{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.264565858082068,
  "global_step": 1800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 0.0001,
      "loss": 1.1655,
      "step": 10
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0001,
      "loss": 1.001,
      "step": 20
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0001,
      "loss": 1.0287,
      "step": 30
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0001,
      "loss": 1.1578,
      "step": 40
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0001,
      "loss": 1.2146,
      "step": 50
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0001,
      "loss": 0.997,
      "step": 60
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0001,
      "loss": 0.9024,
      "step": 70
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0001,
      "loss": 0.9901,
      "step": 80
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0001,
      "loss": 1.1264,
      "step": 90
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0001,
      "loss": 1.2038,
      "step": 100
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0001,
      "loss": 0.8935,
      "step": 110
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0001,
      "loss": 0.9178,
      "step": 120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0001,
      "loss": 0.9746,
      "step": 130
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0001,
      "loss": 1.1566,
      "step": 140
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0001,
      "loss": 1.2877,
      "step": 150
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0001,
      "loss": 0.9146,
      "step": 160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0001,
      "loss": 0.8895,
      "step": 170
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0001,
      "loss": 1.0121,
      "step": 180
    },
    {
      "epoch": 0.34,
      "eval_loss": 1.0215636491775513,
      "eval_runtime": 950.138,
      "eval_samples_per_second": 1.052,
      "eval_steps_per_second": 1.052,
      "step": 187
    },
    {
      "epoch": 0.34,
      "mmlu_eval_accuracy": 0.731892294851104,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.7857142857142857,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.896551724137931,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.7272727272727273,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6538461538461539,
      "mmlu_eval_accuracy_econometrics": 0.8333333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.7073170731707317,
      "mmlu_eval_accuracy_formal_logic": 0.5714285714285714,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.8125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7441860465116279,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9615384615384616,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.9333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.6956521739130435,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.8076923076923077,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7906976744186046,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.59,
      "mmlu_eval_accuracy_nutrition": 0.7878787878787878,
      "mmlu_eval_accuracy_philosophy": 0.7941176470588235,
      "mmlu_eval_accuracy_prehistory": 0.8285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_law": 0.6294117647058823,
      "mmlu_eval_accuracy_professional_medicine": 0.8064516129032258,
      "mmlu_eval_accuracy_professional_psychology": 0.8115942028985508,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.326305795171384,
      "step": 187
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0001,
      "loss": 1.1133,
      "step": 190
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0001,
      "loss": 1.2485,
      "step": 200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0001,
      "loss": 0.9653,
      "step": 210
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0001,
      "loss": 0.9455,
      "step": 220
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0001,
      "loss": 1.0373,
      "step": 230
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0001,
      "loss": 1.1425,
      "step": 240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0001,
      "loss": 1.3136,
      "step": 250
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0001,
      "loss": 0.8695,
      "step": 260
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0001,
      "loss": 0.872,
      "step": 270
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0001,
      "loss": 1.0152,
      "step": 280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0001,
      "loss": 1.1309,
      "step": 290
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0001,
      "loss": 1.267,
      "step": 300
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0001,
      "loss": 0.9249,
      "step": 310
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0001,
      "loss": 0.9148,
      "step": 320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001,
      "loss": 0.9864,
      "step": 330
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001,
      "loss": 1.2312,
      "step": 340
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001,
      "loss": 1.2354,
      "step": 350
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001,
      "loss": 0.9126,
      "step": 360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001,
      "loss": 0.9213,
      "step": 370
    },
    {
      "epoch": 0.68,
      "eval_loss": 1.0163359642028809,
      "eval_runtime": 948.1151,
      "eval_samples_per_second": 1.055,
      "eval_steps_per_second": 1.055,
      "step": 374
    },
    {
      "epoch": 0.68,
      "mmlu_eval_accuracy": 0.7395476061435284,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.7857142857142857,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.896551724137931,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.7272727272727273,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.8333333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.6829268292682927,
      "mmlu_eval_accuracy_formal_logic": 0.6428571428571429,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.8125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7674418604651163,
      "mmlu_eval_accuracy_high_school_mathematics": 0.41379310344827586,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9615384615384616,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.95,
      "mmlu_eval_accuracy_high_school_statistics": 0.7391304347826086,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.7692307692307693,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7790697674418605,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.57,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7941176470588235,
      "mmlu_eval_accuracy_prehistory": 0.8571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_law": 0.6411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.8387096774193549,
      "mmlu_eval_accuracy_professional_psychology": 0.8115942028985508,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.2796503596061355,
      "step": 374
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001,
      "loss": 0.9737,
      "step": 380
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001,
      "loss": 1.157,
      "step": 390
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001,
      "loss": 1.2106,
      "step": 400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001,
      "loss": 0.8687,
      "step": 410
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001,
      "loss": 0.8742,
      "step": 420
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001,
      "loss": 0.9901,
      "step": 430
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001,
      "loss": 1.2238,
      "step": 440
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001,
      "loss": 1.2604,
      "step": 450
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001,
      "loss": 0.8756,
      "step": 460
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001,
      "loss": 0.8683,
      "step": 470
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0001,
      "loss": 0.9824,
      "step": 480
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0001,
      "loss": 1.1574,
      "step": 490
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0001,
      "loss": 1.2687,
      "step": 500
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0001,
      "loss": 0.8657,
      "step": 510
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0001,
      "loss": 0.9207,
      "step": 520
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0001,
      "loss": 1.012,
      "step": 530
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0001,
      "loss": 1.1517,
      "step": 540
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0001,
      "loss": 1.1654,
      "step": 550
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0001,
      "loss": 0.8931,
      "step": 560
    },
    {
      "epoch": 1.02,
      "eval_loss": 1.0150845050811768,
      "eval_runtime": 949.8392,
      "eval_samples_per_second": 1.053,
      "eval_steps_per_second": 1.053,
      "step": 561
    },
    {
      "epoch": 1.02,
      "mmlu_eval_accuracy": 0.7346397374699287,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.7142857142857143,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.8181818181818182,
      "mmlu_eval_accuracy_clinical_knowledge": 0.8620689655172413,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.7272727272727273,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.8333333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.6829268292682927,
      "mmlu_eval_accuracy_formal_logic": 0.6428571428571429,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.84375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7674418604651163,
      "mmlu_eval_accuracy_high_school_mathematics": 0.41379310344827586,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.9666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.7391304347826086,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.8076923076923077,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.56,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.7941176470588235,
      "mmlu_eval_accuracy_prehistory": 0.8571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_law": 0.6294117647058823,
      "mmlu_eval_accuracy_professional_medicine": 0.8064516129032258,
      "mmlu_eval_accuracy_professional_psychology": 0.8260869565217391,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.186674291658526,
      "step": 561
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0001,
      "loss": 0.8507,
      "step": 570
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0001,
      "loss": 0.9164,
      "step": 580
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0001,
      "loss": 1.0908,
      "step": 590
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0001,
      "loss": 1.0431,
      "step": 600
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0001,
      "loss": 0.8567,
      "step": 610
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0001,
      "loss": 0.8818,
      "step": 620
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0001,
      "loss": 0.9499,
      "step": 630
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0001,
      "loss": 1.0437,
      "step": 640
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.0001,
      "loss": 1.0487,
      "step": 650
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.0001,
      "loss": 0.8405,
      "step": 660
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.0001,
      "loss": 0.8818,
      "step": 670
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.0001,
      "loss": 0.9619,
      "step": 680
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.0001,
      "loss": 1.0753,
      "step": 690
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.0001,
      "loss": 1.0218,
      "step": 700
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.0001,
      "loss": 0.8763,
      "step": 710
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.0001,
      "loss": 0.8789,
      "step": 720
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.0001,
      "loss": 0.8631,
      "step": 730
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.0001,
      "loss": 0.9846,
      "step": 740
    },
    {
      "epoch": 1.36,
      "eval_loss": 1.0305067300796509,
      "eval_runtime": 948.7106,
      "eval_samples_per_second": 1.054,
      "eval_steps_per_second": 1.054,
      "step": 748
    },
    {
      "epoch": 1.36,
      "mmlu_eval_accuracy": 0.7324229372189777,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 1.0,
      "mmlu_eval_accuracy_clinical_knowledge": 0.896551724137931,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.8636363636363636,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.8333333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.7073170731707317,
      "mmlu_eval_accuracy_formal_logic": 0.6428571428571429,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.8125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7674418604651163,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9615384615384616,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.9333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.9545454545454546,
      "mmlu_eval_accuracy_high_school_world_history": 0.7692307692307693,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.96,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.57,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7941176470588235,
      "mmlu_eval_accuracy_prehistory": 0.8285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_law": 0.6411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.8064516129032258,
      "mmlu_eval_accuracy_professional_psychology": 0.8260869565217391,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.2988067958029479,
      "step": 748
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.0001,
      "loss": 1.0735,
      "step": 750
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.0001,
      "loss": 0.9066,
      "step": 760
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.0001,
      "loss": 0.8716,
      "step": 770
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.0001,
      "loss": 0.9144,
      "step": 780
    },
    {
      "epoch": 1.43,
      "learning_rate": 0.0001,
      "loss": 1.0338,
      "step": 790
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.0001,
      "loss": 1.0275,
      "step": 800
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.0001,
      "loss": 0.8382,
      "step": 810
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.0001,
      "loss": 0.8489,
      "step": 820
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.0001,
      "loss": 0.8931,
      "step": 830
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.0001,
      "loss": 1.0515,
      "step": 840
    },
    {
      "epoch": 1.54,
      "learning_rate": 0.0001,
      "loss": 1.0965,
      "step": 850
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.0001,
      "loss": 0.8928,
      "step": 860
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.0001,
      "loss": 0.8608,
      "step": 870
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.0001,
      "loss": 0.8831,
      "step": 880
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.0001,
      "loss": 1.0253,
      "step": 890
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.0001,
      "loss": 0.9905,
      "step": 900
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.0001,
      "loss": 0.8487,
      "step": 910
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.0001,
      "loss": 0.8568,
      "step": 920
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.0001,
      "loss": 0.9047,
      "step": 930
    },
    {
      "epoch": 1.7,
      "eval_loss": 1.0250624418258667,
      "eval_runtime": 946.4035,
      "eval_samples_per_second": 1.057,
      "eval_steps_per_second": 1.057,
      "step": 935
    },
    {
      "epoch": 1.7,
      "mmlu_eval_accuracy": 0.7288948695878031,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.9090909090909091,
      "mmlu_eval_accuracy_clinical_knowledge": 0.8620689655172413,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.7272727272727273,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.8333333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.6097560975609756,
      "mmlu_eval_accuracy_formal_logic": 0.6428571428571429,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.84375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7906976744186046,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.9333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.7391304347826086,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.7692307692307693,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.96,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7790697674418605,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.57,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.8235294117647058,
      "mmlu_eval_accuracy_prehistory": 0.8857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_law": 0.6235294117647059,
      "mmlu_eval_accuracy_professional_medicine": 0.8387096774193549,
      "mmlu_eval_accuracy_professional_psychology": 0.8115942028985508,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.243813282909306,
      "step": 935
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.0001,
      "loss": 1.0174,
      "step": 940
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.0001,
      "loss": 1.0302,
      "step": 950
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.0001,
      "loss": 0.8799,
      "step": 960
    },
    {
      "epoch": 1.76,
      "learning_rate": 0.0001,
      "loss": 0.8447,
      "step": 970
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.0001,
      "loss": 0.9053,
      "step": 980
    },
    {
      "epoch": 1.8,
      "learning_rate": 0.0001,
      "loss": 1.0331,
      "step": 990
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.0001,
      "loss": 1.0412,
      "step": 1000
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.0001,
      "loss": 0.8753,
      "step": 1010
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.0001,
      "loss": 0.8744,
      "step": 1020
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.0001,
      "loss": 0.8899,
      "step": 1030
    },
    {
      "epoch": 1.89,
      "learning_rate": 0.0001,
      "loss": 1.0053,
      "step": 1040
    },
    {
      "epoch": 1.9,
      "learning_rate": 0.0001,
      "loss": 1.0127,
      "step": 1050
    },
    {
      "epoch": 1.92,
      "learning_rate": 0.0001,
      "loss": 0.8023,
      "step": 1060
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.0001,
      "loss": 0.8349,
      "step": 1070
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.0001,
      "loss": 0.9742,
      "step": 1080
    },
    {
      "epoch": 1.98,
      "learning_rate": 0.0001,
      "loss": 1.0971,
      "step": 1090
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0001,
      "loss": 1.0728,
      "step": 1100
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0001,
      "loss": 0.7724,
      "step": 1110
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0001,
      "loss": 0.7675,
      "step": 1120
    },
    {
      "epoch": 2.03,
      "eval_loss": 1.052681565284729,
      "eval_runtime": 942.0722,
      "eval_samples_per_second": 1.061,
      "eval_steps_per_second": 1.061,
      "step": 1122
    },
    {
      "epoch": 2.03,
      "mmlu_eval_accuracy": 0.7373981967098951,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.9090909090909091,
      "mmlu_eval_accuracy_clinical_knowledge": 0.896551724137931,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.8333333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5853658536585366,
      "mmlu_eval_accuracy_formal_logic": 0.7142857142857143,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.84375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7674418604651163,
      "mmlu_eval_accuracy_high_school_mathematics": 0.41379310344827586,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.95,
      "mmlu_eval_accuracy_high_school_statistics": 0.782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.7692307692307693,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7906976744186046,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.62,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.8235294117647058,
      "mmlu_eval_accuracy_prehistory": 0.8857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_law": 0.6294117647058823,
      "mmlu_eval_accuracy_professional_medicine": 0.8709677419354839,
      "mmlu_eval_accuracy_professional_psychology": 0.782608695652174,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.2340081441760609,
      "step": 1122
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0001,
      "loss": 0.7194,
      "step": 1130
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0001,
      "loss": 0.8236,
      "step": 1140
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0001,
      "loss": 0.6652,
      "step": 1150
    },
    {
      "epoch": 2.1,
      "learning_rate": 0.0001,
      "loss": 0.7177,
      "step": 1160
    },
    {
      "epoch": 2.12,
      "learning_rate": 0.0001,
      "loss": 0.7788,
      "step": 1170
    },
    {
      "epoch": 2.14,
      "learning_rate": 0.0001,
      "loss": 0.8117,
      "step": 1180
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.0001,
      "loss": 0.8145,
      "step": 1190
    },
    {
      "epoch": 2.18,
      "learning_rate": 0.0001,
      "loss": 0.6984,
      "step": 1200
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.0001,
      "loss": 0.7011,
      "step": 1210
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.0001,
      "loss": 0.769,
      "step": 1220
    },
    {
      "epoch": 2.23,
      "learning_rate": 0.0001,
      "loss": 0.7705,
      "step": 1230
    },
    {
      "epoch": 2.25,
      "learning_rate": 0.0001,
      "loss": 0.8066,
      "step": 1240
    },
    {
      "epoch": 2.27,
      "learning_rate": 0.0001,
      "loss": 0.6622,
      "step": 1250
    },
    {
      "epoch": 2.29,
      "learning_rate": 0.0001,
      "loss": 0.6641,
      "step": 1260
    },
    {
      "epoch": 2.3,
      "learning_rate": 0.0001,
      "loss": 0.7239,
      "step": 1270
    },
    {
      "epoch": 2.32,
      "learning_rate": 0.0001,
      "loss": 0.7618,
      "step": 1280
    },
    {
      "epoch": 2.34,
      "learning_rate": 0.0001,
      "loss": 0.7845,
      "step": 1290
    },
    {
      "epoch": 2.36,
      "learning_rate": 0.0001,
      "loss": 0.719,
      "step": 1300
    },
    {
      "epoch": 2.37,
      "eval_loss": 1.1104822158813477,
      "eval_runtime": 948.1299,
      "eval_samples_per_second": 1.055,
      "eval_steps_per_second": 1.055,
      "step": 1309
    },
    {
      "epoch": 2.37,
      "mmlu_eval_accuracy": 0.7369285730399766,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 1.0,
      "mmlu_eval_accuracy_clinical_knowledge": 0.8275862068965517,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.8636363636363636,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.6923076923076923,
      "mmlu_eval_accuracy_econometrics": 0.75,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.6341463414634146,
      "mmlu_eval_accuracy_formal_logic": 0.7857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.84375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.8888888888888888,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7906976744186046,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.95,
      "mmlu_eval_accuracy_high_school_statistics": 0.6956521739130435,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7906976744186046,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.57,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.8529411764705882,
      "mmlu_eval_accuracy_prehistory": 0.8571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_law": 0.6058823529411764,
      "mmlu_eval_accuracy_professional_medicine": 0.8709677419354839,
      "mmlu_eval_accuracy_professional_psychology": 0.7681159420289855,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.8148148148148148,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.0866562834095908,
      "step": 1309
    },
    {
      "epoch": 2.38,
      "learning_rate": 0.0001,
      "loss": 0.7093,
      "step": 1310
    },
    {
      "epoch": 2.39,
      "learning_rate": 0.0001,
      "loss": 0.7684,
      "step": 1320
    },
    {
      "epoch": 2.41,
      "learning_rate": 0.0001,
      "loss": 0.7501,
      "step": 1330
    },
    {
      "epoch": 2.43,
      "learning_rate": 0.0001,
      "loss": 0.8043,
      "step": 1340
    },
    {
      "epoch": 2.45,
      "learning_rate": 0.0001,
      "loss": 0.6927,
      "step": 1350
    },
    {
      "epoch": 2.47,
      "learning_rate": 0.0001,
      "loss": 0.7278,
      "step": 1360
    },
    {
      "epoch": 2.48,
      "learning_rate": 0.0001,
      "loss": 0.8095,
      "step": 1370
    },
    {
      "epoch": 2.5,
      "learning_rate": 0.0001,
      "loss": 0.7463,
      "step": 1380
    },
    {
      "epoch": 2.52,
      "learning_rate": 0.0001,
      "loss": 0.7707,
      "step": 1390
    },
    {
      "epoch": 2.54,
      "learning_rate": 0.0001,
      "loss": 0.7152,
      "step": 1400
    },
    {
      "epoch": 2.56,
      "learning_rate": 0.0001,
      "loss": 0.687,
      "step": 1410
    },
    {
      "epoch": 2.58,
      "learning_rate": 0.0001,
      "loss": 0.7529,
      "step": 1420
    },
    {
      "epoch": 2.59,
      "learning_rate": 0.0001,
      "loss": 0.7565,
      "step": 1430
    },
    {
      "epoch": 2.61,
      "learning_rate": 0.0001,
      "loss": 0.8066,
      "step": 1440
    },
    {
      "epoch": 2.63,
      "learning_rate": 0.0001,
      "loss": 0.7623,
      "step": 1450
    },
    {
      "epoch": 2.65,
      "learning_rate": 0.0001,
      "loss": 0.6947,
      "step": 1460
    },
    {
      "epoch": 2.67,
      "learning_rate": 0.0001,
      "loss": 0.7756,
      "step": 1470
    },
    {
      "epoch": 2.68,
      "learning_rate": 0.0001,
      "loss": 0.8453,
      "step": 1480
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.0001,
      "loss": 0.8306,
      "step": 1490
    },
    {
      "epoch": 2.71,
      "eval_loss": 1.100826621055603,
      "eval_runtime": 940.4488,
      "eval_samples_per_second": 1.063,
      "eval_steps_per_second": 1.063,
      "step": 1496
    },
    {
      "epoch": 2.71,
      "mmlu_eval_accuracy": 0.7363077307176445,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.9090909090909091,
      "mmlu_eval_accuracy_clinical_knowledge": 0.8620689655172413,
      "mmlu_eval_accuracy_college_biology": 0.8125,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.75,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.6585365853658537,
      "mmlu_eval_accuracy_formal_logic": 0.6428571428571429,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "mmlu_eval_accuracy_high_school_biology": 0.78125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9545454545454546,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7906976744186046,
      "mmlu_eval_accuracy_high_school_mathematics": 0.4827586206896552,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9615384615384616,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.9166666666666666,
      "mmlu_eval_accuracy_high_school_statistics": 0.6521739130434783,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.8076923076923077,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.75,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7906976744186046,
      "mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
      "mmlu_eval_accuracy_moral_scenarios": 0.64,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.8235294117647058,
      "mmlu_eval_accuracy_prehistory": 0.8571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.7096774193548387,
      "mmlu_eval_accuracy_professional_law": 0.6176470588235294,
      "mmlu_eval_accuracy_professional_medicine": 0.8709677419354839,
      "mmlu_eval_accuracy_professional_psychology": 0.7971014492753623,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.8518518518518519,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.2313211129857853,
      "step": 1496
    },
    {
      "epoch": 2.72,
      "learning_rate": 0.0001,
      "loss": 0.6937,
      "step": 1500
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.0001,
      "loss": 0.6997,
      "step": 1510
    },
    {
      "epoch": 2.76,
      "learning_rate": 0.0001,
      "loss": 0.7588,
      "step": 1520
    },
    {
      "epoch": 2.77,
      "learning_rate": 0.0001,
      "loss": 0.7731,
      "step": 1530
    },
    {
      "epoch": 2.79,
      "learning_rate": 0.0001,
      "loss": 0.7914,
      "step": 1540
    },
    {
      "epoch": 2.81,
      "learning_rate": 0.0001,
      "loss": 0.7175,
      "step": 1550
    },
    {
      "epoch": 2.83,
      "learning_rate": 0.0001,
      "loss": 0.7046,
      "step": 1560
    },
    {
      "epoch": 2.85,
      "learning_rate": 0.0001,
      "loss": 0.7597,
      "step": 1570
    },
    {
      "epoch": 2.87,
      "learning_rate": 0.0001,
      "loss": 0.7932,
      "step": 1580
    },
    {
      "epoch": 2.88,
      "learning_rate": 0.0001,
      "loss": 0.8059,
      "step": 1590
    },
    {
      "epoch": 2.9,
      "learning_rate": 0.0001,
      "loss": 0.7258,
      "step": 1600
    },
    {
      "epoch": 2.92,
      "learning_rate": 0.0001,
      "loss": 0.7486,
      "step": 1610
    },
    {
      "epoch": 2.94,
      "learning_rate": 0.0001,
      "loss": 0.7233,
      "step": 1620
    },
    {
      "epoch": 2.96,
      "learning_rate": 0.0001,
      "loss": 0.7945,
      "step": 1630
    },
    {
      "epoch": 2.97,
      "learning_rate": 0.0001,
      "loss": 0.8324,
      "step": 1640
    },
    {
      "epoch": 2.99,
      "learning_rate": 0.0001,
      "loss": 0.7294,
      "step": 1650
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0001,
      "loss": 0.6117,
      "step": 1660
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0001,
      "loss": 0.6464,
      "step": 1670
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0001,
      "loss": 0.6156,
      "step": 1680
    },
    {
      "epoch": 3.05,
      "eval_loss": 1.1478718519210815,
      "eval_runtime": 932.4225,
      "eval_samples_per_second": 1.072,
      "eval_steps_per_second": 1.072,
      "step": 1683
    },
    {
      "epoch": 3.05,
      "mmlu_eval_accuracy": 0.745366643285036,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 1.0,
      "mmlu_eval_accuracy_clinical_knowledge": 0.8620689655172413,
      "mmlu_eval_accuracy_college_biology": 0.875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.7272727272727273,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6538461538461539,
      "mmlu_eval_accuracy_econometrics": 0.75,
      "mmlu_eval_accuracy_electrical_engineering": 0.875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.7073170731707317,
      "mmlu_eval_accuracy_formal_logic": 0.7857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.8,
      "mmlu_eval_accuracy_high_school_biology": 0.8125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5454545454545454,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.7674418604651163,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.9230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.9333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.6521739130434783,
      "mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_world_history": 0.8076923076923077,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.75,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.96,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7906976744186046,
      "mmlu_eval_accuracy_moral_disputes": 0.7894736842105263,
      "mmlu_eval_accuracy_moral_scenarios": 0.61,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.8571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_law": 0.6176470588235294,
      "mmlu_eval_accuracy_professional_medicine": 0.9032258064516129,
      "mmlu_eval_accuracy_professional_psychology": 0.7391304347826086,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.7777777777777778,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4050485734687297,
      "step": 1683
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0001,
      "loss": 0.5237,
      "step": 1690
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0001,
      "loss": 0.3516,
      "step": 1700
    },
    {
      "epoch": 3.1,
      "learning_rate": 0.0001,
      "loss": 0.4976,
      "step": 1710
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.0001,
      "loss": 0.6535,
      "step": 1720
    },
    {
      "epoch": 3.14,
      "learning_rate": 0.0001,
      "loss": 0.5926,
      "step": 1730
    },
    {
      "epoch": 3.16,
      "learning_rate": 0.0001,
      "loss": 0.5476,
      "step": 1740
    },
    {
      "epoch": 3.17,
      "learning_rate": 0.0001,
      "loss": 0.368,
      "step": 1750
    },
    {
      "epoch": 3.19,
      "learning_rate": 0.0001,
      "loss": 0.5043,
      "step": 1760
    },
    {
      "epoch": 3.21,
      "learning_rate": 0.0001,
      "loss": 0.5907,
      "step": 1770
    },
    {
      "epoch": 3.23,
      "learning_rate": 0.0001,
      "loss": 0.5609,
      "step": 1780
    },
    {
      "epoch": 3.25,
      "learning_rate": 0.0001,
      "loss": 0.5272,
      "step": 1790
    },
    {
      "epoch": 3.26,
      "learning_rate": 0.0001,
      "loss": 0.3672,
      "step": 1800
    }
  ],
  "max_steps": 1875,
  "num_train_epochs": 4,
  "total_flos": 2.0859408156226683e+18,
  "trial_name": null,
  "trial_params": null
}