{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.7110536522301228,
  "global_step": 2200,
  "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": 0.9529,
      "step": 10
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7173,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7776,
      "step": 40
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6779,
      "step": 50
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7011,
      "step": 60
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7156,
      "step": 70
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7418,
      "step": 80
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7465,
      "step": 90
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7595,
      "step": 100
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6816,
      "step": 110
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7981,
      "step": 120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6494,
      "step": 130
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7423,
      "step": 140
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7555,
      "step": 150
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5799,
      "step": 160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.699,
      "step": 170
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7154,
      "step": 180
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6096,
      "step": 190
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5713,
      "step": 200
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.6557502150535583,
      "eval_runtime": 92.7109,
      "eval_samples_per_second": 10.786,
      "eval_steps_per_second": 5.393,
      "step": 200
    },
    {
      "epoch": 0.06,
      "mmlu_eval_accuracy": 0.47294392264572244,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.3181818181818182,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "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": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.35714285714285715,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.40625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.3333333333333333,
      "mmlu_eval_accuracy_high_school_european_history": 0.4444444444444444,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "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": 0.7166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.5,
      "mmlu_eval_accuracy_high_school_world_history": 0.5769230769230769,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.18181818181818182,
      "mmlu_eval_accuracy_logical_fallacies": 0.5,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.5454545454545454,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6395348837209303,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3176470588235294,
      "mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8432865626915472,
      "step": 200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6339,
      "step": 210
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.701,
      "step": 220
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.689,
      "step": 230
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6955,
      "step": 240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6124,
      "step": 250
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6648,
      "step": 260
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7364,
      "step": 270
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.558,
      "step": 280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7089,
      "step": 290
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5691,
      "step": 300
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7079,
      "step": 310
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6558,
      "step": 320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5887,
      "step": 330
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.704,
      "step": 340
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.638,
      "step": 350
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.708,
      "step": 360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6203,
      "step": 370
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6288,
      "step": 380
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6725,
      "step": 390
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6849,
      "step": 400
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.6309903860092163,
      "eval_runtime": 92.7675,
      "eval_samples_per_second": 10.78,
      "eval_steps_per_second": 5.39,
      "step": 400
    },
    {
      "epoch": 0.13,
      "mmlu_eval_accuracy": 0.46759845160702646,
      "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": 0.375,
      "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": 0.4090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "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": 0.2926829268292683,
      "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": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
      "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.23076923076923078,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.75,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_world_history": 0.5,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.6363636363636364,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.5,
      "mmlu_eval_accuracy_prehistory": 0.45714285714285713,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3176470588235294,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8323084239996289,
      "step": 400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5261,
      "step": 410
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6614,
      "step": 420
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6213,
      "step": 430
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7206,
      "step": 440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6944,
      "step": 450
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.675,
      "step": 460
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 470
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5778,
      "step": 480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5583,
      "step": 490
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7171,
      "step": 500
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.728,
      "step": 510
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6037,
      "step": 520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6598,
      "step": 530
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6913,
      "step": 540
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6092,
      "step": 550
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.589,
      "step": 560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5608,
      "step": 570
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6251,
      "step": 580
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6527,
      "step": 590
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6518,
      "step": 600
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.618711531162262,
      "eval_runtime": 92.7537,
      "eval_samples_per_second": 10.781,
      "eval_steps_per_second": 5.391,
      "step": 600
    },
    {
      "epoch": 0.19,
      "mmlu_eval_accuracy": 0.4660645174643213,
      "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": 0.4827586206896552,
      "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": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "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": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.46875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.37209302325581395,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.7333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_world_history": 0.5769230769230769,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
      "mmlu_eval_accuracy_logical_fallacies": 0.5,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.6363636363636364,
      "mmlu_eval_accuracy_marketing": 0.68,
      "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": 0.26,
      "mmlu_eval_accuracy_nutrition": 0.5454545454545454,
      "mmlu_eval_accuracy_philosophy": 0.5,
      "mmlu_eval_accuracy_prehistory": 0.4,
      "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
      "mmlu_eval_accuracy_professional_law": 0.32941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_psychology": 0.42028985507246375,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.8410176051775718,
      "step": 600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6876,
      "step": 610
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7174,
      "step": 620
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6002,
      "step": 630
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7426,
      "step": 640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5213,
      "step": 650
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.574,
      "step": 660
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6669,
      "step": 670
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5847,
      "step": 680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7233,
      "step": 690
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6199,
      "step": 700
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5977,
      "step": 710
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6853,
      "step": 720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5777,
      "step": 730
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5943,
      "step": 740
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6323,
      "step": 750
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6367,
      "step": 760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6367,
      "step": 770
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6197,
      "step": 780
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5639,
      "step": 790
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6281,
      "step": 800
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.6189825534820557,
      "eval_runtime": 92.99,
      "eval_samples_per_second": 10.754,
      "eval_steps_per_second": 5.377,
      "step": 800
    },
    {
      "epoch": 0.26,
      "mmlu_eval_accuracy": 0.48167538061853127,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
      "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": 0.4090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.08333333333333333,
      "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": 0.40625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.4444444444444444,
      "mmlu_eval_accuracy_high_school_european_history": 0.3888888888888889,
      "mmlu_eval_accuracy_high_school_geography": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
      "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": 0.7666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.5769230769230769,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "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": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.27,
      "mmlu_eval_accuracy_nutrition": 0.5454545454545454,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3352941176470588,
      "mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.9606302572954105,
      "step": 800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.637,
      "step": 810
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6902,
      "step": 820
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5872,
      "step": 830
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6115,
      "step": 840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 850
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5977,
      "step": 860
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5873,
      "step": 870
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6776,
      "step": 880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6804,
      "step": 890
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5914,
      "step": 900
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6313,
      "step": 910
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5869,
      "step": 920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.519,
      "step": 930
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4877,
      "step": 940
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5931,
      "step": 950
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6614,
      "step": 960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6287,
      "step": 970
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5139,
      "step": 980
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6392,
      "step": 990
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7001,
      "step": 1000
    },
    {
      "epoch": 0.32,
      "eval_loss": 0.5796898007392883,
      "eval_runtime": 92.7581,
      "eval_samples_per_second": 10.781,
      "eval_steps_per_second": 5.39,
      "step": 1000
    },
    {
      "epoch": 0.32,
      "mmlu_eval_accuracy": 0.47460361123181055,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.7142857142857143,
      "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": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.25,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.4444444444444444,
      "mmlu_eval_accuracy_high_school_european_history": 0.5,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.37209302325581395,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.34615384615384615,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.75,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.5454545454545454,
      "mmlu_eval_accuracy_high_school_world_history": 0.46153846153846156,
      "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": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.6363636363636364,
      "mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.29,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.5588235294117647,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.3225806451612903,
      "mmlu_eval_accuracy_professional_law": 0.3176470588235294,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.8773848874586836,
      "step": 1000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6671,
      "step": 1010
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5942,
      "step": 1020
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6236,
      "step": 1030
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6162,
      "step": 1040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.734,
      "step": 1050
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6108,
      "step": 1060
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6669,
      "step": 1070
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6991,
      "step": 1080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6696,
      "step": 1090
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6511,
      "step": 1100
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6007,
      "step": 1110
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5543,
      "step": 1120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6399,
      "step": 1130
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.569,
      "step": 1140
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6436,
      "step": 1150
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4886,
      "step": 1160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5431,
      "step": 1170
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5516,
      "step": 1180
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5041,
      "step": 1190
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6737,
      "step": 1200
    },
    {
      "epoch": 0.39,
      "eval_loss": 0.5647590160369873,
      "eval_runtime": 92.7185,
      "eval_samples_per_second": 10.785,
      "eval_steps_per_second": 5.393,
      "step": 1200
    },
    {
      "epoch": 0.39,
      "mmlu_eval_accuracy": 0.47729744414466774,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.5,
      "mmlu_eval_accuracy_business_ethics": 0.36363636363636365,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.08333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.4444444444444444,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.46511627906976744,
      "mmlu_eval_accuracy_high_school_mathematics": 0.13793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.5454545454545454,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.27,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
      "mmlu_eval_accuracy_professional_law": 0.3352941176470588,
      "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_psychology": 0.34782608695652173,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.9123951537490825,
      "step": 1200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4926,
      "step": 1210
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5814,
      "step": 1220
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4985,
      "step": 1230
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5966,
      "step": 1240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6493,
      "step": 1250
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5811,
      "step": 1260
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.563,
      "step": 1270
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6206,
      "step": 1280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5657,
      "step": 1290
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6061,
      "step": 1300
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5776,
      "step": 1310
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6702,
      "step": 1320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5688,
      "step": 1330
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5237,
      "step": 1340
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5883,
      "step": 1350
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5206,
      "step": 1360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6948,
      "step": 1370
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5656,
      "step": 1380
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6026,
      "step": 1390
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6319,
      "step": 1400
    },
    {
      "epoch": 0.45,
      "eval_loss": 0.5552607774734497,
      "eval_runtime": 92.7391,
      "eval_samples_per_second": 10.783,
      "eval_steps_per_second": 5.391,
      "step": 1400
    },
    {
      "epoch": 0.45,
      "mmlu_eval_accuracy": 0.4812787579159337,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "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": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.3333333333333333,
      "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.5238095238095238,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.37209302325581395,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "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": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.47058823529411764,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
      "mmlu_eval_accuracy_professional_law": 0.3,
      "mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.4444444444444444,
      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.8453598066862896,
      "step": 1400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4931,
      "step": 1410
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4544,
      "step": 1420
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.563,
      "step": 1430
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4629,
      "step": 1440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4903,
      "step": 1450
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5581,
      "step": 1460
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6571,
      "step": 1470
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5562,
      "step": 1480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4745,
      "step": 1490
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6383,
      "step": 1500
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6477,
      "step": 1510
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5758,
      "step": 1520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5195,
      "step": 1530
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5691,
      "step": 1540
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5451,
      "step": 1550
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5753,
      "step": 1560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5859,
      "step": 1570
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6325,
      "step": 1580
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7023,
      "step": 1590
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5009,
      "step": 1600
    },
    {
      "epoch": 0.52,
      "eval_loss": 0.5365247130393982,
      "eval_runtime": 92.8248,
      "eval_samples_per_second": 10.773,
      "eval_steps_per_second": 5.386,
      "step": 1600
    },
    {
      "epoch": 0.52,
      "mmlu_eval_accuracy": 0.4625997452457202,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.0,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.08333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.2727272727272727,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.5238095238095238,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.4186046511627907,
      "mmlu_eval_accuracy_high_school_mathematics": 0.13793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.5454545454545454,
      "mmlu_eval_accuracy_high_school_world_history": 0.4230769230769231,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.25,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.6363636363636364,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6162790697674418,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.24,
      "mmlu_eval_accuracy_nutrition": 0.5454545454545454,
      "mmlu_eval_accuracy_philosophy": 0.4411764705882353,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
      "mmlu_eval_accuracy_professional_law": 0.34705882352941175,
      "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.5454545454545454,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8016339474243377,
      "step": 1600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5698,
      "step": 1610
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5115,
      "step": 1620
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5908,
      "step": 1630
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6114,
      "step": 1640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5335,
      "step": 1650
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6275,
      "step": 1660
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4862,
      "step": 1670
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6334,
      "step": 1680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6494,
      "step": 1690
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5298,
      "step": 1700
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5359,
      "step": 1710
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5049,
      "step": 1720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5015,
      "step": 1730
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6523,
      "step": 1740
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5047,
      "step": 1750
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5321,
      "step": 1760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5595,
      "step": 1770
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7094,
      "step": 1780
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6828,
      "step": 1790
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5355,
      "step": 1800
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.533126711845398,
      "eval_runtime": 92.8481,
      "eval_samples_per_second": 10.77,
      "eval_steps_per_second": 5.385,
      "step": 1800
    },
    {
      "epoch": 0.58,
      "mmlu_eval_accuracy": 0.4894584764122198,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.3793103448275862,
      "mmlu_eval_accuracy_college_biology": 0.375,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "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": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "mmlu_eval_accuracy_high_school_biology": 0.40625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "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": 0.5,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.7333333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
      "mmlu_eval_accuracy_high_school_world_history": 0.46153846153846156,
      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.5,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.4411764705882353,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.3411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.37681159420289856,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.9346268427978919,
      "step": 1800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4742,
      "step": 1810
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5088,
      "step": 1820
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.459,
      "step": 1830
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5195,
      "step": 1840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.617,
      "step": 1850
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4523,
      "step": 1860
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.646,
      "step": 1870
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4423,
      "step": 1880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5904,
      "step": 1890
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5508,
      "step": 1900
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4987,
      "step": 1910
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5466,
      "step": 1920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5342,
      "step": 1930
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6875,
      "step": 1940
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5443,
      "step": 1950
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5367,
      "step": 1960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5529,
      "step": 1970
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.438,
      "step": 1980
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4422,
      "step": 1990
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5624,
      "step": 2000
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.5018913745880127,
      "eval_runtime": 92.7994,
      "eval_samples_per_second": 10.776,
      "eval_steps_per_second": 5.388,
      "step": 2000
    },
    {
      "epoch": 0.65,
      "mmlu_eval_accuracy": 0.4873839805836957,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
      "mmlu_eval_accuracy_college_biology": 0.375,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.08333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.4375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.72,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.27,
      "mmlu_eval_accuracy_nutrition": 0.5454545454545454,
      "mmlu_eval_accuracy_philosophy": 0.4411764705882353,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.3588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_psychology": 0.4492753623188406,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
      "mmlu_eval_accuracy_sociology": 0.5909090909090909,
      "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8898517729972268,
      "step": 2000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4835,
      "step": 2010
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5138,
      "step": 2020
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4579,
      "step": 2030
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6192,
      "step": 2040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5606,
      "step": 2050
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5427,
      "step": 2060
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5302,
      "step": 2070
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5107,
      "step": 2080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5849,
      "step": 2090
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4736,
      "step": 2100
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5274,
      "step": 2110
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5004,
      "step": 2120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4559,
      "step": 2130
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4954,
      "step": 2140
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5255,
      "step": 2150
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3956,
      "step": 2160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.446,
      "step": 2170
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4923,
      "step": 2180
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5505,
      "step": 2190
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6295,
      "step": 2200
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.49923792481422424,
      "eval_runtime": 92.7241,
      "eval_samples_per_second": 10.785,
      "eval_steps_per_second": 5.392,
      "step": 2200
    },
    {
      "epoch": 0.71,
      "mmlu_eval_accuracy": 0.46882014178817705,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
      "mmlu_eval_accuracy_college_biology": 0.4375,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
      "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": 0.08333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.25,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.40625,
      "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": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.37209302325581395,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.75,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "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": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
      "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.68,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.627906976744186,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.23,
      "mmlu_eval_accuracy_nutrition": 0.5454545454545454,
      "mmlu_eval_accuracy_philosophy": 0.38235294117647056,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.3225806451612903,
      "mmlu_eval_accuracy_professional_law": 0.36470588235294116,
      "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_psychology": 0.42028985507246375,
      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
      "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": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8042322283090328,
      "step": 2200
    }
  ],
  "max_steps": 5000,
  "num_train_epochs": 2,
  "total_flos": 1.803354924659835e+17,
  "trial_name": null,
  "trial_params": null
}