{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8254041040926287,
  "global_step": 1800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.7825,
      "step": 10
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7565,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7408,
      "step": 30
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7496,
      "step": 40
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7424,
      "step": 50
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7112,
      "step": 60
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7409,
      "step": 70
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7646,
      "step": 80
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7129,
      "step": 90
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7671,
      "step": 100
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7166,
      "step": 110
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7113,
      "step": 120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6682,
      "step": 130
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7644,
      "step": 140
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6813,
      "step": 150
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6447,
      "step": 160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 170
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6657,
      "step": 180
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.681,
      "step": 190
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7142,
      "step": 200
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.6789492964744568,
      "eval_runtime": 280.115,
      "eval_samples_per_second": 3.57,
      "eval_steps_per_second": 0.892,
      "step": 200
    },
    {
      "epoch": 0.09,
      "mmlu_eval_accuracy": 0.4601645000494307,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.2727272727272727,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3170731707317073,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "mmlu_eval_accuracy_high_school_biology": 0.34375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.3488372093023256,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.75,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.5454545454545454,
      "mmlu_eval_accuracy_marketing": 0.68,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.24,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.42028985507246375,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.5909090909090909,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.3888888888888889,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.9642877595465115,
      "step": 200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6854,
      "step": 210
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7256,
      "step": 220
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7505,
      "step": 230
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.618,
      "step": 240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6726,
      "step": 250
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6882,
      "step": 260
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6982,
      "step": 270
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.661,
      "step": 280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.699,
      "step": 290
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6867,
      "step": 300
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7127,
      "step": 310
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7141,
      "step": 320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6483,
      "step": 330
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6532,
      "step": 340
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6474,
      "step": 350
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6728,
      "step": 360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6736,
      "step": 370
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7164,
      "step": 380
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6844,
      "step": 390
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6797,
      "step": 400
    },
    {
      "epoch": 0.18,
      "eval_loss": 0.6707000732421875,
      "eval_runtime": 280.204,
      "eval_samples_per_second": 3.569,
      "eval_steps_per_second": 0.892,
      "step": 400
    },
    {
      "epoch": 0.18,
      "mmlu_eval_accuracy": 0.4521886129310749,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.0,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.2727272727272727,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.35714285714285715,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.6923076923076923,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.45454545454545453,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
      "mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
      "mmlu_eval_accuracy_moral_scenarios": 0.24,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6363636363636364,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.92568634446545,
      "step": 400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7059,
      "step": 410
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.79,
      "step": 420
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7714,
      "step": 430
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.644,
      "step": 440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7102,
      "step": 450
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6229,
      "step": 460
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6742,
      "step": 470
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6997,
      "step": 480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6598,
      "step": 490
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7154,
      "step": 500
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6796,
      "step": 510
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6769,
      "step": 520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6663,
      "step": 530
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6758,
      "step": 540
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7022,
      "step": 550
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.706,
      "step": 560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 570
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7077,
      "step": 580
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6352,
      "step": 590
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6203,
      "step": 600
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.664973795413971,
      "eval_runtime": 280.1433,
      "eval_samples_per_second": 3.57,
      "eval_steps_per_second": 0.892,
      "step": 600
    },
    {
      "epoch": 0.28,
      "mmlu_eval_accuracy": 0.44929002048717553,
      "mmlu_eval_accuracy_abstract_algebra": 0.09090909090909091,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.375,
      "mmlu_eval_accuracy_college_chemistry": 0.0,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.2727272727272727,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.21739130434782608,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.5,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.6923076923076923,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.5454545454545454,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.23,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.5909090909090909,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.9583329685985575,
      "step": 600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6652,
      "step": 610
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 620
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6722,
      "step": 630
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7112,
      "step": 640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6976,
      "step": 650
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6842,
      "step": 660
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6913,
      "step": 670
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6879,
      "step": 680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6684,
      "step": 690
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7271,
      "step": 700
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6683,
      "step": 710
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6531,
      "step": 720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6948,
      "step": 730
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.68,
      "step": 740
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6339,
      "step": 750
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.602,
      "step": 760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6817,
      "step": 770
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6686,
      "step": 780
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6869,
      "step": 790
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.655,
      "step": 800
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.6593752503395081,
      "eval_runtime": 280.0748,
      "eval_samples_per_second": 3.57,
      "eval_steps_per_second": 0.893,
      "step": 800
    },
    {
      "epoch": 0.37,
      "mmlu_eval_accuracy": 0.46269190662697585,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.5,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.0,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3170731707317073,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.34375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.3488372093023256,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.7,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_world_history": 0.5,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.5454545454545454,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
      "mmlu_eval_accuracy_moral_disputes": 0.47368421052631576,
      "mmlu_eval_accuracy_moral_scenarios": 0.22,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.5,
      "mmlu_eval_accuracy_prehistory": 0.45714285714285713,
      "mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
      "mmlu_eval_accuracy_professional_law": 0.32941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.6363636363636364,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.1274755495957856,
      "step": 800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7119,
      "step": 810
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6581,
      "step": 820
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6475,
      "step": 830
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6944,
      "step": 840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6469,
      "step": 850
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7231,
      "step": 860
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6182,
      "step": 870
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7032,
      "step": 880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6291,
      "step": 890
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6845,
      "step": 900
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7016,
      "step": 910
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6921,
      "step": 920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6908,
      "step": 930
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7402,
      "step": 940
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6553,
      "step": 950
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6525,
      "step": 960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7003,
      "step": 970
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 980
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6757,
      "step": 990
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6803,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.656200647354126,
      "eval_runtime": 279.9415,
      "eval_samples_per_second": 3.572,
      "eval_steps_per_second": 0.893,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "mmlu_eval_accuracy": 0.46074490234022447,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.5,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.34615384615384615,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.6363636363636364,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.23,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.5,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.34705882352941175,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.5909090909090909,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.016490816680943,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6489,
      "step": 1010
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5919,
      "step": 1020
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.721,
      "step": 1030
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6696,
      "step": 1040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6976,
      "step": 1050
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6665,
      "step": 1060
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6343,
      "step": 1070
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6133,
      "step": 1080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6259,
      "step": 1090
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6446,
      "step": 1100
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.705,
      "step": 1110
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6445,
      "step": 1120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7156,
      "step": 1130
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6842,
      "step": 1140
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6528,
      "step": 1150
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7015,
      "step": 1160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.679,
      "step": 1170
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 1180
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6369,
      "step": 1190
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6512,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "eval_loss": 0.6535513997077942,
      "eval_runtime": 279.9064,
      "eval_samples_per_second": 3.573,
      "eval_steps_per_second": 0.893,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "mmlu_eval_accuracy": 0.4563086650055728,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.0,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.3181818181818182,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.3488372093023256,
      "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.3076923076923077,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.5454545454545454,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
      "mmlu_eval_accuracy_moral_disputes": 0.3684210526315789,
      "mmlu_eval_accuracy_moral_scenarios": 0.23,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.34705882352941175,
      "mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.5909090909090909,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.0143195821006366,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6658,
      "step": 1210
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7283,
      "step": 1220
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6944,
      "step": 1230
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6326,
      "step": 1240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6574,
      "step": 1250
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6239,
      "step": 1260
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7219,
      "step": 1270
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6704,
      "step": 1280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6708,
      "step": 1290
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6862,
      "step": 1300
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6771,
      "step": 1310
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.661,
      "step": 1320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6742,
      "step": 1330
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6734,
      "step": 1340
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6798,
      "step": 1350
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6152,
      "step": 1360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.671,
      "step": 1370
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7562,
      "step": 1380
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7571,
      "step": 1390
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6926,
      "step": 1400
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.6509103178977966,
      "eval_runtime": 280.3444,
      "eval_samples_per_second": 3.567,
      "eval_steps_per_second": 0.892,
      "step": 1400
    },
    {
      "epoch": 0.64,
      "mmlu_eval_accuracy": 0.4563542680952826,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.375,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
      "mmlu_eval_accuracy_formal_logic": 0.35714285714285715,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.4444444444444444,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.3488372093023256,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.3076923076923077,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_world_history": 0.5,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.5454545454545454,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
      "mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
      "mmlu_eval_accuracy_moral_scenarios": 0.23,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.5294117647058824,
      "mmlu_eval_accuracy_prehistory": 0.4,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.5909090909090909,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.092854514290085,
      "step": 1400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6375,
      "step": 1410
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6123,
      "step": 1420
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6628,
      "step": 1430
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7702,
      "step": 1440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7076,
      "step": 1450
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.704,
      "step": 1460
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6406,
      "step": 1470
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.669,
      "step": 1480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6253,
      "step": 1490
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6544,
      "step": 1500
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6018,
      "step": 1510
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6673,
      "step": 1520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.705,
      "step": 1530
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6875,
      "step": 1540
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6864,
      "step": 1550
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6465,
      "step": 1560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6714,
      "step": 1570
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.627,
      "step": 1580
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5958,
      "step": 1590
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.62,
      "step": 1600
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.6470592617988586,
      "eval_runtime": 280.3872,
      "eval_samples_per_second": 3.566,
      "eval_steps_per_second": 0.892,
      "step": 1600
    },
    {
      "epoch": 0.73,
      "mmlu_eval_accuracy": 0.448089860540604,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
      "mmlu_eval_accuracy_college_biology": 0.3125,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.3181818181818182,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.2727272727272727,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3170731707317073,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.34375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.3076923076923077,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.45454545454545453,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
      "mmlu_eval_accuracy_moral_disputes": 0.47368421052631576,
      "mmlu_eval_accuracy_moral_scenarios": 0.23,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.5294117647058824,
      "mmlu_eval_accuracy_prehistory": 0.4,
      "mmlu_eval_accuracy_professional_accounting": 0.3225806451612903,
      "mmlu_eval_accuracy_professional_law": 0.3411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_psychology": 0.42028985507246375,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6363636363636364,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.9696583512090827,
      "step": 1600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.722,
      "step": 1610
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6225,
      "step": 1620
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7043,
      "step": 1630
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6671,
      "step": 1640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6424,
      "step": 1650
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.641,
      "step": 1660
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6494,
      "step": 1670
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6845,
      "step": 1680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6646,
      "step": 1690
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6706,
      "step": 1700
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6993,
      "step": 1710
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6672,
      "step": 1720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6114,
      "step": 1730
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7194,
      "step": 1740
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6669,
      "step": 1750
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6482,
      "step": 1760
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.681,
      "step": 1770
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.718,
      "step": 1780
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6948,
      "step": 1790
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6835,
      "step": 1800
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.6446049809455872,
      "eval_runtime": 280.3145,
      "eval_samples_per_second": 3.567,
      "eval_steps_per_second": 0.892,
      "step": 1800
    },
    {
      "epoch": 0.83,
      "mmlu_eval_accuracy": 0.45510585655253843,
      "mmlu_eval_accuracy_abstract_algebra": 0.09090909090909091,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
      "mmlu_eval_accuracy_college_biology": 0.375,
      "mmlu_eval_accuracy_college_chemistry": 0.0,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.2727272727272727,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
      "mmlu_eval_accuracy_formal_logic": 0.35714285714285715,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.34375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.3076923076923077,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7,
      "mmlu_eval_accuracy_high_school_statistics": 0.21739130434782608,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.5,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.6923076923076923,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.45454545454545453,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6395348837209303,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.24,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.5294117647058824,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.3225806451612903,
      "mmlu_eval_accuracy_professional_law": 0.3235294117647059,
      "mmlu_eval_accuracy_professional_medicine": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6363636363636364,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 0.9942787007936921,
      "step": 1800
    }
  ],
  "max_steps": 5000,
  "num_train_epochs": 3,
  "total_flos": 5.1765837924836966e+17,
  "trial_name": null,
  "trial_params": null
}