learning_rate": 1e-05,
      "loss": 0.4452,
      "mean_token_accuracy": 0.855238139629364,
      "num_tokens": 918072489.0,
      "step": 5761
    },
    {
      "epoch": 2.930824008138352,
      "grad_norm": 1.0064209699630737,
      "learning_rate": 1e-05,
      "loss": 0.4518,
      "mean_token_accuracy": 0.8538899421691895,
      "num_tokens": 918219033.0,
      "step": 5762
    },
    {
      "epoch": 2.931332655137335,
      "grad_norm": 0.9989521503448486,
      "learning_rate": 1e-05,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8602008819580078,
      "num_tokens": 918368824.0,
      "step": 5763
    },
    {
      "epoch": 2.931841302136317,
      "grad_norm": 0.8290295004844666,
      "learning_rate": 1e-05,
      "loss": 0.438,
      "mean_token_accuracy": 0.8582369685173035,
      "num_tokens": 918538998.0,
      "step": 5764
    },
    {
      "epoch": 2.9323499491353,
      "grad_norm": 0.9657419919967651,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.8566818237304688,
      "num_tokens": 918692837.0,
      "step": 5765
    },
    {
      "epoch": 2.932858596134283,
      "grad_norm": 0.9440754055976868,
      "learning_rate": 1e-05,
      "loss": 0.4077,
      "mean_token_accuracy": 0.8666754961013794,
      "num_tokens": 918848120.0,
      "step": 5766
    },
    {
      "epoch": 2.9333672431332656,
      "grad_norm": 0.843254566192627,
      "learning_rate": 1e-05,
      "loss": 0.4247,
      "mean_token_accuracy": 0.8619252443313599,
      "num_tokens": 919000637.0,
      "step": 5767
    },
    {
      "epoch": 2.9338758901322484,
      "grad_norm": 0.9444208741188049,
      "learning_rate": 1e-05,
      "loss": 0.4594,
      "mean_token_accuracy": 0.8507851362228394,
      "num_tokens": 919158162.0,
      "step": 5768
    },
    {
      "epoch": 2.9343845371312307,
      "grad_norm": 0.9736772179603577,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "mean_token_accuracy": 0.8532898426055908,
      "num_tokens": 919298786.0,
      "step": 5769
    },
    {
      "epoch": 2.9348931841302135,
      "grad_norm": 0.8969355225563049,
      "learning_rate": 1e-05,
      "loss": 0.4455,
      "mean_token_accuracy": 0.855862021446228,
      "num_tokens": 919454435.0,
      "step": 5770
    },
    {
      "epoch": 2.9354018311291963,
      "grad_norm": 0.9506131410598755,
      "learning_rate": 1e-05,
      "loss": 0.4198,
      "mean_token_accuracy": 0.8641059398651123,
      "num_tokens": 919611410.0,
      "step": 5771
    },
    {
      "epoch": 2.935910478128179,
      "grad_norm": 0.923133909702301,
      "learning_rate": 1e-05,
      "loss": 0.4606,
      "mean_token_accuracy": 0.8519150018692017,
      "num_tokens": 919767371.0,
      "step": 5772
    },
    {
      "epoch": 2.936419125127162,
      "grad_norm": 0.8677843809127808,
      "learning_rate": 1e-05,
      "loss": 0.4098,
      "mean_token_accuracy": 0.8654341101646423,
      "num_tokens": 919927418.0,
      "step": 5773
    },
    {
      "epoch": 2.936927772126144,
      "grad_norm": 0.8953611850738525,
      "learning_rate": 1e-05,
      "loss": 0.4254,
      "mean_token_accuracy": 0.8617711067199707,
      "num_tokens": 920097231.0,
      "step": 5774
    },
    {
      "epoch": 2.9374364191251274,
      "grad_norm": 0.9595847725868225,
      "learning_rate": 1e-05,
      "loss": 0.4293,
      "mean_token_accuracy": 0.8610657453536987,
      "num_tokens": 920251209.0,
      "step": 5775
    },
    {
      "epoch": 2.9379450661241098,
      "grad_norm": 0.839107871055603,
      "learning_rate": 1e-05,
      "loss": 0.4336,
      "mean_token_accuracy": 0.8592088222503662,
      "num_tokens": 920404124.0,
      "step": 5776
    },
    {
      "epoch": 2.9384537131230926,
      "grad_norm": 0.9945467114448547,
      "learning_rate": 1e-05,
      "loss": 0.4794,
      "mean_token_accuracy": 0.8458770513534546,
      "num_tokens": 920562123.0,
      "step": 5777
    },
    {
      "epoch": 2.9389623601220753,
      "grad_norm": 0.9366150498390198,
      "learning_rate": 1e-05,
      "loss": 0.4482,
      "mean_token_accuracy": 0.854699969291687,
      "num_tokens": 920723542.0,
      "step": 5778
    },
    {
      "epoch": 2.939471007121058,
      "grad_norm": 0.9452741742134094,
      "learning_rate": 1e-05,
      "loss": 0.4341,
      "mean_token_accuracy": 0.8599119782447815,
      "num_tokens": 920881486.0,
      "step": 5779
    },
    {
      "epoch": 2.939979654120041,
      "grad_norm": 0.9388673305511475,
      "learning_rate": 1e-05,
      "loss": 0.4526,
      "mean_token_accuracy": 0.853776216506958,
      "num_tokens": 921032468.0,
      "step": 5780
    },
    {
      "epoch": 2.9404883011190233,
      "grad_norm": 0.9121649265289307,
      "learning_rate": 1e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8614283800125122,
      "num_tokens": 921185787.0,
      "step": 5781
    },
    {
      "epoch": 2.940996948118006,
      "grad_norm": 0.8633955717086792,
      "learning_rate": 1e-05,
      "loss": 0.4254,
      "mean_token_accuracy": 0.8609992861747742,
      "num_tokens": 921343839.0,
      "step": 5782
    },
    {
      "epoch": 2.941505595116989,
      "grad_norm": 0.9059745073318481,
      "learning_rate": 1e-05,
      "loss": 0.4727,
      "mean_token_accuracy": 0.8484917879104614,
      "num_tokens": 921504957.0,
      "step": 5783
    },
    {
      "epoch": 2.9420142421159716,
      "grad_norm": 0.8350028991699219,
      "learning_rate": 1e-05,
      "loss": 0.4218,
      "mean_token_accuracy": 0.8621200323104858,
      "num_tokens": 921677191.0,
      "step": 5784
    },
    {
      "epoch": 2.9425228891149544,
      "grad_norm": 0.9048314094543457,
      "learning_rate": 1e-05,
      "loss": 0.4706,
      "mean_token_accuracy": 0.8496427536010742,
      "num_tokens": 921834278.0,
      "step": 5785
    },
    {
      "epoch": 2.9430315361139368,
      "grad_norm": 0.8689739108085632,
      "learning_rate": 1e-05,
      "loss": 0.4199,
      "mean_token_accuracy": 0.8630783557891846,
      "num_tokens": 921985325.0,
      "step": 5786
    },
    {
      "epoch": 2.9435401831129195,
      "grad_norm": 0.8724319934844971,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "mean_token_accuracy": 0.8558523654937744,
      "num_tokens": 922150194.0,
      "step": 5787
    },
    {
      "epoch": 2.9440488301119023,
      "grad_norm": 0.8524068593978882,
      "learning_rate": 1e-05,
      "loss": 0.391,
      "mean_token_accuracy": 0.8696946501731873,
      "num_tokens": 922307545.0,
      "step": 5788
    },
    {
      "epoch": 2.944557477110885,
      "grad_norm": 0.9184873104095459,
      "learning_rate": 1e-05,
      "loss": 0.4684,
      "mean_token_accuracy": 0.8503162860870361,
      "num_tokens": 922467567.0,
      "step": 5789
    },
    {
      "epoch": 2.945066124109868,
      "grad_norm": 0.8801780939102173,
      "learning_rate": 1e-05,
      "loss": 0.4286,
      "mean_token_accuracy": 0.8609876036643982,
      "num_tokens": 922627414.0,
      "step": 5790
    },
    {
      "epoch": 2.9455747711088502,
      "grad_norm": 0.9340073466300964,
      "learning_rate": 1e-05,
      "loss": 0.4448,
      "mean_token_accuracy": 0.856844961643219,
      "num_tokens": 922782327.0,
      "step": 5791
    },
    {
      "epoch": 2.946083418107833,
      "grad_norm": 0.9026318192481995,
      "learning_rate": 1e-05,
      "loss": 0.4559,
      "mean_token_accuracy": 0.8538775444030762,
      "num_tokens": 922940862.0,
      "step": 5792
    },
    {
      "epoch": 2.946592065106816,
      "grad_norm": 0.8768016695976257,
      "learning_rate": 1e-05,
      "loss": 0.4637,
      "mean_token_accuracy": 0.8493372797966003,
      "num_tokens": 923105553.0,
      "step": 5793
    },
    {
      "epoch": 2.9471007121057986,
      "grad_norm": 0.8517777919769287,
      "learning_rate": 1e-05,
      "loss": 0.4204,
      "mean_token_accuracy": 0.8641206622123718,
      "num_tokens": 923268455.0,
      "step": 5794
    },
    {
      "epoch": 2.9476093591047814,
      "grad_norm": 0.8783465027809143,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "mean_token_accuracy": 0.8550516963005066,
      "num_tokens": 923426151.0,
      "step": 5795
    },
    {
      "epoch": 2.9481180061037637,
      "grad_norm": 0.8488904237747192,
      "learning_rate": 1e-05,
      "loss": 0.4081,
      "mean_token_accuracy": 0.8654050230979919,
      "num_tokens": 923581013.0,
      "step": 5796
    },
    {
      "epoch": 2.948626653102747,
      "grad_norm": 0.8633197546005249,
      "learning_rate": 1e-05,
      "loss": 0.4261,
      "mean_token_accuracy": 0.860765278339386,
      "num_tokens": 923735620.0,
      "step": 5797
    },
    {
      "epoch": 2.9491353001017293,
      "grad_norm": 0.8653395771980286,
      "learning_rate": 1e-05,
      "loss": 0.4483,
      "mean_token_accuracy": 0.8558655381202698,
      "num_tokens": 923889387.0,
      "step": 5798
    },
    {
      "epoch": 2.949643947100712,
      "grad_norm": 0.9279692769050598,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "mean_token_accuracy": 0.8564894199371338,
      "num_tokens": 924043776.0,
      "step": 5799
    },
    {
      "epoch": 2.950152594099695,
      "grad_norm": 0.8566408753395081,
      "learning_rate": 1e-05,
      "loss": 0.4252,
      "mean_token_accuracy": 0.8626484274864197,
      "num_tokens": 924193594.0,
      "step": 5800
    },
    {
      "epoch": 2.9506612410986777,
      "grad_norm": 0.8193241953849792,
      "learning_rate": 1e-05,
      "loss": 0.4094,
      "mean_token_accuracy": 0.8671364784240723,
      "num_tokens": 924362420.0,
      "step": 5801
    },
    {
      "epoch": 2.9511698880976605,
      "grad_norm": 0.8682567477226257,
      "learning_rate": 1e-05,
      "loss": 0.4405,
      "mean_token_accuracy": 0.857014536857605,
      "num_tokens": 924523122.0,
      "step": 5802
    },
    {
      "epoch": 2.951678535096643,
      "grad_norm": 0.8148035407066345,
      "learning_rate": 1e-05,
      "loss": 0.3869,
      "mean_token_accuracy": 0.8721718788146973,
      "num_tokens": 924684963.0,
      "step": 5803
    },
    {
      "epoch": 2.9521871820956256,
      "grad_norm": 0.9142512083053589,
      "learning_rate": 1e-05,
      "loss": 0.4299,
      "mean_token_accuracy": 0.8605676889419556,
      "num_tokens": 924829073.0,
      "step": 5804
    },
    {
      "epoch": 2.9526958290946084,
      "grad_norm": 0.9010807275772095,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "mean_token_accuracy": 0.8550687432289124,
      "num_tokens": 925002218.0,
      "step": 5805
    },
    {
      "epoch": 2.953204476093591,
      "grad_norm": 0.8682645559310913,
      "learning_rate": 1e-05,
      "loss": 0.4288,
      "mean_token_accuracy": 0.8604177236557007,
      "num_tokens": 925176143.0,
      "step": 5806
    },
    {
      "epoch": 2.953713123092574,
      "grad_norm": 0.875255286693573,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "mean_token_accuracy": 0.8656608462333679,
      "num_tokens": 925331149.0,
      "step": 5807
    },
    {
      "epoch": 2.9542217700915563,
      "grad_norm": 0.9133750796318054,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "mean_token_accuracy": 0.8645917773246765,
      "num_tokens": 925477479.0,
      "step": 5808
    },
    {
      "epoch": 2.954730417090539,
      "grad_norm": 0.8490424752235413,
      "learning_rate": 1e-05,
      "loss": 0.4319,
      "mean_token_accuracy": 0.8602641820907593,
      "num_tokens": 925643872.0,
      "step": 5809
    },
    {
      "epoch": 2.955239064089522,
      "grad_norm": 0.8649740219116211,
      "learning_rate": 1e-05,
      "loss": 0.4087,
      "mean_token_accuracy": 0.8658615946769714,
      "num_tokens": 925794642.0,
      "step": 5810
    },
    {
      "epoch": 2.9557477110885046,
      "grad_norm": 0.8914206624031067,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "mean_token_accuracy": 0.8545023798942566,
      "num_tokens": 925945760.0,
      "step": 5811
    },
    {
      "epoch": 2.9562563580874874,
      "grad_norm": 0.8733227252960205,
      "learning_rate": 1e-05,
      "loss": 0.4024,
      "mean_token_accuracy": 0.8668968677520752,
      "num_tokens": 926085112.0,
      "step": 5812
    },
    {
      "epoch": 2.9567650050864698,
      "grad_norm": 0.8863427042961121,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "mean_token_accuracy": 0.8588989973068237,
      "num_tokens": 926240188.0,
      "step": 5813
    },
    {
      "epoch": 2.9572736520854526,
      "grad_norm": 0.9870784878730774,
      "learning_rate": 1e-05,
      "loss": 0.471,
      "mean_token_accuracy": 0.8487330079078674,
      "num_tokens": 926404523.0,
      "step": 5814
    },
    {
      "epoch": 2.9577822990844354,
      "grad_norm": 0.8670083284378052,
      "learning_rate": 1e-05,
      "loss": 0.4,
      "mean_token_accuracy": 0.867817759513855,
      "num_tokens": 926553199.0,
      "step": 5815
    },
    {
      "epoch": 2.958290946083418,
      "grad_norm": 0.8427149653434753,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "mean_token_accuracy": 0.8652034997940063,
      "num_tokens": 926703380.0,
      "step": 5816
    },
    {
      "epoch": 2.958799593082401,
      "grad_norm": 0.8586266040802002,
      "learning_rate": 1e-05,
      "loss": 0.4348,
      "mean_token_accuracy": 0.8596380352973938,
      "num_tokens": 926863378.0,
      "step": 5817
    },
    {
      "epoch": 2.9593082400813833,
      "grad_norm": 0.9027202129364014,
      "learning_rate": 1e-05,
      "loss": 0.4329,
      "mean_token_accuracy": 0.8599502444267273,
      "num_tokens": 927009636.0,
      "step": 5818
    },
    {
      "epoch": 2.9598168870803665,
      "grad_norm": 0.8678000569343567,
      "learning_rate": 1e-05,
      "loss": 0.4253,
      "mean_token_accuracy": 0.8616846799850464,
      "num_tokens": 927168274.0,
      "step": 5819
    },
    {
      "epoch": 2.960325534079349,
      "grad_norm": 0.8938841819763184,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8489087820053101,
      "num_tokens": 927349363.0,
      "step": 5820
    },
    {
      "epoch": 2.9608341810783316,
      "grad_norm": 0.8878791928291321,
      "learning_rate": 1e-05,
      "loss": 0.4292,
      "mean_token_accuracy": 0.8604944348335266,
      "num_tokens": 927504872.0,
      "step": 5821
    },
    {
      "epoch": 2.9613428280773144,
      "grad_norm": 0.8782253265380859,
      "learning_rate": 1e-05,
      "loss": 0.4121,
      "mean_token_accuracy": 0.8634744882583618,
      "num_tokens": 927661962.0,
      "step": 5822
    },
    {
      "epoch": 2.961851475076297,
      "grad_norm": 0.8695517778396606,
      "learning_rate": 1e-05,
      "loss": 0.4543,
      "mean_token_accuracy": 0.8522966504096985,
      "num_tokens": 927820356.0,
      "step": 5823
    },
    {
      "epoch": 2.96236012207528,
      "grad_norm": 1.1400654315948486,
      "learning_rate": 1e-05,
      "loss": 0.4255,
      "mean_token_accuracy": 0.8615685701370239,
      "num_tokens": 927984140.0,
      "step": 5824
    },
    {
      "epoch": 2.9628687690742623,
      "grad_norm": 0.8361784219741821,
      "learning_rate": 1e-05,
      "loss": 0.4135,
      "mean_token_accuracy": 0.8660613298416138,
      "num_tokens": 928149741.0,
      "step": 5825
    },
    {
      "epoch": 2.963377416073245,
      "grad_norm": 0.8206629157066345,
      "learning_rate": 1e-05,
      "loss": 0.44,
      "mean_token_accuracy": 0.8572777509689331,
      "num_tokens": 928321531.0,
      "step": 5826
    },
    {
      "epoch": 2.963886063072228,
      "grad_norm": 0.889646589756012,
      "learning_rate": 1e-05,
      "loss": 0.4235,
      "mean_token_accuracy": 0.8619962930679321,
      "num_tokens": 928470935.0,
      "step": 5827
    },
    {
      "epoch": 2.9643947100712107,
      "grad_norm": 0.9425367712974548,
      "learning_rate": 1e-05,
      "loss": 0.407,
      "mean_token_accuracy": 0.8678266406059265,
      "num_tokens": 928624216.0,
      "step": 5828
    },
    {
      "epoch": 2.9649033570701935,
      "grad_norm": 0.8638216257095337,
      "learning_rate": 1e-05,
      "loss": 0.4432,
      "mean_token_accuracy": 0.8562783002853394,
      "num_tokens": 928787184.0,
      "step": 5829
    },
    {
      "epoch": 2.965412004069176,
      "grad_norm": 0.8920972943305969,
      "learning_rate": 1e-05,
      "loss": 0.4364,
      "mean_token_accuracy": 0.8587175607681274,
      "num_tokens": 928950416.0,
      "step": 5830
    },
    {
      "epoch": 2.9659206510681586,
      "grad_norm": 0.8766757845878601,
      "learning_rate": 1e-05,
      "loss": 0.4639,
      "mean_token_accuracy": 0.8482862114906311,
      "num_tokens": 929103335.0,
      "step": 5831
    },
    {
      "epoch": 2.9664292980671414,
      "grad_norm": 0.9349969029426575,
      "learning_rate": 1e-05,
      "loss": 0.4122,
      "mean_token_accuracy": 0.8652468919754028,
      "num_tokens": 929254662.0,
      "step": 5832
    },
    {
      "epoch": 2.966937945066124,
      "grad_norm": 0.9228084683418274,
      "learning_rate": 1e-05,
      "loss": 0.4211,
      "mean_token_accuracy": 0.8630307912826538,
      "num_tokens": 929403805.0,
      "step": 5833
    },
    {
      "epoch": 2.967446592065107,
      "grad_norm": 0.860146701335907,
      "learning_rate": 1e-05,
      "loss": 0.4109,
      "mean_token_accuracy": 0.8636744618415833,
      "num_tokens": 929559750.0,
      "step": 5834
    },
    {
      "epoch": 2.9679552390640893,
      "grad_norm": 0.9154155254364014,
      "learning_rate": 1e-05,
      "loss": 0.4011,
      "mean_token_accuracy": 0.8682236075401306,
      "num_tokens": 929708224.0,
      "step": 5835
    },
    {
      "epoch": 2.968463886063072,
      "grad_norm": 0.8699504733085632,
      "learning_rate": 1e-05,
      "loss": 0.4218,
      "mean_token_accuracy": 0.8614717721939087,
      "num_tokens": 929861530.0,
      "step": 5836
    },
    {
      "epoch": 2.968972533062055,
      "grad_norm": 0.9429341554641724,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8554191589355469,
      "num_tokens": 930013475.0,
      "step": 5837
    },
    {
      "epoch": 2.9694811800610377,
      "grad_norm": 0.837781548500061,
      "learning_rate": 1e-05,
      "loss": 0.427,
      "mean_token_accuracy": 0.8609052896499634,
      "num_tokens": 930177144.0,
      "step": 5838
    },
    {
      "epoch": 2.9699898270600205,
      "grad_norm": 0.9412606358528137,
      "learning_rate": 1e-05,
      "loss": 0.4665,
      "mean_token_accuracy": 0.8494040966033936,
      "num_tokens": 930339758.0,
      "step": 5839
    },
    {
      "epoch": 2.970498474059003,
      "grad_norm": 0.8832743167877197,
      "learning_rate": 1e-05,
      "loss": 0.437,
      "mean_token_accuracy": 0.8587670922279358,
      "num_tokens": 930500065.0,
      "step": 5840
    },
    {
      "epoch": 2.971007121057986,
      "grad_norm": 0.9219328165054321,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "mean_token_accuracy": 0.8565832376480103,
      "num_tokens": 930650312.0,
      "step": 5841
    },
    {
      "epoch": 2.9715157680569684,
      "grad_norm": 0.9061831831932068,
      "learning_rate": 1e-05,
      "loss": 0.4494,
      "mean_token_accuracy": 0.8533433675765991,
      "num_tokens": 930809980.0,
      "step": 5842
    },
    {
      "epoch": 2.972024415055951,
      "grad_norm": 0.8891438245773315,
      "learning_rate": 1e-05,
      "loss": 0.4757,
      "mean_token_accuracy": 0.8493027687072754,
      "num_tokens": 930965995.0,
      "step": 5843
    },
    {
      "epoch": 2.972533062054934,
      "grad_norm": 0.869954526424408,
      "learning_rate": 1e-05,
      "loss": 0.3767,
      "mean_token_accuracy": 0.8745460510253906,
      "num_tokens": 931124035.0,
      "step": 5844
    },
    {
      "epoch": 2.9730417090539167,
      "grad_norm": 0.8907552361488342,
      "learning_rate": 1e-05,
      "loss": 0.4332,
      "mean_token_accuracy": 0.8616758584976196,
      "num_tokens": 931271106.0,
      "step": 5845
    },
    {
      "epoch": 2.9735503560528995,
      "grad_norm": 0.9448345303535461,
      "learning_rate": 1e-05,
      "loss": 0.4254,
      "mean_token_accuracy": 0.8614962100982666,
      "num_tokens": 931415682.0,
      "step": 5846
    },
    {
      "epoch": 2.974059003051882,
      "grad_norm": 0.8706139922142029,
      "learning_rate": 1e-05,
      "loss": 0.4197,
      "mean_token_accuracy": 0.8637193441390991,
      "num_tokens": 931571305.0,
      "step": 5847
    },
    {
      "epoch": 2.9745676500508647,
      "grad_norm": 0.9198446869850159,
      "learning_rate": 1e-05,
      "loss": 0.4755,
      "mean_token_accuracy": 0.8487168550491333,
      "num_tokens": 931735254.0,
      "step": 5848
    },
    {
      "epoch": 2.9750762970498474,
      "grad_norm": 0.9850619435310364,
      "learning_rate": 1e-05,
      "loss": 0.4645,
      "mean_token_accuracy": 0.850271463394165,
      "num_tokens": 931891725.0,
      "step": 5849
    },
    {
      "epoch": 2.9755849440488302,
      "grad_norm": 0.9214954376220703,
      "learning_rate": 1e-05,
      "loss": 0.4462,
      "mean_token_accuracy": 0.856054425239563,
      "num_tokens": 932044046.0,
      "step": 5850
    },
    {
      "epoch": 2.976093591047813,
      "grad_norm": 0.9269073009490967,
      "learning_rate": 1e-05,
      "loss": 0.422,
      "mean_token_accuracy": 0.8621110320091248,
      "num_tokens": 932197593.0,
      "step": 5851
    },
    {
      "epoch": 2.9766022380467954,
      "grad_norm": 0.8905263543128967,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "mean_token_accuracy": 0.8598513603210449,
      "num_tokens": 932350730.0,
      "step": 5852
    },
    {
      "epoch": 2.977110885045778,
      "grad_norm": 0.8312326669692993,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "mean_token_accuracy": 0.8548107743263245,
      "num_tokens": 932513750.0,
      "step": 5853
    },
    {
      "epoch": 2.977619532044761,
      "grad_norm": 0.89973384141922,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "mean_token_accuracy": 0.8633626699447632,
      "num_tokens": 932673626.0,
      "step": 5854
    },
    {
      "epoch": 2.9781281790437437,
      "grad_norm": 0.8836349844932556,
      "learning_rate": 1e-05,
      "loss": 0.4065,
      "mean_token_accuracy": 0.8674150109291077,
      "num_tokens": 932846477.0,
      "step": 5855
    },
    {
      "epoch": 2.9786368260427265,
      "grad_norm": 0.8640139698982239,
      "learning_rate": 1e-05,
      "loss": 0.4461,
      "mean_token_accuracy": 0.8566869497299194,
      "num_tokens": 933021539.0,
      "step": 5856
    },
    {
      "epoch": 2.979145473041709,
      "grad_norm": 0.9574412107467651,
      "learning_rate": 1e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.8520747423171997,
      "num_tokens": 933173949.0,
      "step": 5857
    },
    {
      "epoch": 2.9796541200406916,
      "grad_norm": 0.8743879199028015,
      "learning_rate": 1e-05,
      "loss": 0.474,
      "mean_token_accuracy": 0.8490179181098938,
      "num_tokens": 933342649.0,
      "step": 5858
    },
    {
      "epoch": 2.9801627670396744,
      "grad_norm": 1.0385737419128418,
      "learning_rate": 1e-05,
      "loss": 0.4288,
      "mean_token_accuracy": 0.8602662086486816,
      "num_tokens": 933509483.0,
      "step": 5859
    },
    {
      "epoch": 2.980671414038657,
      "grad_norm": 0.8679269552230835,
      "learning_rate": 1e-05,
      "loss": 0.4137,
      "mean_token_accuracy": 0.8649242520332336,
      "num_tokens": 933683564.0,
      "step": 5860
    },
    {
      "epoch": 2.98118006103764,
      "grad_norm": 0.8368974924087524,
      "learning_rate": 1e-05,
      "loss": 0.4153,
      "mean_token_accuracy": 0.8648080229759216,
      "num_tokens": 933842448.0,
      "step": 5861
    },
    {
      "epoch": 2.9816887080366223,
      "grad_norm": 0.9177777767181396,
      "learning_rate": 1e-05,
      "loss": 0.4774,
      "mean_token_accuracy": 0.8469770550727844,
      "num_tokens": 933996328.0,
      "step": 5862
    },
    {
      "epoch": 2.982197355035605,
      "grad_norm": 0.9060273766517639,
      "learning_rate": 1e-05,
      "loss": 0.3981,
      "mean_token_accuracy": 0.8685603141784668,
      "num_tokens": 934146606.0,
      "step": 5863
    },
    {
      "epoch": 2.982706002034588,
      "grad_norm": 0.8164167404174805,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "mean_token_accuracy": 0.8569616079330444,
      "num_tokens": 934323985.0,
      "step": 5864
    },
    {
      "epoch": 2.9832146490335707,
      "grad_norm": 0.9017447233200073,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "mean_token_accuracy": 0.8646686673164368,
      "num_tokens": 934492502.0,
      "step": 5865
    },
    {
      "epoch": 2.9837232960325535,
      "grad_norm": 0.8669429421424866,
      "learning_rate": 1e-05,
      "loss": 0.4405,
      "mean_token_accuracy": 0.8561424016952515,
      "num_tokens": 934651946.0,
      "step": 5866
    },
    {
      "epoch": 2.9842319430315363,
      "grad_norm": 0.8700833320617676,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "mean_token_accuracy": 0.8638380169868469,
      "num_tokens": 934817411.0,
      "step": 5867
    },
    {
      "epoch": 2.984740590030519,
      "grad_norm": 0.8727999925613403,
      "learning_rate": 1e-05,
      "loss": 0.4343,
      "mean_token_accuracy": 0.8592092990875244,
      "num_tokens": 934983408.0,
      "step": 5868
    },
    {
      "epoch": 2.9852492370295014,
      "grad_norm": 0.8172105550765991,
      "learning_rate": 1e-05,
      "loss": 0.3962,
      "mean_token_accuracy": 0.8701355457305908,
      "num_tokens": 935138600.0,
      "step": 5869
    },
    {
      "epoch": 2.985757884028484,
      "grad_norm": 0.9610244035720825,
      "learning_rate": 1e-05,
      "loss": 0.4521,
      "mean_token_accuracy": 0.8546727895736694,
      "num_tokens": 935295080.0,
      "step": 5870
    },
    {
      "epoch": 2.986266531027467,
      "grad_norm": 0.8617233633995056,
      "learning_rate": 1e-05,
      "loss": 0.4283,
      "mean_token_accuracy": 0.8610825538635254,
      "num_tokens": 935460191.0,
      "step": 5871
    },
    {
      "epoch": 2.9867751780264498,
      "grad_norm": 0.8361219763755798,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.8569613695144653,
      "num_tokens": 935616845.0,
      "step": 5872
    },
    {
      "epoch": 2.9872838250254325,
      "grad_norm": 0.9874329566955566,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8622537851333618,
      "num_tokens": 935768955.0,
      "step": 5873
    },
    {
      "epoch": 2.987792472024415,
      "grad_norm": 0.9735478758811951,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "mean_token_accuracy": 0.8583329319953918,
      "num_tokens": 935925552.0,
      "step": 5874
    },
    {
      "epoch": 2.9883011190233977,
      "grad_norm": 0.9315745234489441,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "mean_token_accuracy": 0.8649601340293884,
      "num_tokens": 936086816.0,
      "step": 5875
    },
    {
      "epoch": 2.9888097660223805,
      "grad_norm": 0.9147635698318481,
      "learning_rate": 1e-05,
      "loss": 0.4417,
      "mean_token_accuracy": 0.8556721806526184,
      "num_tokens": 936253904.0,
      "step": 5876
    },
    {
      "epoch": 2.9893184130213633,
      "grad_norm": 0.935129702091217,
      "learning_rate": 1e-05,
      "loss": 0.4219,
      "mean_token_accuracy": 0.8635666370391846,
      "num_tokens": 936409194.0,
      "step": 5877
    },
    {
      "epoch": 2.989827060020346,
      "grad_norm": 0.9388399720191956,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "mean_token_accuracy": 0.8653729557991028,
      "num_tokens": 936577899.0,
      "step": 5878
    },
    {
      "epoch": 2.9903357070193284,
      "grad_norm": 0.8908611536026001,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "mean_token_accuracy": 0.856574296951294,
      "num_tokens": 936734381.0,
      "step": 5879
    },
    {
      "epoch": 2.990844354018311,
      "grad_norm": 0.8607943654060364,
      "learning_rate": 1e-05,
      "loss": 0.4315,
      "mean_token_accuracy": 0.8605170249938965,
      "num_tokens": 936910362.0,
      "step": 5880
    },
    {
      "epoch": 2.991353001017294,
      "grad_norm": 0.8888345956802368,
      "learning_rate": 1e-05,
      "loss": 0.4425,
      "mean_token_accuracy": 0.8557511568069458,
      "num_tokens": 937078132.0,
      "step": 5881
    },
    {
      "epoch": 2.9918616480162767,
      "grad_norm": 0.9754486680030823,
      "learning_rate": 1e-05,
      "loss": 0.4191,
      "mean_token_accuracy": 0.8625936508178711,
      "num_tokens": 937231947.0,
      "step": 5882
    },
    {
      "epoch": 2.9923702950152595,
      "grad_norm": 0.8599565029144287,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "mean_token_accuracy": 0.8562726378440857,
      "num_tokens": 937393536.0,
      "step": 5883
    },
    {
      "epoch": 2.992878942014242,
      "grad_norm": 0.9650074243545532,
      "learning_rate": 1e-05,
      "loss": 0.4699,
      "mean_token_accuracy": 0.8488419055938721,
      "num_tokens": 937547528.0,
      "step": 5884
    },
    {
      "epoch": 2.9933875890132247,
      "grad_norm": 0.9295585751533508,
      "learning_rate": 1e-05,
      "loss": 0.4081,
      "mean_token_accuracy": 0.8679110407829285,
      "num_tokens": 937697898.0,
      "step": 5885
    },
    {
      "epoch": 2.9938962360122074,
      "grad_norm": 0.8543989062309265,
      "learning_rate": 1e-05,
      "loss": 0.4233,
      "mean_token_accuracy": 0.8614925146102905,
      "num_tokens": 937866031.0,
      "step": 5886
    },
    {
      "epoch": 2.9944048830111902,
      "grad_norm": 0.9370037913322449,
      "learning_rate": 1e-05,
      "loss": 0.4275,
      "mean_token_accuracy": 0.8597909808158875,
      "num_tokens": 938023104.0,
      "step": 5887
    },
    {
      "epoch": 2.994913530010173,
      "grad_norm": 0.8910934329032898,
      "learning_rate": 1e-05,
      "loss": 0.4226,
      "mean_token_accuracy": 0.8629887104034424,
      "num_tokens": 938184792.0,
      "step": 5888
    },
    {
      "epoch": 2.995422177009156,
      "grad_norm": 1.5907645225524902,
      "learning_rate": 1e-05,
      "loss": 0.428,
      "mean_token_accuracy": 0.8609790205955505,
      "num_tokens": 938339130.0,
      "step": 5889
    },
    {
      "epoch": 2.9959308240081386,
      "grad_norm": 0.8379800915718079,
      "learning_rate": 1e-05,
      "loss": 0.441,
      "mean_token_accuracy": 0.8579897284507751,
      "num_tokens": 938504623.0,
      "step": 5890
    },
    {
      "epoch": 2.996439471007121,
      "grad_norm": 0.9151134490966797,
      "learning_rate": 1e-05,
      "loss": 0.4417,
      "mean_token_accuracy": 0.8580859899520874,
      "num_tokens": 938672321.0,
      "step": 5891
    },
    {
      "epoch": 2.9969481180061037,
      "grad_norm": 0.9333588480949402,
      "learning_rate": 1e-05,
      "loss": 0.43,
      "mean_token_accuracy": 0.8601522445678711,
      "num_tokens": 938827631.0,
      "step": 5892
    },
    {
      "epoch": 2.9974567650050865,
      "grad_norm": 0.8548453450202942,
      "learning_rate": 1e-05,
      "loss": 0.4148,
      "mean_token_accuracy": 0.86507648229599,
      "num_tokens": 938986452.0,
      "step": 5893
    },
    {
      "epoch": 2.9979654120040693,
      "grad_norm": 0.8833407759666443,
      "learning_rate": 1e-05,
      "loss": 0.4301,
      "mean_token_accuracy": 0.860978364944458,
      "num_tokens": 939140656.0,
      "step": 5894
    },
    {
      "epoch": 2.998474059003052,
      "grad_norm": 0.8966116309165955,
      "learning_rate": 1e-05,
      "loss": 0.4235,
      "mean_token_accuracy": 0.8621483445167542,
      "num_tokens": 939281857.0,
      "step": 5895
    },
    {
      "epoch": 2.9989827060020344,
      "grad_norm": 0.8217146396636963,
      "learning_rate": 1e-05,
      "loss": 0.4287,
      "mean_token_accuracy": 0.8596178293228149,
      "num_tokens": 939437597.0,
      "step": 5896
    },
    {
      "epoch": 2.999491353001017,
      "grad_norm": 0.8681790828704834,
      "learning_rate": 1e-05,
      "loss": 0.423,
      "mean_token_accuracy": 0.8632126450538635,
      "num_tokens": 939611691.0,
      "step": 5897
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8482578992843628,
      "learning_rate": 1e-05,
      "loss": 0.4063,
      "mean_token_accuracy": 0.8673560619354248,
      "num_tokens": 939774271.0,
      "step": 5898
    },
    {
      "epoch": 3.0,
      "step": 5898,
      "total_flos": 6.289145677364593e+18,
      "train_loss": 0.497733977299579,
      "train_runtime": 8529.1423,
      "train_samples_per_second": 44.238,
      "train_steps_per_second": 0.692
    }
  ],
  "logging_steps": 1,
  "max_steps": 5898,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 2949,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.289145677364593e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}