{
  "best_metric": 0.8125561475753784,
  "best_model_checkpoint": "experts/mistralic-expert-22/checkpoint-3400",
  "epoch": 0.885424185554112,
  "eval_steps": 200,
  "global_step": 4600,
  "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": 0.9917,
      "step": 1
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8752,
      "step": 2
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8807,
      "step": 3
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.7598,
      "step": 4
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6723,
      "step": 5
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 1.1704,
      "step": 6
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.7503,
      "step": 7
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.624,
      "step": 8
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8103,
      "step": 9
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.9605,
      "step": 10
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.905,
      "step": 11
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.759,
      "step": 12
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8095,
      "step": 13
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 1.1747,
      "step": 14
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8053,
      "step": 15
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6902,
      "step": 16
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.878,
      "step": 17
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.89,
      "step": 18
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.899,
      "step": 19
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 1.0926,
      "step": 20
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.9986,
      "step": 21
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8805,
      "step": 22
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6184,
      "step": 23
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.9538,
      "step": 24
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 1.0293,
      "step": 25
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4661,
      "step": 26
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8878,
      "step": 27
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8821,
      "step": 28
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.85,
      "step": 29
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.9784,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7107,
      "step": 31
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7971,
      "step": 32
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8528,
      "step": 33
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8603,
      "step": 34
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7917,
      "step": 35
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6604,
      "step": 36
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8287,
      "step": 37
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 1.1047,
      "step": 38
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6991,
      "step": 39
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.9996,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8583,
      "step": 41
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7114,
      "step": 42
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7781,
      "step": 43
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.9923,
      "step": 44
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8158,
      "step": 45
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 46
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8207,
      "step": 47
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7159,
      "step": 48
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8143,
      "step": 49
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.9955,
      "step": 50
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6577,
      "step": 51
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7208,
      "step": 52
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.9416,
      "step": 53
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7941,
      "step": 54
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7234,
      "step": 55
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 1.0252,
      "step": 56
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7091,
      "step": 57
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7707,
      "step": 58
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 1.157,
      "step": 59
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5682,
      "step": 60
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.807,
      "step": 61
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6721,
      "step": 62
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8324,
      "step": 63
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8405,
      "step": 64
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7994,
      "step": 65
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7385,
      "step": 66
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4077,
      "step": 67
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8142,
      "step": 68
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6694,
      "step": 69
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8789,
      "step": 70
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 1.058,
      "step": 71
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 1.1531,
      "step": 72
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.9339,
      "step": 73
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7771,
      "step": 74
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 1.0121,
      "step": 75
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7184,
      "step": 76
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6577,
      "step": 77
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9318,
      "step": 78
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 1.105,
      "step": 79
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6653,
      "step": 80
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.631,
      "step": 81
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9929,
      "step": 82
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 1.03,
      "step": 83
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7572,
      "step": 84
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8269,
      "step": 85
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6807,
      "step": 86
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9773,
      "step": 87
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.866,
      "step": 88
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9673,
      "step": 89
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8735,
      "step": 90
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8992,
      "step": 91
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8388,
      "step": 92
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8848,
      "step": 93
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6855,
      "step": 94
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7848,
      "step": 95
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 1.1573,
      "step": 96
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8753,
      "step": 97
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 1.016,
      "step": 98
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8481,
      "step": 99
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8461,
      "step": 100
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8009,
      "step": 101
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8147,
      "step": 102
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.818,
      "step": 103
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8583,
      "step": 105
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8898,
      "step": 106
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5784,
      "step": 107
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6973,
      "step": 108
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8995,
      "step": 109
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8895,
      "step": 110
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7302,
      "step": 111
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.975,
      "step": 112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7004,
      "step": 113
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6645,
      "step": 114
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7651,
      "step": 115
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8179,
      "step": 116
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8355,
      "step": 117
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8815,
      "step": 118
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8535,
      "step": 119
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7703,
      "step": 120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8666,
      "step": 121
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9532,
      "step": 122
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7817,
      "step": 123
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 1.0935,
      "step": 124
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7434,
      "step": 125
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8263,
      "step": 126
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9015,
      "step": 127
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 1.5507,
      "step": 128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.923,
      "step": 129
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 1.0015,
      "step": 130
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8959,
      "step": 131
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.996,
      "step": 132
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9214,
      "step": 133
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6441,
      "step": 134
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8035,
      "step": 135
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8738,
      "step": 136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6934,
      "step": 137
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7723,
      "step": 138
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8101,
      "step": 139
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6368,
      "step": 140
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7078,
      "step": 141
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8225,
      "step": 142
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7866,
      "step": 143
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8628,
      "step": 144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6387,
      "step": 145
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7479,
      "step": 146
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8173,
      "step": 147
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8456,
      "step": 148
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7523,
      "step": 149
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6802,
      "step": 150
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7845,
      "step": 151
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.622,
      "step": 152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.842,
      "step": 153
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 1.0277,
      "step": 154
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 1.0264,
      "step": 155
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9195,
      "step": 156
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.797,
      "step": 157
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.87,
      "step": 158
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 1.6444,
      "step": 159
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8486,
      "step": 160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9643,
      "step": 161
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6787,
      "step": 162
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7781,
      "step": 163
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6973,
      "step": 164
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7924,
      "step": 165
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9079,
      "step": 166
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7007,
      "step": 167
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8613,
      "step": 168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.608,
      "step": 169
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9465,
      "step": 170
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9435,
      "step": 171
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7896,
      "step": 172
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6892,
      "step": 173
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8605,
      "step": 174
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7067,
      "step": 175
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6283,
      "step": 176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.9131,
      "step": 177
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 1.0426,
      "step": 178
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7652,
      "step": 179
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 1.0208,
      "step": 180
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7484,
      "step": 181
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8419,
      "step": 182
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9176,
      "step": 183
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.861,
      "step": 184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9,
      "step": 185
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.907,
      "step": 186
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.651,
      "step": 187
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6038,
      "step": 188
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7511,
      "step": 189
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6037,
      "step": 190
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.592,
      "step": 191
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5935,
      "step": 192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9313,
      "step": 193
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8181,
      "step": 194
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.792,
      "step": 195
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7445,
      "step": 196
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6853,
      "step": 197
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9783,
      "step": 198
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9599,
      "step": 199
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9033,
      "step": 200
    },
    {
      "epoch": 0.04,
      "eval_loss": 0.8272572159767151,
      "eval_runtime": 108.0162,
      "eval_samples_per_second": 9.258,
      "eval_steps_per_second": 4.629,
      "step": 200
    },
    {
      "epoch": 0.04,
      "mmlu_eval_accuracy": 0.599579385714013,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.6363636363636364,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.5,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6511627906976745,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_law": 0.38823529411764707,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.6231884057971014,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.1056992180898357,
      "step": 200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8147,
      "step": 201
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 1.2314,
      "step": 202
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9407,
      "step": 203
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7344,
      "step": 204
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7506,
      "step": 205
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9563,
      "step": 206
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9926,
      "step": 207
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8277,
      "step": 208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7435,
      "step": 209
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9858,
      "step": 210
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8615,
      "step": 211
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7798,
      "step": 212
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8055,
      "step": 213
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8161,
      "step": 214
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8922,
      "step": 215
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7699,
      "step": 216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9098,
      "step": 217
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8848,
      "step": 218
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7967,
      "step": 219
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7214,
      "step": 220
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6714,
      "step": 221
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8672,
      "step": 222
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7304,
      "step": 223
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7675,
      "step": 224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5663,
      "step": 225
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9466,
      "step": 226
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 1.0255,
      "step": 227
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.9056,
      "step": 228
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8921,
      "step": 229
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7836,
      "step": 230
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8398,
      "step": 231
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8849,
      "step": 232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8133,
      "step": 233
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9293,
      "step": 234
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6496,
      "step": 235
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5932,
      "step": 236
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9838,
      "step": 237
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8546,
      "step": 238
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9101,
      "step": 239
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.859,
      "step": 240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9602,
      "step": 241
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9327,
      "step": 242
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5277,
      "step": 243
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.748,
      "step": 244
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7524,
      "step": 245
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9312,
      "step": 246
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 1.2737,
      "step": 247
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 1.0512,
      "step": 248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7334,
      "step": 249
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7493,
      "step": 250
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5577,
      "step": 251
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 1.0016,
      "step": 252
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9124,
      "step": 253
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8657,
      "step": 254
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9144,
      "step": 255
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9532,
      "step": 256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7411,
      "step": 257
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7931,
      "step": 258
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6414,
      "step": 259
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7083,
      "step": 260
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9409,
      "step": 261
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.856,
      "step": 262
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9955,
      "step": 263
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5929,
      "step": 264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7547,
      "step": 265
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6542,
      "step": 266
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8503,
      "step": 267
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8429,
      "step": 268
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.857,
      "step": 269
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8845,
      "step": 270
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8411,
      "step": 271
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9812,
      "step": 272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8003,
      "step": 273
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7384,
      "step": 274
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7263,
      "step": 275
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.78,
      "step": 276
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.9858,
      "step": 277
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7142,
      "step": 278
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7677,
      "step": 279
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6429,
      "step": 280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8833,
      "step": 281
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8403,
      "step": 282
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5925,
      "step": 283
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7705,
      "step": 284
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7843,
      "step": 285
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8162,
      "step": 286
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7978,
      "step": 287
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7364,
      "step": 288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8184,
      "step": 289
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8204,
      "step": 290
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9493,
      "step": 291
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8844,
      "step": 292
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7213,
      "step": 293
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7367,
      "step": 294
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8065,
      "step": 295
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7912,
      "step": 296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9889,
      "step": 297
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6346,
      "step": 298
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8347,
      "step": 299
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.821,
      "step": 300
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 1.0193,
      "step": 301
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7745,
      "step": 302
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5443,
      "step": 303
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 1.0296,
      "step": 304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8065,
      "step": 305
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7023,
      "step": 306
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7963,
      "step": 307
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6676,
      "step": 308
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7119,
      "step": 309
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8342,
      "step": 310
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8657,
      "step": 311
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9189,
      "step": 312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9195,
      "step": 313
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.737,
      "step": 314
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6512,
      "step": 315
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9037,
      "step": 316
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.942,
      "step": 317
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9569,
      "step": 318
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7858,
      "step": 319
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 1.0143,
      "step": 320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 1.0598,
      "step": 321
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8547,
      "step": 322
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6792,
      "step": 323
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9757,
      "step": 324
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7605,
      "step": 325
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7029,
      "step": 326
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9174,
      "step": 327
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.908,
      "step": 328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 1.0953,
      "step": 329
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6913,
      "step": 330
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7406,
      "step": 331
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6988,
      "step": 332
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4839,
      "step": 333
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.9061,
      "step": 334
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7339,
      "step": 335
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7077,
      "step": 336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 1.0788,
      "step": 337
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7634,
      "step": 338
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.652,
      "step": 339
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.802,
      "step": 340
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7895,
      "step": 341
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8487,
      "step": 342
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7397,
      "step": 343
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.857,
      "step": 344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9782,
      "step": 345
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.688,
      "step": 346
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5103,
      "step": 347
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.762,
      "step": 348
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9286,
      "step": 349
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7606,
      "step": 350
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8981,
      "step": 351
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7572,
      "step": 352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7552,
      "step": 353
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6557,
      "step": 354
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7197,
      "step": 355
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6654,
      "step": 356
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8249,
      "step": 357
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9703,
      "step": 358
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9126,
      "step": 359
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6395,
      "step": 360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8437,
      "step": 361
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8885,
      "step": 362
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9231,
      "step": 363
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5622,
      "step": 364
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9886,
      "step": 365
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6807,
      "step": 366
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7713,
      "step": 367
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8205,
      "step": 368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6163,
      "step": 369
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6523,
      "step": 370
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7218,
      "step": 371
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7817,
      "step": 372
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 1.1076,
      "step": 373
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9519,
      "step": 374
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9615,
      "step": 375
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8734,
      "step": 376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7385,
      "step": 377
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6975,
      "step": 378
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6935,
      "step": 379
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6487,
      "step": 380
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8057,
      "step": 381
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7861,
      "step": 382
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6614,
      "step": 383
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6929,
      "step": 384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9957,
      "step": 385
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7739,
      "step": 386
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.953,
      "step": 387
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5405,
      "step": 388
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8276,
      "step": 389
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6577,
      "step": 390
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6967,
      "step": 391
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7781,
      "step": 392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8823,
      "step": 393
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5715,
      "step": 394
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9113,
      "step": 395
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8301,
      "step": 396
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9687,
      "step": 397
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.76,
      "step": 398
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7402,
      "step": 399
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7616,
      "step": 400
    },
    {
      "epoch": 0.08,
      "eval_loss": 0.8319438099861145,
      "eval_runtime": 107.9019,
      "eval_samples_per_second": 9.268,
      "eval_steps_per_second": 4.634,
      "step": 400
    },
    {
      "epoch": 0.08,
      "mmlu_eval_accuracy": 0.5644849203853226,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.18181818181818182,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.631578947368421,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.7878787878787878,
      "mmlu_eval_accuracy_philosophy": 0.6176470588235294,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.3124866759886007,
      "step": 400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8433,
      "step": 401
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9205,
      "step": 402
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7859,
      "step": 403
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8488,
      "step": 404
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6307,
      "step": 405
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5087,
      "step": 406
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8378,
      "step": 407
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8898,
      "step": 408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8309,
      "step": 409
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8289,
      "step": 410
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8561,
      "step": 411
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8791,
      "step": 412
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6545,
      "step": 413
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9316,
      "step": 414
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7037,
      "step": 415
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6658,
      "step": 416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9517,
      "step": 417
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8115,
      "step": 418
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8659,
      "step": 419
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6427,
      "step": 420
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6879,
      "step": 421
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 1.0384,
      "step": 422
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5888,
      "step": 423
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9127,
      "step": 424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7456,
      "step": 425
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8999,
      "step": 426
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9658,
      "step": 427
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9099,
      "step": 428
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8155,
      "step": 429
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5489,
      "step": 430
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7532,
      "step": 431
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7318,
      "step": 432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8388,
      "step": 433
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7363,
      "step": 434
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8658,
      "step": 435
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.87,
      "step": 436
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8169,
      "step": 437
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.9461,
      "step": 438
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.777,
      "step": 439
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8922,
      "step": 440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6376,
      "step": 441
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 442
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.845,
      "step": 443
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.1059,
      "step": 444
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9725,
      "step": 445
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.001,
      "step": 446
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9184,
      "step": 447
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7949,
      "step": 448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6397,
      "step": 449
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8519,
      "step": 450
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7851,
      "step": 451
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7957,
      "step": 452
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.2149,
      "step": 453
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9572,
      "step": 454
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9525,
      "step": 455
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.0605,
      "step": 456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7367,
      "step": 457
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8119,
      "step": 458
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6568,
      "step": 459
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7564,
      "step": 460
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7836,
      "step": 461
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8655,
      "step": 462
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9538,
      "step": 463
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7256,
      "step": 464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6022,
      "step": 465
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5525,
      "step": 466
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8683,
      "step": 467
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5313,
      "step": 468
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7129,
      "step": 469
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9313,
      "step": 470
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9505,
      "step": 471
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8025,
      "step": 472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8977,
      "step": 473
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8055,
      "step": 474
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9525,
      "step": 475
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.0739,
      "step": 476
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.0571,
      "step": 477
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8615,
      "step": 478
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8694,
      "step": 479
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8516,
      "step": 480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 1.0282,
      "step": 481
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9444,
      "step": 482
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6297,
      "step": 483
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9982,
      "step": 484
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.814,
      "step": 485
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8239,
      "step": 486
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8397,
      "step": 487
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7726,
      "step": 488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.8604,
      "step": 489
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9932,
      "step": 490
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9045,
      "step": 491
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9513,
      "step": 492
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.9425,
      "step": 493
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6765,
      "step": 494
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8696,
      "step": 495
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8834,
      "step": 496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8216,
      "step": 497
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8824,
      "step": 498
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7342,
      "step": 499
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8486,
      "step": 500
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 1.1737,
      "step": 501
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7487,
      "step": 502
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5858,
      "step": 503
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8642,
      "step": 504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8384,
      "step": 505
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.9313,
      "step": 506
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8897,
      "step": 507
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7311,
      "step": 508
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7789,
      "step": 509
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8578,
      "step": 510
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 1.0066,
      "step": 511
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6273,
      "step": 512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6804,
      "step": 513
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7417,
      "step": 514
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8504,
      "step": 515
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.856,
      "step": 516
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8277,
      "step": 517
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7622,
      "step": 518
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6808,
      "step": 519
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.9331,
      "step": 520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7157,
      "step": 521
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6574,
      "step": 522
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.9497,
      "step": 523
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7128,
      "step": 524
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7914,
      "step": 525
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7403,
      "step": 526
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8889,
      "step": 527
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8651,
      "step": 528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6389,
      "step": 529
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.9403,
      "step": 530
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7894,
      "step": 531
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.882,
      "step": 532
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8738,
      "step": 533
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 1.1446,
      "step": 534
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7493,
      "step": 535
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8229,
      "step": 536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8648,
      "step": 537
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.9932,
      "step": 538
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.8452,
      "step": 539
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7869,
      "step": 540
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 1.1482,
      "step": 541
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7878,
      "step": 542
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.788,
      "step": 543
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6392,
      "step": 544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.9142,
      "step": 545
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5735,
      "step": 546
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7491,
      "step": 547
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8111,
      "step": 548
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 1.1825,
      "step": 549
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7141,
      "step": 550
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7931,
      "step": 551
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7072,
      "step": 552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6284,
      "step": 553
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7792,
      "step": 554
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8095,
      "step": 555
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7571,
      "step": 556
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7008,
      "step": 557
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 1.0028,
      "step": 558
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.84,
      "step": 559
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8219,
      "step": 560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6631,
      "step": 561
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4663,
      "step": 562
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8435,
      "step": 563
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.9562,
      "step": 564
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6672,
      "step": 565
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6709,
      "step": 566
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8419,
      "step": 567
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7651,
      "step": 568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.699,
      "step": 569
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.9775,
      "step": 570
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8303,
      "step": 571
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7964,
      "step": 572
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6591,
      "step": 573
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8658,
      "step": 574
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.9896,
      "step": 575
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4332,
      "step": 576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8514,
      "step": 577
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8869,
      "step": 578
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8641,
      "step": 579
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8047,
      "step": 580
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6639,
      "step": 581
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 1.0241,
      "step": 582
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8246,
      "step": 583
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8675,
      "step": 584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8013,
      "step": 585
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7683,
      "step": 586
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6556,
      "step": 587
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.77,
      "step": 588
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7938,
      "step": 589
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6935,
      "step": 590
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.9622,
      "step": 591
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.87,
      "step": 592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7516,
      "step": 593
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8894,
      "step": 594
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7505,
      "step": 595
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.9013,
      "step": 596
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7466,
      "step": 597
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 1.0704,
      "step": 598
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7459,
      "step": 599
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6249,
      "step": 600
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.8239523768424988,
      "eval_runtime": 108.0006,
      "eval_samples_per_second": 9.259,
      "eval_steps_per_second": 4.63,
      "step": 600
    },
    {
      "epoch": 0.12,
      "mmlu_eval_accuracy": 0.5767479059350066,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.34,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.3333333333333333,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.0626621378230674,
      "step": 600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8937,
      "step": 601
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6484,
      "step": 602
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 1.1631,
      "step": 603
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9261,
      "step": 604
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8708,
      "step": 605
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7974,
      "step": 606
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.593,
      "step": 607
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7858,
      "step": 608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.853,
      "step": 609
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8493,
      "step": 610
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.767,
      "step": 611
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7043,
      "step": 612
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6778,
      "step": 613
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8997,
      "step": 614
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6661,
      "step": 615
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6566,
      "step": 616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 1.1303,
      "step": 617
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6522,
      "step": 618
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7512,
      "step": 619
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8616,
      "step": 620
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6807,
      "step": 621
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7494,
      "step": 622
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9582,
      "step": 623
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9157,
      "step": 624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.712,
      "step": 625
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7072,
      "step": 626
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9214,
      "step": 627
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8604,
      "step": 628
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7721,
      "step": 629
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6244,
      "step": 630
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 1.0174,
      "step": 631
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.933,
      "step": 632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8022,
      "step": 633
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6461,
      "step": 634
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8137,
      "step": 635
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.644,
      "step": 636
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.97,
      "step": 637
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 1.0134,
      "step": 638
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7135,
      "step": 639
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8643,
      "step": 640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9489,
      "step": 641
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9385,
      "step": 642
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6434,
      "step": 643
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8027,
      "step": 644
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5102,
      "step": 645
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7558,
      "step": 646
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8466,
      "step": 647
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9113,
      "step": 648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.8914,
      "step": 649
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8724,
      "step": 650
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7429,
      "step": 651
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.797,
      "step": 652
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9116,
      "step": 653
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 1.0321,
      "step": 654
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6855,
      "step": 655
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8333,
      "step": 656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8465,
      "step": 657
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.841,
      "step": 658
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7407,
      "step": 659
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.671,
      "step": 660
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8655,
      "step": 661
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7531,
      "step": 662
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.797,
      "step": 663
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7406,
      "step": 664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8582,
      "step": 665
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7556,
      "step": 666
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 1.0,
      "step": 667
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9225,
      "step": 668
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7745,
      "step": 669
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.672,
      "step": 670
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8807,
      "step": 671
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9416,
      "step": 672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9291,
      "step": 673
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 1.07,
      "step": 674
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.865,
      "step": 675
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5955,
      "step": 676
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9277,
      "step": 677
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7636,
      "step": 678
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6985,
      "step": 679
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7898,
      "step": 680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8206,
      "step": 681
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8341,
      "step": 682
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6745,
      "step": 683
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5721,
      "step": 684
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8839,
      "step": 685
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8846,
      "step": 686
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8544,
      "step": 687
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9011,
      "step": 688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 1.039,
      "step": 689
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9518,
      "step": 690
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7659,
      "step": 691
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7467,
      "step": 692
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7914,
      "step": 693
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8516,
      "step": 694
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9462,
      "step": 695
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7714,
      "step": 696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8217,
      "step": 697
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 1.0775,
      "step": 698
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6857,
      "step": 699
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8516,
      "step": 700
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.9827,
      "step": 701
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 702
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8473,
      "step": 703
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7903,
      "step": 704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 1.1412,
      "step": 705
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9854,
      "step": 706
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9383,
      "step": 707
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9324,
      "step": 708
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.772,
      "step": 709
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6148,
      "step": 710
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8209,
      "step": 711
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5247,
      "step": 712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7913,
      "step": 713
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8309,
      "step": 714
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.868,
      "step": 715
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8826,
      "step": 716
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8212,
      "step": 717
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.991,
      "step": 718
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7799,
      "step": 719
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5986,
      "step": 720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7601,
      "step": 721
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8981,
      "step": 722
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.672,
      "step": 723
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7597,
      "step": 724
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7075,
      "step": 725
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7639,
      "step": 726
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5174,
      "step": 727
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9761,
      "step": 728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8007,
      "step": 729
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9283,
      "step": 730
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8481,
      "step": 731
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.876,
      "step": 732
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9838,
      "step": 733
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.873,
      "step": 734
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6769,
      "step": 735
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8994,
      "step": 736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6173,
      "step": 737
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9697,
      "step": 738
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9337,
      "step": 739
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6562,
      "step": 740
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9336,
      "step": 741
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9479,
      "step": 742
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7138,
      "step": 743
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9331,
      "step": 744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8263,
      "step": 745
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7274,
      "step": 746
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8048,
      "step": 747
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8069,
      "step": 748
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7556,
      "step": 749
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9893,
      "step": 750
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7542,
      "step": 751
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8254,
      "step": 752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6808,
      "step": 753
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8977,
      "step": 754
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 1.0146,
      "step": 755
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8072,
      "step": 756
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7988,
      "step": 757
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9259,
      "step": 758
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7222,
      "step": 759
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6838,
      "step": 760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7277,
      "step": 761
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8344,
      "step": 762
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7687,
      "step": 763
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7731,
      "step": 764
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6727,
      "step": 765
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8028,
      "step": 766
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7871,
      "step": 767
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7153,
      "step": 768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7676,
      "step": 769
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9727,
      "step": 770
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6709,
      "step": 771
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8173,
      "step": 772
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7286,
      "step": 773
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.789,
      "step": 774
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5131,
      "step": 775
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8337,
      "step": 776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8912,
      "step": 777
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7301,
      "step": 778
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 1.0566,
      "step": 779
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8174,
      "step": 780
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7458,
      "step": 781
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7841,
      "step": 782
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8716,
      "step": 783
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7436,
      "step": 784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6822,
      "step": 785
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8617,
      "step": 786
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8575,
      "step": 787
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7825,
      "step": 788
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9225,
      "step": 789
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9179,
      "step": 790
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6989,
      "step": 791
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9076,
      "step": 792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.8997,
      "step": 793
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9768,
      "step": 794
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7592,
      "step": 795
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9849,
      "step": 796
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7044,
      "step": 797
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7252,
      "step": 798
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.785,
      "step": 799
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7535,
      "step": 800
    },
    {
      "epoch": 0.15,
      "eval_loss": 0.8314322233200073,
      "eval_runtime": 107.7767,
      "eval_samples_per_second": 9.278,
      "eval_steps_per_second": 4.639,
      "step": 800
    },
    {
      "epoch": 0.15,
      "mmlu_eval_accuracy": 0.5852864021594637,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.26,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.068377559941369,
      "step": 800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9542,
      "step": 801
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.9369,
      "step": 802
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.722,
      "step": 803
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5851,
      "step": 804
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 1.0236,
      "step": 805
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5474,
      "step": 806
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9945,
      "step": 807
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 1.0455,
      "step": 808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7946,
      "step": 809
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7688,
      "step": 810
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9149,
      "step": 811
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7135,
      "step": 812
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7656,
      "step": 813
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7422,
      "step": 814
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8751,
      "step": 815
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8876,
      "step": 816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 1.0541,
      "step": 817
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.993,
      "step": 818
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.77,
      "step": 819
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8067,
      "step": 820
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9629,
      "step": 821
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9683,
      "step": 822
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.722,
      "step": 823
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7577,
      "step": 824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6923,
      "step": 825
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8585,
      "step": 826
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8164,
      "step": 827
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7618,
      "step": 828
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8823,
      "step": 829
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9447,
      "step": 830
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9193,
      "step": 831
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7581,
      "step": 832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7875,
      "step": 833
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7074,
      "step": 834
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7871,
      "step": 835
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9631,
      "step": 836
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9075,
      "step": 837
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9951,
      "step": 838
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7714,
      "step": 839
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7699,
      "step": 840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7919,
      "step": 841
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9562,
      "step": 842
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9486,
      "step": 843
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7622,
      "step": 844
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7765,
      "step": 845
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6023,
      "step": 846
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7426,
      "step": 847
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6728,
      "step": 848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.825,
      "step": 849
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9433,
      "step": 850
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9373,
      "step": 851
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9356,
      "step": 852
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9599,
      "step": 853
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.825,
      "step": 854
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9899,
      "step": 855
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.9104,
      "step": 856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7832,
      "step": 857
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7825,
      "step": 858
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.661,
      "step": 859
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6978,
      "step": 860
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7528,
      "step": 861
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.553,
      "step": 862
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7198,
      "step": 863
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7452,
      "step": 864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4933,
      "step": 865
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5857,
      "step": 866
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8252,
      "step": 867
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8617,
      "step": 868
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.64,
      "step": 869
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7116,
      "step": 870
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9711,
      "step": 871
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9874,
      "step": 872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8595,
      "step": 873
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7947,
      "step": 874
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6689,
      "step": 875
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9601,
      "step": 876
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.834,
      "step": 877
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5666,
      "step": 878
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7491,
      "step": 879
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7362,
      "step": 880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6809,
      "step": 881
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8278,
      "step": 882
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8141,
      "step": 883
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 884
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7019,
      "step": 885
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9741,
      "step": 886
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9711,
      "step": 887
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7295,
      "step": 888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8351,
      "step": 889
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8951,
      "step": 890
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6802,
      "step": 891
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8653,
      "step": 892
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9284,
      "step": 893
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8818,
      "step": 894
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7905,
      "step": 895
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8017,
      "step": 896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7307,
      "step": 897
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9555,
      "step": 898
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8246,
      "step": 899
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8623,
      "step": 900
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7665,
      "step": 901
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7087,
      "step": 902
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7632,
      "step": 903
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8463,
      "step": 904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.67,
      "step": 905
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9897,
      "step": 906
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.839,
      "step": 907
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6107,
      "step": 908
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 1.0102,
      "step": 909
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5473,
      "step": 910
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9059,
      "step": 911
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9688,
      "step": 912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6527,
      "step": 913
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7323,
      "step": 914
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7994,
      "step": 915
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 1.0027,
      "step": 916
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9274,
      "step": 917
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5995,
      "step": 918
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8599,
      "step": 919
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7599,
      "step": 920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.601,
      "step": 921
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8202,
      "step": 922
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8857,
      "step": 923
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5073,
      "step": 924
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 1.0193,
      "step": 925
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8295,
      "step": 926
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9962,
      "step": 927
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 1.1134,
      "step": 928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.833,
      "step": 929
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9124,
      "step": 930
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 1.0302,
      "step": 931
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9182,
      "step": 932
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.532,
      "step": 933
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7662,
      "step": 934
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8333,
      "step": 935
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6508,
      "step": 936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.603,
      "step": 937
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8424,
      "step": 938
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8288,
      "step": 939
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8063,
      "step": 940
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7095,
      "step": 941
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9294,
      "step": 942
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9205,
      "step": 943
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8253,
      "step": 944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8448,
      "step": 945
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7412,
      "step": 946
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9216,
      "step": 947
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.822,
      "step": 948
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8435,
      "step": 949
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8546,
      "step": 950
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.575,
      "step": 951
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 1.0692,
      "step": 952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8348,
      "step": 953
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9301,
      "step": 954
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7265,
      "step": 955
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8753,
      "step": 956
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8021,
      "step": 957
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7412,
      "step": 958
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7696,
      "step": 959
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.9858,
      "step": 960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8307,
      "step": 961
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7608,
      "step": 962
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8428,
      "step": 963
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9388,
      "step": 964
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5026,
      "step": 965
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6896,
      "step": 966
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7954,
      "step": 967
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9893,
      "step": 968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6204,
      "step": 969
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8292,
      "step": 970
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6268,
      "step": 971
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9288,
      "step": 972
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8863,
      "step": 973
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8266,
      "step": 974
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8311,
      "step": 975
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7688,
      "step": 976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9308,
      "step": 977
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.0505,
      "step": 978
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8594,
      "step": 979
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.927,
      "step": 980
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6642,
      "step": 981
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8662,
      "step": 982
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8863,
      "step": 983
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7942,
      "step": 984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9385,
      "step": 985
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.1678,
      "step": 986
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.0085,
      "step": 987
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.0007,
      "step": 988
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.1553,
      "step": 989
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8498,
      "step": 990
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.1541,
      "step": 991
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8701,
      "step": 992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.691,
      "step": 993
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7448,
      "step": 994
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9344,
      "step": 995
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 1.1617,
      "step": 996
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6735,
      "step": 997
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9245,
      "step": 998
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6245,
      "step": 999
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7727,
      "step": 1000
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.8221133947372437,
      "eval_runtime": 107.7614,
      "eval_samples_per_second": 9.28,
      "eval_steps_per_second": 4.64,
      "step": 1000
    },
    {
      "epoch": 0.19,
      "mmlu_eval_accuracy": 0.5768050908161603,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.18181818181818182,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.7272727272727273,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.35,
      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
      "mmlu_eval_accuracy_philosophy": 0.5882352941176471,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.6666666666666666,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.0894444307443678,
      "step": 1000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7983,
      "step": 1001
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7486,
      "step": 1002
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8172,
      "step": 1003
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6772,
      "step": 1004
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7724,
      "step": 1005
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8207,
      "step": 1006
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7904,
      "step": 1007
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8351,
      "step": 1008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7827,
      "step": 1009
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9647,
      "step": 1010
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.9669,
      "step": 1011
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7974,
      "step": 1012
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7287,
      "step": 1013
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9687,
      "step": 1014
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9239,
      "step": 1015
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9859,
      "step": 1016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6932,
      "step": 1017
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.762,
      "step": 1018
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8419,
      "step": 1019
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8942,
      "step": 1020
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6363,
      "step": 1021
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 1.0143,
      "step": 1022
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7529,
      "step": 1023
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9264,
      "step": 1024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6283,
      "step": 1025
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7133,
      "step": 1026
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7825,
      "step": 1027
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7986,
      "step": 1028
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6617,
      "step": 1029
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8952,
      "step": 1030
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9269,
      "step": 1031
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8428,
      "step": 1032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9976,
      "step": 1033
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9051,
      "step": 1034
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.704,
      "step": 1035
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6956,
      "step": 1036
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7071,
      "step": 1037
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8563,
      "step": 1038
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5753,
      "step": 1039
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7108,
      "step": 1040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7634,
      "step": 1041
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9367,
      "step": 1042
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9209,
      "step": 1043
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7426,
      "step": 1044
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7212,
      "step": 1045
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 1.172,
      "step": 1046
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6434,
      "step": 1047
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8692,
      "step": 1048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7776,
      "step": 1049
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7858,
      "step": 1050
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7578,
      "step": 1051
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7549,
      "step": 1052
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8046,
      "step": 1053
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8274,
      "step": 1054
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 1.0264,
      "step": 1055
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8592,
      "step": 1056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5615,
      "step": 1057
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7084,
      "step": 1058
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8377,
      "step": 1059
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9046,
      "step": 1060
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9777,
      "step": 1061
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.598,
      "step": 1062
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7353,
      "step": 1063
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7785,
      "step": 1064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7571,
      "step": 1065
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8154,
      "step": 1066
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6559,
      "step": 1067
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8339,
      "step": 1068
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9225,
      "step": 1069
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9206,
      "step": 1070
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8587,
      "step": 1071
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6064,
      "step": 1072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.839,
      "step": 1073
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7312,
      "step": 1074
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8338,
      "step": 1075
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6873,
      "step": 1076
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9526,
      "step": 1077
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7698,
      "step": 1078
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8345,
      "step": 1079
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.836,
      "step": 1080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6012,
      "step": 1081
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8149,
      "step": 1082
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.621,
      "step": 1083
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9225,
      "step": 1084
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8493,
      "step": 1085
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6558,
      "step": 1086
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7498,
      "step": 1087
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.985,
      "step": 1088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.891,
      "step": 1089
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8744,
      "step": 1090
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8277,
      "step": 1091
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6847,
      "step": 1092
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6816,
      "step": 1093
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6313,
      "step": 1094
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7649,
      "step": 1095
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9559,
      "step": 1096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 1.0069,
      "step": 1097
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8068,
      "step": 1098
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.826,
      "step": 1099
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9864,
      "step": 1100
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8207,
      "step": 1101
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6934,
      "step": 1102
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9991,
      "step": 1103
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7827,
      "step": 1104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 1.0234,
      "step": 1105
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9579,
      "step": 1106
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6391,
      "step": 1107
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8214,
      "step": 1108
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 1.0001,
      "step": 1109
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9233,
      "step": 1110
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5531,
      "step": 1111
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7158,
      "step": 1112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8638,
      "step": 1113
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8938,
      "step": 1114
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8129,
      "step": 1115
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.89,
      "step": 1116
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5949,
      "step": 1117
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6555,
      "step": 1118
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8881,
      "step": 1119
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9674,
      "step": 1120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8433,
      "step": 1121
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8145,
      "step": 1122
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7719,
      "step": 1123
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 1.0407,
      "step": 1124
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7751,
      "step": 1125
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9276,
      "step": 1126
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.885,
      "step": 1127
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9224,
      "step": 1128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.964,
      "step": 1129
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.822,
      "step": 1130
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9673,
      "step": 1131
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8695,
      "step": 1132
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9036,
      "step": 1133
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 1.0815,
      "step": 1134
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8784,
      "step": 1135
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8444,
      "step": 1136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9152,
      "step": 1137
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7576,
      "step": 1138
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8372,
      "step": 1139
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9207,
      "step": 1140
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9243,
      "step": 1141
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6946,
      "step": 1142
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5998,
      "step": 1143
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6083,
      "step": 1144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 1.0207,
      "step": 1145
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8205,
      "step": 1146
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7422,
      "step": 1147
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.763,
      "step": 1148
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8019,
      "step": 1149
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9221,
      "step": 1150
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8362,
      "step": 1151
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8024,
      "step": 1152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8264,
      "step": 1153
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6174,
      "step": 1154
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6716,
      "step": 1155
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.798,
      "step": 1156
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8662,
      "step": 1157
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7987,
      "step": 1158
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9795,
      "step": 1159
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8001,
      "step": 1160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7329,
      "step": 1161
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8504,
      "step": 1162
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9516,
      "step": 1163
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.9072,
      "step": 1164
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7216,
      "step": 1165
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8843,
      "step": 1166
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7729,
      "step": 1167
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8456,
      "step": 1168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7582,
      "step": 1169
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7492,
      "step": 1170
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9831,
      "step": 1171
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7332,
      "step": 1172
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7229,
      "step": 1173
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7622,
      "step": 1174
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5888,
      "step": 1175
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8685,
      "step": 1176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.778,
      "step": 1177
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9537,
      "step": 1178
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6841,
      "step": 1179
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8068,
      "step": 1180
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 1.0086,
      "step": 1181
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8699,
      "step": 1182
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7676,
      "step": 1183
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8789,
      "step": 1184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 1.375,
      "step": 1185
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.759,
      "step": 1186
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8446,
      "step": 1187
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8148,
      "step": 1188
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6919,
      "step": 1189
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.608,
      "step": 1190
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.749,
      "step": 1191
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7304,
      "step": 1192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 1.1062,
      "step": 1193
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7372,
      "step": 1194
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9267,
      "step": 1195
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9035,
      "step": 1196
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8798,
      "step": 1197
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9881,
      "step": 1198
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.872,
      "step": 1199
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.93,
      "step": 1200
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.8195345401763916,
      "eval_runtime": 107.5294,
      "eval_samples_per_second": 9.3,
      "eval_steps_per_second": 4.65,
      "step": 1200
    },
    {
      "epoch": 0.23,
      "mmlu_eval_accuracy": 0.5435130031626643,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.09090909090909091,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "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": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.6176470588235294,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_psychology": 0.5362318840579711,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.2997911337461547,
      "step": 1200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8528,
      "step": 1201
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9547,
      "step": 1202
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8032,
      "step": 1203
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.679,
      "step": 1204
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.824,
      "step": 1205
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.964,
      "step": 1206
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6154,
      "step": 1207
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5239,
      "step": 1208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7988,
      "step": 1209
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5223,
      "step": 1210
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7433,
      "step": 1211
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7831,
      "step": 1212
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7253,
      "step": 1213
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8159,
      "step": 1214
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6605,
      "step": 1215
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5967,
      "step": 1216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8218,
      "step": 1217
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.9243,
      "step": 1218
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8622,
      "step": 1219
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8015,
      "step": 1220
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6271,
      "step": 1221
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8378,
      "step": 1222
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8605,
      "step": 1223
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7579,
      "step": 1224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8241,
      "step": 1225
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7355,
      "step": 1226
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7556,
      "step": 1227
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 1.054,
      "step": 1228
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8148,
      "step": 1229
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.878,
      "step": 1230
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.939,
      "step": 1231
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.624,
      "step": 1232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 1.0049,
      "step": 1233
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8586,
      "step": 1234
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8201,
      "step": 1235
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8713,
      "step": 1236
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7515,
      "step": 1237
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5959,
      "step": 1238
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8113,
      "step": 1239
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5581,
      "step": 1240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7962,
      "step": 1241
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8917,
      "step": 1242
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 1.0076,
      "step": 1243
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8805,
      "step": 1244
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8305,
      "step": 1245
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8017,
      "step": 1246
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 1.1182,
      "step": 1247
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8506,
      "step": 1248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8716,
      "step": 1249
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7659,
      "step": 1250
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6689,
      "step": 1251
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 1.0241,
      "step": 1252
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.9119,
      "step": 1253
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8612,
      "step": 1254
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6326,
      "step": 1255
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7158,
      "step": 1256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.9439,
      "step": 1257
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7628,
      "step": 1258
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7597,
      "step": 1259
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7789,
      "step": 1260
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7511,
      "step": 1261
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6601,
      "step": 1262
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7342,
      "step": 1263
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7831,
      "step": 1264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8261,
      "step": 1265
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7506,
      "step": 1266
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5815,
      "step": 1267
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 1268
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7392,
      "step": 1269
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6834,
      "step": 1270
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8545,
      "step": 1271
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 1.0233,
      "step": 1272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7578,
      "step": 1273
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.2378,
      "step": 1274
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9399,
      "step": 1275
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.668,
      "step": 1276
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7894,
      "step": 1277
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6307,
      "step": 1278
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7293,
      "step": 1279
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.629,
      "step": 1280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8629,
      "step": 1281
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9917,
      "step": 1282
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7331,
      "step": 1283
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8318,
      "step": 1284
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.725,
      "step": 1285
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7665,
      "step": 1286
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8015,
      "step": 1287
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7864,
      "step": 1288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8412,
      "step": 1289
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9087,
      "step": 1290
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7792,
      "step": 1291
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.0072,
      "step": 1292
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7536,
      "step": 1293
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.0158,
      "step": 1294
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9052,
      "step": 1295
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8222,
      "step": 1296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8344,
      "step": 1297
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.825,
      "step": 1298
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6271,
      "step": 1299
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9525,
      "step": 1300
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6363,
      "step": 1301
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7225,
      "step": 1302
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9168,
      "step": 1303
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8018,
      "step": 1304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6198,
      "step": 1305
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8333,
      "step": 1306
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8377,
      "step": 1307
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7898,
      "step": 1308
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.9539,
      "step": 1309
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.0578,
      "step": 1310
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.0193,
      "step": 1311
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8825,
      "step": 1312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8261,
      "step": 1313
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.846,
      "step": 1314
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 1315
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.2466,
      "step": 1316
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.0793,
      "step": 1317
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7391,
      "step": 1318
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6585,
      "step": 1319
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4362,
      "step": 1320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 1321
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 1.1133,
      "step": 1322
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8416,
      "step": 1323
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5723,
      "step": 1324
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9606,
      "step": 1325
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6929,
      "step": 1326
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.846,
      "step": 1327
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4776,
      "step": 1328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.641,
      "step": 1329
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8026,
      "step": 1330
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8442,
      "step": 1331
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5105,
      "step": 1332
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.778,
      "step": 1333
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8846,
      "step": 1334
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8107,
      "step": 1335
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.844,
      "step": 1336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8696,
      "step": 1337
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7486,
      "step": 1338
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8199,
      "step": 1339
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9825,
      "step": 1340
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8303,
      "step": 1341
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7669,
      "step": 1342
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8261,
      "step": 1343
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8445,
      "step": 1344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7926,
      "step": 1345
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.735,
      "step": 1346
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8486,
      "step": 1347
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6096,
      "step": 1348
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6806,
      "step": 1349
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8527,
      "step": 1350
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.718,
      "step": 1351
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8803,
      "step": 1352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9253,
      "step": 1353
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6331,
      "step": 1354
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7672,
      "step": 1355
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8612,
      "step": 1356
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7002,
      "step": 1357
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7032,
      "step": 1358
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9663,
      "step": 1359
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8368,
      "step": 1360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6596,
      "step": 1361
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 1.0682,
      "step": 1362
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6945,
      "step": 1363
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7514,
      "step": 1364
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.727,
      "step": 1365
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9341,
      "step": 1366
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8012,
      "step": 1367
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7649,
      "step": 1368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8664,
      "step": 1369
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8466,
      "step": 1370
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9392,
      "step": 1371
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8166,
      "step": 1372
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9135,
      "step": 1373
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8317,
      "step": 1374
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9675,
      "step": 1375
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8231,
      "step": 1376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5382,
      "step": 1377
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8736,
      "step": 1378
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8208,
      "step": 1379
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 1380
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8718,
      "step": 1381
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6804,
      "step": 1382
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7236,
      "step": 1383
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 1.119,
      "step": 1384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8956,
      "step": 1385
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.9316,
      "step": 1386
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8747,
      "step": 1387
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8069,
      "step": 1388
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8237,
      "step": 1389
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7534,
      "step": 1390
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7788,
      "step": 1391
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 1.0281,
      "step": 1392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8001,
      "step": 1393
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8231,
      "step": 1394
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7892,
      "step": 1395
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.941,
      "step": 1396
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5953,
      "step": 1397
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.9596,
      "step": 1398
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7858,
      "step": 1399
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6043,
      "step": 1400
    },
    {
      "epoch": 0.27,
      "eval_loss": 0.825057864189148,
      "eval_runtime": 107.5613,
      "eval_samples_per_second": 9.297,
      "eval_steps_per_second": 4.649,
      "step": 1400
    },
    {
      "epoch": 0.27,
      "mmlu_eval_accuracy": 0.5261002332209418,
      "mmlu_eval_accuracy_abstract_algebra": 0.09090909090909091,
      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
      "mmlu_eval_accuracy_astronomy": 0.375,
      "mmlu_eval_accuracy_business_ethics": 0.36363636363636365,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.7272727272727273,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.2682926829268293,
      "mmlu_eval_accuracy_formal_logic": 0.42857142857142855,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.6875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_geography": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.7692307692307693,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.4782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.7692307692307693,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.68,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
      "mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.5882352941176471,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
      "mmlu_eval_accuracy_professional_law": 0.36470588235294116,
      "mmlu_eval_accuracy_professional_medicine": 0.3225806451612903,
      "mmlu_eval_accuracy_professional_psychology": 0.5072463768115942,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.8181818181818182,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.3333333333333333,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.0671936689484525,
      "step": 1400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8498,
      "step": 1401
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5981,
      "step": 1402
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7947,
      "step": 1403
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7712,
      "step": 1404
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.854,
      "step": 1405
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8629,
      "step": 1406
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6836,
      "step": 1407
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6803,
      "step": 1408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.558,
      "step": 1409
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8982,
      "step": 1410
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.914,
      "step": 1411
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5412,
      "step": 1412
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7332,
      "step": 1413
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.9627,
      "step": 1414
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8164,
      "step": 1415
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5844,
      "step": 1416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8248,
      "step": 1417
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 1.2164,
      "step": 1418
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8451,
      "step": 1419
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7261,
      "step": 1420
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6247,
      "step": 1421
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.837,
      "step": 1422
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 1.0168,
      "step": 1423
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8536,
      "step": 1424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7846,
      "step": 1425
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7055,
      "step": 1426
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6929,
      "step": 1427
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.717,
      "step": 1428
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7653,
      "step": 1429
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8804,
      "step": 1430
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7957,
      "step": 1431
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9933,
      "step": 1432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5636,
      "step": 1433
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7725,
      "step": 1434
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5726,
      "step": 1435
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8531,
      "step": 1436
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8233,
      "step": 1437
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 1.0856,
      "step": 1438
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.686,
      "step": 1439
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.673,
      "step": 1440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6611,
      "step": 1441
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6216,
      "step": 1442
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9692,
      "step": 1443
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7918,
      "step": 1444
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7039,
      "step": 1445
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7026,
      "step": 1446
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7067,
      "step": 1447
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.848,
      "step": 1448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9224,
      "step": 1449
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 1.0266,
      "step": 1450
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8069,
      "step": 1451
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8536,
      "step": 1452
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7476,
      "step": 1453
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8078,
      "step": 1454
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9525,
      "step": 1455
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.649,
      "step": 1456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8569,
      "step": 1457
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9695,
      "step": 1458
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 1.0852,
      "step": 1459
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8398,
      "step": 1460
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7028,
      "step": 1461
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7751,
      "step": 1462
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7162,
      "step": 1463
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 1.0478,
      "step": 1464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9565,
      "step": 1465
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.9746,
      "step": 1466
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6842,
      "step": 1467
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 1.0284,
      "step": 1468
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7857,
      "step": 1469
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6987,
      "step": 1470
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8753,
      "step": 1471
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7316,
      "step": 1472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7754,
      "step": 1473
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8902,
      "step": 1474
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7006,
      "step": 1475
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8815,
      "step": 1476
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8732,
      "step": 1477
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8545,
      "step": 1478
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7443,
      "step": 1479
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7169,
      "step": 1480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.898,
      "step": 1481
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8762,
      "step": 1482
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7574,
      "step": 1483
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9288,
      "step": 1484
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8225,
      "step": 1485
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.589,
      "step": 1486
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 1487
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8877,
      "step": 1488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8212,
      "step": 1489
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.938,
      "step": 1490
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8443,
      "step": 1491
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8144,
      "step": 1492
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5863,
      "step": 1493
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8429,
      "step": 1494
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8094,
      "step": 1495
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.889,
      "step": 1496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7291,
      "step": 1497
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9432,
      "step": 1498
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7819,
      "step": 1499
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7555,
      "step": 1500
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7687,
      "step": 1501
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8002,
      "step": 1502
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8006,
      "step": 1503
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8311,
      "step": 1504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.687,
      "step": 1505
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8844,
      "step": 1506
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.964,
      "step": 1507
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6618,
      "step": 1508
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8479,
      "step": 1509
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9227,
      "step": 1510
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8525,
      "step": 1511
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7547,
      "step": 1512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7651,
      "step": 1513
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8823,
      "step": 1514
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8152,
      "step": 1515
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8002,
      "step": 1516
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8979,
      "step": 1517
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7671,
      "step": 1518
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9666,
      "step": 1519
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9958,
      "step": 1520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7264,
      "step": 1521
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8294,
      "step": 1522
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8473,
      "step": 1523
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4256,
      "step": 1524
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 1.0817,
      "step": 1525
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8366,
      "step": 1526
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9516,
      "step": 1527
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5898,
      "step": 1528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7336,
      "step": 1529
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8132,
      "step": 1530
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7671,
      "step": 1531
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7801,
      "step": 1532
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 1.0315,
      "step": 1533
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5838,
      "step": 1534
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9304,
      "step": 1535
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.966,
      "step": 1536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7656,
      "step": 1537
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6905,
      "step": 1538
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.754,
      "step": 1539
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9846,
      "step": 1540
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8398,
      "step": 1541
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9174,
      "step": 1542
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6372,
      "step": 1543
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7296,
      "step": 1544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9294,
      "step": 1545
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7385,
      "step": 1546
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7534,
      "step": 1547
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7553,
      "step": 1548
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 1.0317,
      "step": 1549
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8087,
      "step": 1550
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6293,
      "step": 1551
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7301,
      "step": 1552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9018,
      "step": 1553
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.787,
      "step": 1554
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6241,
      "step": 1555
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5388,
      "step": 1556
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6282,
      "step": 1557
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8723,
      "step": 1558
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7209,
      "step": 1559
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7809,
      "step": 1560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5963,
      "step": 1561
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8324,
      "step": 1562
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6484,
      "step": 1563
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8241,
      "step": 1564
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5794,
      "step": 1565
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9628,
      "step": 1566
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7016,
      "step": 1567
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 1.0955,
      "step": 1568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8643,
      "step": 1569
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6446,
      "step": 1570
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9405,
      "step": 1571
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8703,
      "step": 1572
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6421,
      "step": 1573
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8491,
      "step": 1574
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7685,
      "step": 1575
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 1.0141,
      "step": 1576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8018,
      "step": 1577
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7598,
      "step": 1578
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9007,
      "step": 1579
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8525,
      "step": 1580
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8368,
      "step": 1581
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7355,
      "step": 1582
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6625,
      "step": 1583
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.841,
      "step": 1584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5787,
      "step": 1585
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8488,
      "step": 1586
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.3169,
      "step": 1587
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9284,
      "step": 1588
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8948,
      "step": 1589
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7806,
      "step": 1590
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.0476,
      "step": 1591
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5354,
      "step": 1592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8623,
      "step": 1593
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9032,
      "step": 1594
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7472,
      "step": 1595
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.1098,
      "step": 1596
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.412,
      "step": 1597
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.869,
      "step": 1598
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7577,
      "step": 1599
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8903,
      "step": 1600
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.8221507668495178,
      "eval_runtime": 107.8499,
      "eval_samples_per_second": 9.272,
      "eval_steps_per_second": 4.636,
      "step": 1600
    },
    {
      "epoch": 0.31,
      "mmlu_eval_accuracy": 0.5689444031497279,
      "mmlu_eval_accuracy_abstract_algebra": 0.09090909090909091,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.2727272727272727,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.6875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8571428571428571,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.8461538461538461,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7093023255813954,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.38,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 0.8437367426936683,
      "step": 1600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6816,
      "step": 1601
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6536,
      "step": 1602
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7733,
      "step": 1603
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.01,
      "step": 1604
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.0122,
      "step": 1605
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7618,
      "step": 1606
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8975,
      "step": 1607
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.818,
      "step": 1608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8361,
      "step": 1609
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5973,
      "step": 1610
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.057,
      "step": 1611
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.888,
      "step": 1612
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8184,
      "step": 1613
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7498,
      "step": 1614
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9438,
      "step": 1615
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8675,
      "step": 1616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7394,
      "step": 1617
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9155,
      "step": 1618
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.878,
      "step": 1619
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9278,
      "step": 1620
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 1.0698,
      "step": 1621
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6272,
      "step": 1622
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7526,
      "step": 1623
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9892,
      "step": 1624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8515,
      "step": 1625
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7535,
      "step": 1626
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9034,
      "step": 1627
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8185,
      "step": 1628
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8484,
      "step": 1629
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8227,
      "step": 1630
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9481,
      "step": 1631
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8076,
      "step": 1632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7329,
      "step": 1633
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6009,
      "step": 1634
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.9391,
      "step": 1635
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8281,
      "step": 1636
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9875,
      "step": 1637
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.705,
      "step": 1638
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6577,
      "step": 1639
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9632,
      "step": 1640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9511,
      "step": 1641
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9213,
      "step": 1642
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9555,
      "step": 1643
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7763,
      "step": 1644
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8044,
      "step": 1645
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8144,
      "step": 1646
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7764,
      "step": 1647
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7505,
      "step": 1648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9509,
      "step": 1649
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5935,
      "step": 1650
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9426,
      "step": 1651
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9142,
      "step": 1652
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 1.006,
      "step": 1653
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8634,
      "step": 1654
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 1.0491,
      "step": 1655
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8147,
      "step": 1656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6093,
      "step": 1657
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5288,
      "step": 1658
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7538,
      "step": 1659
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8122,
      "step": 1660
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8889,
      "step": 1661
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7966,
      "step": 1662
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8071,
      "step": 1663
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8796,
      "step": 1664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6675,
      "step": 1665
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8975,
      "step": 1666
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 1.0125,
      "step": 1667
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7006,
      "step": 1668
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9563,
      "step": 1669
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7269,
      "step": 1670
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9715,
      "step": 1671
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 1.0151,
      "step": 1672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6572,
      "step": 1673
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5888,
      "step": 1674
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9198,
      "step": 1675
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.9,
      "step": 1676
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7485,
      "step": 1677
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5904,
      "step": 1678
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6453,
      "step": 1679
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6435,
      "step": 1680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7521,
      "step": 1681
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8277,
      "step": 1682
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7807,
      "step": 1683
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7882,
      "step": 1684
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7859,
      "step": 1685
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 1.0744,
      "step": 1686
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 1687
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.8735,
      "step": 1688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5536,
      "step": 1689
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7903,
      "step": 1690
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 1.0194,
      "step": 1691
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7377,
      "step": 1692
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8358,
      "step": 1693
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.524,
      "step": 1694
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7802,
      "step": 1695
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6332,
      "step": 1696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9569,
      "step": 1697
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9413,
      "step": 1698
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 1.0077,
      "step": 1699
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7344,
      "step": 1700
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6809,
      "step": 1701
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7994,
      "step": 1702
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7187,
      "step": 1703
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7705,
      "step": 1704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9343,
      "step": 1705
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9431,
      "step": 1706
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7519,
      "step": 1707
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7143,
      "step": 1708
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8279,
      "step": 1709
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5734,
      "step": 1710
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9228,
      "step": 1711
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9653,
      "step": 1712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7466,
      "step": 1713
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9613,
      "step": 1714
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7572,
      "step": 1715
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7698,
      "step": 1716
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7907,
      "step": 1717
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7198,
      "step": 1718
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.866,
      "step": 1719
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7706,
      "step": 1720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6453,
      "step": 1721
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9814,
      "step": 1722
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5207,
      "step": 1723
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7474,
      "step": 1724
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.585,
      "step": 1725
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9092,
      "step": 1726
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8555,
      "step": 1727
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8701,
      "step": 1728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9159,
      "step": 1729
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9023,
      "step": 1730
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7989,
      "step": 1731
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7434,
      "step": 1732
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7934,
      "step": 1733
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8429,
      "step": 1734
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7717,
      "step": 1735
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7028,
      "step": 1736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7803,
      "step": 1737
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7464,
      "step": 1738
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 1.0388,
      "step": 1739
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8894,
      "step": 1740
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.812,
      "step": 1741
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8362,
      "step": 1742
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8536,
      "step": 1743
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9904,
      "step": 1744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7413,
      "step": 1745
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9807,
      "step": 1746
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8341,
      "step": 1747
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8223,
      "step": 1748
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7822,
      "step": 1749
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 1750
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4812,
      "step": 1751
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8285,
      "step": 1752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8178,
      "step": 1753
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6253,
      "step": 1754
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7521,
      "step": 1755
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8063,
      "step": 1756
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 1.0525,
      "step": 1757
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6469,
      "step": 1758
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5751,
      "step": 1759
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9385,
      "step": 1760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6278,
      "step": 1761
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8664,
      "step": 1762
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7919,
      "step": 1763
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6463,
      "step": 1764
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9496,
      "step": 1765
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5857,
      "step": 1766
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8012,
      "step": 1767
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7307,
      "step": 1768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 1.0323,
      "step": 1769
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9045,
      "step": 1770
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4746,
      "step": 1771
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7484,
      "step": 1772
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7377,
      "step": 1773
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9423,
      "step": 1774
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8146,
      "step": 1775
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8995,
      "step": 1776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5627,
      "step": 1777
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6484,
      "step": 1778
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9091,
      "step": 1779
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7198,
      "step": 1780
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6813,
      "step": 1781
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9187,
      "step": 1782
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9621,
      "step": 1783
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6419,
      "step": 1784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 1.2731,
      "step": 1785
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8213,
      "step": 1786
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8048,
      "step": 1787
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5645,
      "step": 1788
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8758,
      "step": 1789
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9589,
      "step": 1790
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7625,
      "step": 1791
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7643,
      "step": 1792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8189,
      "step": 1793
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8299,
      "step": 1794
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.789,
      "step": 1795
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7415,
      "step": 1796
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6801,
      "step": 1797
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 1.0205,
      "step": 1798
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8781,
      "step": 1799
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 1.0804,
      "step": 1800
    },
    {
      "epoch": 0.35,
      "eval_loss": 0.8146001100540161,
      "eval_runtime": 108.0941,
      "eval_samples_per_second": 9.251,
      "eval_steps_per_second": 4.626,
      "step": 1800
    },
    {
      "epoch": 0.35,
      "mmlu_eval_accuracy": 0.5759611012722413,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "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": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5121951219512195,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.4482758620689655,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.7666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.8,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.28,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.31176470588235294,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.0327665438831786,
      "step": 1800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8016,
      "step": 1801
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8261,
      "step": 1802
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6574,
      "step": 1803
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8124,
      "step": 1804
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5276,
      "step": 1805
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7502,
      "step": 1806
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 1.0469,
      "step": 1807
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9149,
      "step": 1808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.889,
      "step": 1809
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6277,
      "step": 1810
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7256,
      "step": 1811
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8466,
      "step": 1812
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8437,
      "step": 1813
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8968,
      "step": 1814
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8476,
      "step": 1815
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 1816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7074,
      "step": 1817
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9776,
      "step": 1818
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7861,
      "step": 1819
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9726,
      "step": 1820
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7608,
      "step": 1821
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8355,
      "step": 1822
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7307,
      "step": 1823
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7717,
      "step": 1824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.908,
      "step": 1825
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6057,
      "step": 1826
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6752,
      "step": 1827
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9067,
      "step": 1828
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8395,
      "step": 1829
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6725,
      "step": 1830
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7249,
      "step": 1831
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7055,
      "step": 1832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9172,
      "step": 1833
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9367,
      "step": 1834
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8796,
      "step": 1835
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8621,
      "step": 1836
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8216,
      "step": 1837
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 1.0159,
      "step": 1838
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9952,
      "step": 1839
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8941,
      "step": 1840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.9377,
      "step": 1841
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7245,
      "step": 1842
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7017,
      "step": 1843
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 1.0832,
      "step": 1844
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6605,
      "step": 1845
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8471,
      "step": 1846
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6213,
      "step": 1847
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8034,
      "step": 1848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7434,
      "step": 1849
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7257,
      "step": 1850
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6828,
      "step": 1851
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 1.0467,
      "step": 1852
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 1.0019,
      "step": 1853
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8737,
      "step": 1854
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6707,
      "step": 1855
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6337,
      "step": 1856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8495,
      "step": 1857
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7764,
      "step": 1858
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7458,
      "step": 1859
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 1860
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.9008,
      "step": 1861
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.843,
      "step": 1862
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8604,
      "step": 1863
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7122,
      "step": 1864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.856,
      "step": 1865
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.9627,
      "step": 1866
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8974,
      "step": 1867
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.986,
      "step": 1868
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8162,
      "step": 1869
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7787,
      "step": 1870
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8206,
      "step": 1871
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6129,
      "step": 1872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.9801,
      "step": 1873
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7778,
      "step": 1874
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6498,
      "step": 1875
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8788,
      "step": 1876
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7876,
      "step": 1877
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7567,
      "step": 1878
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8699,
      "step": 1879
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5723,
      "step": 1880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8008,
      "step": 1881
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.9557,
      "step": 1882
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 1.1339,
      "step": 1883
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7568,
      "step": 1884
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.9133,
      "step": 1885
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7743,
      "step": 1886
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7552,
      "step": 1887
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6108,
      "step": 1888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8523,
      "step": 1889
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7165,
      "step": 1890
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7073,
      "step": 1891
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.8273,
      "step": 1892
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.9308,
      "step": 1893
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5477,
      "step": 1894
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 1.0091,
      "step": 1895
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6657,
      "step": 1896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7054,
      "step": 1897
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.763,
      "step": 1898
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8852,
      "step": 1899
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7191,
      "step": 1900
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.767,
      "step": 1901
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6845,
      "step": 1902
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8597,
      "step": 1903
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0146,
      "step": 1904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8562,
      "step": 1905
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.9003,
      "step": 1906
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7635,
      "step": 1907
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.9138,
      "step": 1908
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6774,
      "step": 1909
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.668,
      "step": 1910
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.9251,
      "step": 1911
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.9255,
      "step": 1912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7345,
      "step": 1913
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8944,
      "step": 1914
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0152,
      "step": 1915
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8073,
      "step": 1916
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8764,
      "step": 1917
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6797,
      "step": 1918
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.639,
      "step": 1919
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6815,
      "step": 1920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0481,
      "step": 1921
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5657,
      "step": 1922
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7028,
      "step": 1923
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8974,
      "step": 1924
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.875,
      "step": 1925
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7509,
      "step": 1926
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.756,
      "step": 1927
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0423,
      "step": 1928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8452,
      "step": 1929
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7343,
      "step": 1930
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8304,
      "step": 1931
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8745,
      "step": 1932
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7954,
      "step": 1933
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7008,
      "step": 1934
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6662,
      "step": 1935
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5971,
      "step": 1936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.789,
      "step": 1937
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6848,
      "step": 1938
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8327,
      "step": 1939
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6904,
      "step": 1940
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7409,
      "step": 1941
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.9888,
      "step": 1942
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0448,
      "step": 1943
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0101,
      "step": 1944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7703,
      "step": 1945
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8223,
      "step": 1946
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0914,
      "step": 1947
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 1.0928,
      "step": 1948
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7851,
      "step": 1949
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6261,
      "step": 1950
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7146,
      "step": 1951
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8096,
      "step": 1952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7789,
      "step": 1953
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9109,
      "step": 1954
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8681,
      "step": 1955
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6915,
      "step": 1956
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7957,
      "step": 1957
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.833,
      "step": 1958
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6663,
      "step": 1959
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7429,
      "step": 1960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7705,
      "step": 1961
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9757,
      "step": 1962
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6649,
      "step": 1963
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6528,
      "step": 1964
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9402,
      "step": 1965
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6765,
      "step": 1966
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6915,
      "step": 1967
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9004,
      "step": 1968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9036,
      "step": 1969
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 1.128,
      "step": 1970
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8759,
      "step": 1971
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6407,
      "step": 1972
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6665,
      "step": 1973
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9832,
      "step": 1974
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.951,
      "step": 1975
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9842,
      "step": 1976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7394,
      "step": 1977
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9248,
      "step": 1978
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6551,
      "step": 1979
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6816,
      "step": 1980
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8979,
      "step": 1981
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8824,
      "step": 1982
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9177,
      "step": 1983
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8617,
      "step": 1984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6392,
      "step": 1985
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7309,
      "step": 1986
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9224,
      "step": 1987
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8798,
      "step": 1988
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7272,
      "step": 1989
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 1.0722,
      "step": 1990
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7289,
      "step": 1991
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9002,
      "step": 1992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8929,
      "step": 1993
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 1.0633,
      "step": 1994
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7856,
      "step": 1995
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9059,
      "step": 1996
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8342,
      "step": 1997
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8652,
      "step": 1998
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.8609,
      "step": 1999
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.9752,
      "step": 2000
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.8149585127830505,
      "eval_runtime": 108.0696,
      "eval_samples_per_second": 9.253,
      "eval_steps_per_second": 4.627,
      "step": 2000
    },
    {
      "epoch": 0.38,
      "mmlu_eval_accuracy": 0.5738805591482586,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.8076923076923077,
      "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": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.34,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5507246376811594,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.105458383538393,
      "step": 2000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7308,
      "step": 2001
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8644,
      "step": 2002
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8867,
      "step": 2003
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9645,
      "step": 2004
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6527,
      "step": 2005
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6629,
      "step": 2006
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8458,
      "step": 2007
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7732,
      "step": 2008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6468,
      "step": 2009
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8299,
      "step": 2010
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6831,
      "step": 2011
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5921,
      "step": 2012
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8091,
      "step": 2013
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7409,
      "step": 2014
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9298,
      "step": 2015
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8473,
      "step": 2016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5417,
      "step": 2017
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9532,
      "step": 2018
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7641,
      "step": 2019
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6915,
      "step": 2020
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7277,
      "step": 2021
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7725,
      "step": 2022
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 1.0329,
      "step": 2023
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8635,
      "step": 2024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.891,
      "step": 2025
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9792,
      "step": 2026
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8687,
      "step": 2027
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7852,
      "step": 2028
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7678,
      "step": 2029
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8241,
      "step": 2030
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7062,
      "step": 2031
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8912,
      "step": 2032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8203,
      "step": 2033
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7255,
      "step": 2034
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7639,
      "step": 2035
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7734,
      "step": 2036
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 1.0269,
      "step": 2037
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 1.1298,
      "step": 2038
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8837,
      "step": 2039
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5968,
      "step": 2040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8809,
      "step": 2041
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 1.0293,
      "step": 2042
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9419,
      "step": 2043
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 1.0109,
      "step": 2044
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9946,
      "step": 2045
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6345,
      "step": 2046
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8327,
      "step": 2047
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5433,
      "step": 2048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7982,
      "step": 2049
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.9287,
      "step": 2050
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7342,
      "step": 2051
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.716,
      "step": 2052
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0033,
      "step": 2053
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8942,
      "step": 2054
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6888,
      "step": 2055
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9211,
      "step": 2056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7433,
      "step": 2057
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.742,
      "step": 2058
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0346,
      "step": 2059
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8035,
      "step": 2060
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8608,
      "step": 2061
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0434,
      "step": 2062
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8506,
      "step": 2063
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8903,
      "step": 2064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 2065
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.1097,
      "step": 2066
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9261,
      "step": 2067
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0323,
      "step": 2068
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7585,
      "step": 2069
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9306,
      "step": 2070
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9677,
      "step": 2071
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5895,
      "step": 2072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6045,
      "step": 2073
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6815,
      "step": 2074
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0023,
      "step": 2075
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0172,
      "step": 2076
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8858,
      "step": 2077
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8824,
      "step": 2078
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7991,
      "step": 2079
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0729,
      "step": 2080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8214,
      "step": 2081
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8925,
      "step": 2082
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.755,
      "step": 2083
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.1572,
      "step": 2084
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.941,
      "step": 2085
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9525,
      "step": 2086
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9721,
      "step": 2087
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7794,
      "step": 2088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7309,
      "step": 2089
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.628,
      "step": 2090
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8088,
      "step": 2091
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7635,
      "step": 2092
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8365,
      "step": 2093
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6425,
      "step": 2094
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8185,
      "step": 2095
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9957,
      "step": 2096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6733,
      "step": 2097
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6154,
      "step": 2098
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 1.0058,
      "step": 2099
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7598,
      "step": 2100
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6487,
      "step": 2101
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7136,
      "step": 2102
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8475,
      "step": 2103
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8936,
      "step": 2104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5463,
      "step": 2105
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9124,
      "step": 2106
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9203,
      "step": 2107
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5896,
      "step": 2108
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7218,
      "step": 2109
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7588,
      "step": 2110
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6434,
      "step": 2111
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7581,
      "step": 2112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 1.1152,
      "step": 2113
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9519,
      "step": 2114
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5531,
      "step": 2115
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8028,
      "step": 2116
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9286,
      "step": 2117
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9637,
      "step": 2118
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9171,
      "step": 2119
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.545,
      "step": 2120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.805,
      "step": 2121
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8533,
      "step": 2122
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7983,
      "step": 2123
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7136,
      "step": 2124
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6182,
      "step": 2125
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9107,
      "step": 2126
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6259,
      "step": 2127
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9567,
      "step": 2128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9105,
      "step": 2129
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8831,
      "step": 2130
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9322,
      "step": 2131
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8539,
      "step": 2132
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7234,
      "step": 2133
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8665,
      "step": 2134
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8813,
      "step": 2135
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.86,
      "step": 2136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9041,
      "step": 2137
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8649,
      "step": 2138
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6517,
      "step": 2139
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8964,
      "step": 2140
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7097,
      "step": 2141
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8851,
      "step": 2142
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8768,
      "step": 2143
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7991,
      "step": 2144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9392,
      "step": 2145
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5467,
      "step": 2146
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9682,
      "step": 2147
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8882,
      "step": 2148
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8174,
      "step": 2149
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7441,
      "step": 2150
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8576,
      "step": 2151
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9202,
      "step": 2152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.9291,
      "step": 2153
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.954,
      "step": 2154
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.93,
      "step": 2155
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6837,
      "step": 2156
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7823,
      "step": 2157
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 1.1254,
      "step": 2158
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5681,
      "step": 2159
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8552,
      "step": 2160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6315,
      "step": 2161
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6988,
      "step": 2162
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7536,
      "step": 2163
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8722,
      "step": 2164
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8341,
      "step": 2165
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9531,
      "step": 2166
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7694,
      "step": 2167
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5121,
      "step": 2168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.741,
      "step": 2169
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8716,
      "step": 2170
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6412,
      "step": 2171
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5901,
      "step": 2172
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9068,
      "step": 2173
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7652,
      "step": 2174
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8821,
      "step": 2175
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9705,
      "step": 2176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8512,
      "step": 2177
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 1.0384,
      "step": 2178
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.706,
      "step": 2179
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.65,
      "step": 2180
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8064,
      "step": 2181
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8046,
      "step": 2182
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9499,
      "step": 2183
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5965,
      "step": 2184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7451,
      "step": 2185
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 2186
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9094,
      "step": 2187
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7085,
      "step": 2188
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8265,
      "step": 2189
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8682,
      "step": 2190
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9477,
      "step": 2191
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9759,
      "step": 2192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8473,
      "step": 2193
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 1.0222,
      "step": 2194
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7179,
      "step": 2195
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.9366,
      "step": 2196
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7656,
      "step": 2197
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8445,
      "step": 2198
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7999,
      "step": 2199
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.894,
      "step": 2200
    },
    {
      "epoch": 0.42,
      "eval_loss": 0.8177380561828613,
      "eval_runtime": 108.0401,
      "eval_samples_per_second": 9.256,
      "eval_steps_per_second": 4.628,
      "step": 2200
    },
    {
      "epoch": 0.42,
      "mmlu_eval_accuracy": 0.5416379928605715,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.29,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.3225806451612903,
      "mmlu_eval_accuracy_professional_law": 0.34705882352941175,
      "mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.005314374892114,
      "step": 2200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8848,
      "step": 2201
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7956,
      "step": 2202
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 1.0595,
      "step": 2203
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6619,
      "step": 2204
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 1.0131,
      "step": 2205
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.8641,
      "step": 2206
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 1.0132,
      "step": 2207
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8946,
      "step": 2208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8528,
      "step": 2209
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9401,
      "step": 2210
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7828,
      "step": 2211
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7281,
      "step": 2212
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7198,
      "step": 2213
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7603,
      "step": 2214
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 1.0791,
      "step": 2215
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9247,
      "step": 2216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3889,
      "step": 2217
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.62,
      "step": 2218
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6235,
      "step": 2219
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.774,
      "step": 2220
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 1.1101,
      "step": 2221
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.772,
      "step": 2222
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9038,
      "step": 2223
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7867,
      "step": 2224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9549,
      "step": 2225
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8205,
      "step": 2226
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8256,
      "step": 2227
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9343,
      "step": 2228
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8522,
      "step": 2229
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 1.0069,
      "step": 2230
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7696,
      "step": 2231
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7041,
      "step": 2232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6866,
      "step": 2233
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9121,
      "step": 2234
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6566,
      "step": 2235
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.718,
      "step": 2236
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7678,
      "step": 2237
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5789,
      "step": 2238
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5531,
      "step": 2239
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9033,
      "step": 2240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.891,
      "step": 2241
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6672,
      "step": 2242
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.653,
      "step": 2243
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 1.0025,
      "step": 2244
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9605,
      "step": 2245
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9797,
      "step": 2246
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6619,
      "step": 2247
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8861,
      "step": 2248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9175,
      "step": 2249
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.9279,
      "step": 2250
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8795,
      "step": 2251
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.848,
      "step": 2252
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 1.0087,
      "step": 2253
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7884,
      "step": 2254
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 1.0761,
      "step": 2255
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8049,
      "step": 2256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.902,
      "step": 2257
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8886,
      "step": 2258
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8743,
      "step": 2259
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 1.1583,
      "step": 2260
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 2261
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8413,
      "step": 2262
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7919,
      "step": 2263
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7922,
      "step": 2264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7493,
      "step": 2265
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6961,
      "step": 2266
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8163,
      "step": 2267
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 1.0211,
      "step": 2268
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5468,
      "step": 2269
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.945,
      "step": 2270
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9233,
      "step": 2271
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.831,
      "step": 2272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9077,
      "step": 2273
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.862,
      "step": 2274
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9526,
      "step": 2275
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6042,
      "step": 2276
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8155,
      "step": 2277
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8278,
      "step": 2278
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8279,
      "step": 2279
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6948,
      "step": 2280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7439,
      "step": 2281
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8155,
      "step": 2282
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7576,
      "step": 2283
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5613,
      "step": 2284
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5458,
      "step": 2285
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6473,
      "step": 2286
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9188,
      "step": 2287
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7833,
      "step": 2288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 1.0366,
      "step": 2289
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8819,
      "step": 2290
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 1.1105,
      "step": 2291
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6863,
      "step": 2292
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 1.0806,
      "step": 2293
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7539,
      "step": 2294
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6476,
      "step": 2295
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8803,
      "step": 2296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7202,
      "step": 2297
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7878,
      "step": 2298
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9729,
      "step": 2299
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8873,
      "step": 2300
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7658,
      "step": 2301
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6554,
      "step": 2302
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8215,
      "step": 2303
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6565,
      "step": 2304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8141,
      "step": 2305
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5487,
      "step": 2306
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.815,
      "step": 2307
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7608,
      "step": 2308
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9029,
      "step": 2309
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.9271,
      "step": 2310
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8714,
      "step": 2311
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5332,
      "step": 2312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7826,
      "step": 2313
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7414,
      "step": 2314
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8224,
      "step": 2315
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 1.003,
      "step": 2316
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8252,
      "step": 2317
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8574,
      "step": 2318
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7941,
      "step": 2319
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9149,
      "step": 2320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9722,
      "step": 2321
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8927,
      "step": 2322
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8357,
      "step": 2323
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6254,
      "step": 2324
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7226,
      "step": 2325
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.759,
      "step": 2326
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.881,
      "step": 2327
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8117,
      "step": 2328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.656,
      "step": 2329
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8083,
      "step": 2330
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 1.2686,
      "step": 2331
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6736,
      "step": 2332
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5367,
      "step": 2333
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8158,
      "step": 2334
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.77,
      "step": 2335
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6481,
      "step": 2336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5888,
      "step": 2337
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.671,
      "step": 2338
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6305,
      "step": 2339
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9366,
      "step": 2340
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9533,
      "step": 2341
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8109,
      "step": 2342
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9133,
      "step": 2343
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8661,
      "step": 2344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 1.1764,
      "step": 2345
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8569,
      "step": 2346
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8668,
      "step": 2347
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8258,
      "step": 2348
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7445,
      "step": 2349
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9121,
      "step": 2350
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8903,
      "step": 2351
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 1.0068,
      "step": 2352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8355,
      "step": 2353
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8963,
      "step": 2354
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 1.1243,
      "step": 2355
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8218,
      "step": 2356
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5972,
      "step": 2357
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.9975,
      "step": 2358
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7838,
      "step": 2359
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6951,
      "step": 2360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.658,
      "step": 2361
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 1.0234,
      "step": 2362
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.8804,
      "step": 2363
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8345,
      "step": 2364
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9587,
      "step": 2365
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7751,
      "step": 2366
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 2367
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9796,
      "step": 2368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4751,
      "step": 2369
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7981,
      "step": 2370
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9075,
      "step": 2371
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.0665,
      "step": 2372
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.0116,
      "step": 2373
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.824,
      "step": 2374
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5336,
      "step": 2375
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7978,
      "step": 2376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9311,
      "step": 2377
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8532,
      "step": 2378
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.753,
      "step": 2379
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7889,
      "step": 2380
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7652,
      "step": 2381
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8783,
      "step": 2382
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5376,
      "step": 2383
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7477,
      "step": 2384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8446,
      "step": 2385
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8288,
      "step": 2386
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.0405,
      "step": 2387
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8691,
      "step": 2388
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4569,
      "step": 2389
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7414,
      "step": 2390
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 2391
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9894,
      "step": 2392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8463,
      "step": 2393
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.0096,
      "step": 2394
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9081,
      "step": 2395
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.0624,
      "step": 2396
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.547,
      "step": 2397
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9159,
      "step": 2398
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8307,
      "step": 2399
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.0275,
      "step": 2400
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.8227358460426331,
      "eval_runtime": 107.9665,
      "eval_samples_per_second": 9.262,
      "eval_steps_per_second": 4.631,
      "step": 2400
    },
    {
      "epoch": 0.46,
      "mmlu_eval_accuracy": 0.5601980072854039,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6551724137931034,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.09090909090909091,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5,
      "mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.8181818181818182,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
      "mmlu_eval_accuracy_moral_scenarios": 0.38,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_law": 0.3588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5507246376811594,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 0.871043609314192,
      "step": 2400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6021,
      "step": 2401
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.611,
      "step": 2402
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8174,
      "step": 2403
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8978,
      "step": 2404
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9202,
      "step": 2405
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9315,
      "step": 2406
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8926,
      "step": 2407
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8469,
      "step": 2408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6786,
      "step": 2409
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8899,
      "step": 2410
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.9438,
      "step": 2411
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8449,
      "step": 2412
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6292,
      "step": 2413
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7881,
      "step": 2414
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8872,
      "step": 2415
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8654,
      "step": 2416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9059,
      "step": 2417
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7544,
      "step": 2418
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6964,
      "step": 2419
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.863,
      "step": 2420
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 1.0355,
      "step": 2421
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9555,
      "step": 2422
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7699,
      "step": 2423
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9966,
      "step": 2424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7141,
      "step": 2425
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8409,
      "step": 2426
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7712,
      "step": 2427
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7263,
      "step": 2428
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9799,
      "step": 2429
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9629,
      "step": 2430
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7028,
      "step": 2431
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 1.0114,
      "step": 2432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9042,
      "step": 2433
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4401,
      "step": 2434
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8192,
      "step": 2435
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7718,
      "step": 2436
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9107,
      "step": 2437
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 1.0165,
      "step": 2438
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7129,
      "step": 2439
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4188,
      "step": 2440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8093,
      "step": 2441
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7248,
      "step": 2442
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6833,
      "step": 2443
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8692,
      "step": 2444
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 1.0222,
      "step": 2445
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9136,
      "step": 2446
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8741,
      "step": 2447
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8395,
      "step": 2448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8223,
      "step": 2449
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.547,
      "step": 2450
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.835,
      "step": 2451
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5637,
      "step": 2452
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9422,
      "step": 2453
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7259,
      "step": 2454
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7345,
      "step": 2455
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.77,
      "step": 2456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7593,
      "step": 2457
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6853,
      "step": 2458
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7403,
      "step": 2459
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7333,
      "step": 2460
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7412,
      "step": 2461
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9952,
      "step": 2462
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8078,
      "step": 2463
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.9261,
      "step": 2464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7814,
      "step": 2465
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7009,
      "step": 2466
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7777,
      "step": 2467
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6305,
      "step": 2468
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9309,
      "step": 2469
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7214,
      "step": 2470
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8001,
      "step": 2471
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7112,
      "step": 2472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6989,
      "step": 2473
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8005,
      "step": 2474
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6524,
      "step": 2475
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6657,
      "step": 2476
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7659,
      "step": 2477
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8311,
      "step": 2478
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8947,
      "step": 2479
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7761,
      "step": 2480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7591,
      "step": 2481
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7501,
      "step": 2482
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8377,
      "step": 2483
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6014,
      "step": 2484
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8565,
      "step": 2485
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8122,
      "step": 2486
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7954,
      "step": 2487
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8426,
      "step": 2488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9018,
      "step": 2489
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6567,
      "step": 2490
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8823,
      "step": 2491
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7025,
      "step": 2492
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9858,
      "step": 2493
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9202,
      "step": 2494
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8154,
      "step": 2495
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6378,
      "step": 2496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7726,
      "step": 2497
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7794,
      "step": 2498
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 1.12,
      "step": 2499
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7789,
      "step": 2500
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8035,
      "step": 2501
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9508,
      "step": 2502
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.78,
      "step": 2503
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9804,
      "step": 2504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6855,
      "step": 2505
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8847,
      "step": 2506
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.848,
      "step": 2507
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.832,
      "step": 2508
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5001,
      "step": 2509
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8495,
      "step": 2510
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9432,
      "step": 2511
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.854,
      "step": 2512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7702,
      "step": 2513
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9013,
      "step": 2514
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6056,
      "step": 2515
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.9455,
      "step": 2516
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8339,
      "step": 2517
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7624,
      "step": 2518
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.8653,
      "step": 2519
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6067,
      "step": 2520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7719,
      "step": 2521
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7582,
      "step": 2522
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8597,
      "step": 2523
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 1.3708,
      "step": 2524
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9456,
      "step": 2525
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9209,
      "step": 2526
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9611,
      "step": 2527
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 1.0301,
      "step": 2528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.597,
      "step": 2529
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6622,
      "step": 2530
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9098,
      "step": 2531
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8371,
      "step": 2532
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.97,
      "step": 2533
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9332,
      "step": 2534
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8993,
      "step": 2535
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9386,
      "step": 2536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.81,
      "step": 2537
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8203,
      "step": 2538
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7876,
      "step": 2539
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9323,
      "step": 2540
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6756,
      "step": 2541
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8619,
      "step": 2542
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7727,
      "step": 2543
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8566,
      "step": 2544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7615,
      "step": 2545
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6994,
      "step": 2546
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.799,
      "step": 2547
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8194,
      "step": 2548
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8458,
      "step": 2549
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8907,
      "step": 2550
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 1.0218,
      "step": 2551
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9161,
      "step": 2552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8088,
      "step": 2553
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.967,
      "step": 2554
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.841,
      "step": 2555
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.659,
      "step": 2556
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7203,
      "step": 2557
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3754,
      "step": 2558
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8392,
      "step": 2559
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 1.0582,
      "step": 2560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7758,
      "step": 2561
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9556,
      "step": 2562
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5456,
      "step": 2563
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8772,
      "step": 2564
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 1.1292,
      "step": 2565
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9821,
      "step": 2566
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.9393,
      "step": 2567
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6412,
      "step": 2568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8271,
      "step": 2569
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6897,
      "step": 2570
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8608,
      "step": 2571
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.896,
      "step": 2572
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7169,
      "step": 2573
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9325,
      "step": 2574
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7816,
      "step": 2575
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.764,
      "step": 2576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9851,
      "step": 2577
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8324,
      "step": 2578
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9043,
      "step": 2579
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 1.1322,
      "step": 2580
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9828,
      "step": 2581
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.69,
      "step": 2582
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6951,
      "step": 2583
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.75,
      "step": 2584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8248,
      "step": 2585
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8326,
      "step": 2586
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.886,
      "step": 2587
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5709,
      "step": 2588
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7761,
      "step": 2589
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8333,
      "step": 2590
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8544,
      "step": 2591
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7103,
      "step": 2592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 2593
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7435,
      "step": 2594
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8399,
      "step": 2595
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8271,
      "step": 2596
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9368,
      "step": 2597
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7112,
      "step": 2598
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7735,
      "step": 2599
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 1.0317,
      "step": 2600
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.8142709136009216,
      "eval_runtime": 107.8934,
      "eval_samples_per_second": 9.268,
      "eval_steps_per_second": 4.634,
      "step": 2600
    },
    {
      "epoch": 0.5,
      "mmlu_eval_accuracy": 0.5584228679719467,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.18181818181818182,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.22727272727272727,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.13793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.36470588235294116,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5362318840579711,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.8181818181818182,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.0159508551692216,
      "step": 2600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7568,
      "step": 2601
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7916,
      "step": 2602
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8148,
      "step": 2603
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7766,
      "step": 2604
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8777,
      "step": 2605
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.491,
      "step": 2606
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9363,
      "step": 2607
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6196,
      "step": 2608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.917,
      "step": 2609
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9716,
      "step": 2610
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6052,
      "step": 2611
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8019,
      "step": 2612
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.89,
      "step": 2613
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7948,
      "step": 2614
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7145,
      "step": 2615
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8018,
      "step": 2616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8606,
      "step": 2617
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6563,
      "step": 2618
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7774,
      "step": 2619
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6276,
      "step": 2620
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6386,
      "step": 2621
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5339,
      "step": 2622
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7127,
      "step": 2623
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8414,
      "step": 2624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9041,
      "step": 2625
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7482,
      "step": 2626
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7791,
      "step": 2627
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9256,
      "step": 2628
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8123,
      "step": 2629
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8268,
      "step": 2630
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7897,
      "step": 2631
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8501,
      "step": 2632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.618,
      "step": 2633
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.75,
      "step": 2634
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7926,
      "step": 2635
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9108,
      "step": 2636
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6465,
      "step": 2637
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7474,
      "step": 2638
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.921,
      "step": 2639
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 1.0146,
      "step": 2640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8337,
      "step": 2641
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9109,
      "step": 2642
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7484,
      "step": 2643
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8368,
      "step": 2644
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8584,
      "step": 2645
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.915,
      "step": 2646
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9244,
      "step": 2647
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8439,
      "step": 2648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6943,
      "step": 2649
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7258,
      "step": 2650
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.705,
      "step": 2651
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7286,
      "step": 2652
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8415,
      "step": 2653
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7969,
      "step": 2654
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8188,
      "step": 2655
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 1.0434,
      "step": 2656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9776,
      "step": 2657
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7722,
      "step": 2658
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5606,
      "step": 2659
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5965,
      "step": 2660
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 1.1235,
      "step": 2661
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9431,
      "step": 2662
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 1.0722,
      "step": 2663
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6177,
      "step": 2664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.9522,
      "step": 2665
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6252,
      "step": 2666
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7976,
      "step": 2667
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7437,
      "step": 2668
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7523,
      "step": 2669
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8138,
      "step": 2670
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 1.0007,
      "step": 2671
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8412,
      "step": 2672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 1.0941,
      "step": 2673
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.7852,
      "step": 2674
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8645,
      "step": 2675
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8187,
      "step": 2676
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9788,
      "step": 2677
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8954,
      "step": 2678
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9454,
      "step": 2679
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8629,
      "step": 2680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9379,
      "step": 2681
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8188,
      "step": 2682
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9039,
      "step": 2683
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7539,
      "step": 2684
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9876,
      "step": 2685
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7586,
      "step": 2686
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8564,
      "step": 2687
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8851,
      "step": 2688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8176,
      "step": 2689
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7864,
      "step": 2690
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7586,
      "step": 2691
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9661,
      "step": 2692
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6785,
      "step": 2693
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7177,
      "step": 2694
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 1.1491,
      "step": 2695
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9702,
      "step": 2696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9576,
      "step": 2697
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6987,
      "step": 2698
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6798,
      "step": 2699
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8404,
      "step": 2700
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8236,
      "step": 2701
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7174,
      "step": 2702
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8287,
      "step": 2703
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7962,
      "step": 2704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8313,
      "step": 2705
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9817,
      "step": 2706
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6993,
      "step": 2707
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6561,
      "step": 2708
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6688,
      "step": 2709
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6386,
      "step": 2710
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8458,
      "step": 2711
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8285,
      "step": 2712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8986,
      "step": 2713
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9368,
      "step": 2714
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5017,
      "step": 2715
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7048,
      "step": 2716
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7693,
      "step": 2717
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8364,
      "step": 2718
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.897,
      "step": 2719
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8525,
      "step": 2720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8309,
      "step": 2721
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8689,
      "step": 2722
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9083,
      "step": 2723
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6953,
      "step": 2724
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9651,
      "step": 2725
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7324,
      "step": 2726
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.9432,
      "step": 2727
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9137,
      "step": 2728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7916,
      "step": 2729
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.714,
      "step": 2730
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7905,
      "step": 2731
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7503,
      "step": 2732
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8167,
      "step": 2733
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9315,
      "step": 2734
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8437,
      "step": 2735
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6618,
      "step": 2736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9736,
      "step": 2737
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7666,
      "step": 2738
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8734,
      "step": 2739
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6573,
      "step": 2740
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8603,
      "step": 2741
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 1.2405,
      "step": 2742
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7858,
      "step": 2743
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6687,
      "step": 2744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8774,
      "step": 2745
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8304,
      "step": 2746
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7555,
      "step": 2747
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7692,
      "step": 2748
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7099,
      "step": 2749
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6054,
      "step": 2750
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6203,
      "step": 2751
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6055,
      "step": 2752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8435,
      "step": 2753
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.688,
      "step": 2754
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.771,
      "step": 2755
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6123,
      "step": 2756
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7265,
      "step": 2757
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6294,
      "step": 2758
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8518,
      "step": 2759
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9845,
      "step": 2760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8441,
      "step": 2761
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8425,
      "step": 2762
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9727,
      "step": 2763
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5493,
      "step": 2764
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 1.0171,
      "step": 2765
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8467,
      "step": 2766
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 1.1141,
      "step": 2767
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8054,
      "step": 2768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9188,
      "step": 2769
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8368,
      "step": 2770
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.768,
      "step": 2771
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7098,
      "step": 2772
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7929,
      "step": 2773
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6201,
      "step": 2774
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8564,
      "step": 2775
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6464,
      "step": 2776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7632,
      "step": 2777
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8253,
      "step": 2778
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9474,
      "step": 2779
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9539,
      "step": 2780
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6101,
      "step": 2781
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7937,
      "step": 2782
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8068,
      "step": 2783
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8304,
      "step": 2784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7049,
      "step": 2785
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.719,
      "step": 2786
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8079,
      "step": 2787
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8154,
      "step": 2788
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9495,
      "step": 2789
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5459,
      "step": 2790
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5969,
      "step": 2791
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.917,
      "step": 2792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9176,
      "step": 2793
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8145,
      "step": 2794
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8318,
      "step": 2795
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8763,
      "step": 2796
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8708,
      "step": 2797
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9576,
      "step": 2798
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8426,
      "step": 2799
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8078,
      "step": 2800
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.8125653266906738,
      "eval_runtime": 107.8904,
      "eval_samples_per_second": 9.269,
      "eval_steps_per_second": 4.634,
      "step": 2800
    },
    {
      "epoch": 0.54,
      "mmlu_eval_accuracy": 0.5293441870662119,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.5,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "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": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "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": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "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": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.06896551724137931,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.5217391304347826,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
      "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.6470588235294118,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.38823529411764707,
      "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.4074074074074074,
      "mmlu_eval_accuracy_sociology": 0.6363636363636364,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.0585868249674069,
      "step": 2800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8149,
      "step": 2801
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7395,
      "step": 2802
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8277,
      "step": 2803
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.664,
      "step": 2804
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6743,
      "step": 2805
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7636,
      "step": 2806
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7867,
      "step": 2807
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6622,
      "step": 2808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.745,
      "step": 2809
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 1.0689,
      "step": 2810
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8589,
      "step": 2811
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 1.1289,
      "step": 2812
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7215,
      "step": 2813
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8739,
      "step": 2814
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8323,
      "step": 2815
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7386,
      "step": 2816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8335,
      "step": 2817
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9259,
      "step": 2818
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 1.0855,
      "step": 2819
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9255,
      "step": 2820
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 1.0472,
      "step": 2821
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6921,
      "step": 2822
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6134,
      "step": 2823
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 1.1317,
      "step": 2824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7024,
      "step": 2825
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6909,
      "step": 2826
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9214,
      "step": 2827
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7086,
      "step": 2828
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.7378,
      "step": 2829
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8698,
      "step": 2830
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8446,
      "step": 2831
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8104,
      "step": 2832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3772,
      "step": 2833
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6191,
      "step": 2834
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.883,
      "step": 2835
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.9992,
      "step": 2836
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6016,
      "step": 2837
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.0174,
      "step": 2838
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.9253,
      "step": 2839
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.731,
      "step": 2840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8446,
      "step": 2841
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7004,
      "step": 2842
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7689,
      "step": 2843
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8618,
      "step": 2844
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.034,
      "step": 2845
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.0211,
      "step": 2846
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.823,
      "step": 2847
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8862,
      "step": 2848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8412,
      "step": 2849
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7605,
      "step": 2850
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.9015,
      "step": 2851
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.0364,
      "step": 2852
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8528,
      "step": 2853
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.9186,
      "step": 2854
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8147,
      "step": 2855
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6097,
      "step": 2856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.9267,
      "step": 2857
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7924,
      "step": 2858
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8592,
      "step": 2859
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7057,
      "step": 2860
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8973,
      "step": 2861
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8334,
      "step": 2862
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7568,
      "step": 2863
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8868,
      "step": 2864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.782,
      "step": 2865
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8502,
      "step": 2866
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8374,
      "step": 2867
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7538,
      "step": 2868
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6918,
      "step": 2869
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6043,
      "step": 2870
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.0515,
      "step": 2871
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.1502,
      "step": 2872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7786,
      "step": 2873
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.81,
      "step": 2874
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6991,
      "step": 2875
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8271,
      "step": 2876
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6753,
      "step": 2877
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6544,
      "step": 2878
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8559,
      "step": 2879
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7543,
      "step": 2880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5401,
      "step": 2881
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.9702,
      "step": 2882
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8001,
      "step": 2883
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6811,
      "step": 2884
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9198,
      "step": 2885
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7131,
      "step": 2886
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9092,
      "step": 2887
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8042,
      "step": 2888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7018,
      "step": 2889
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7641,
      "step": 2890
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8303,
      "step": 2891
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.704,
      "step": 2892
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 1.0971,
      "step": 2893
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.661,
      "step": 2894
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7684,
      "step": 2895
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7317,
      "step": 2896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8827,
      "step": 2897
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7972,
      "step": 2898
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.873,
      "step": 2899
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8535,
      "step": 2900
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.759,
      "step": 2901
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8282,
      "step": 2902
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7869,
      "step": 2903
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7434,
      "step": 2904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.863,
      "step": 2905
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9593,
      "step": 2906
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7688,
      "step": 2907
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7351,
      "step": 2908
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5857,
      "step": 2909
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6734,
      "step": 2910
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9114,
      "step": 2911
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6888,
      "step": 2912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9141,
      "step": 2913
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9618,
      "step": 2914
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9241,
      "step": 2915
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9624,
      "step": 2916
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6741,
      "step": 2917
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8834,
      "step": 2918
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 1.1509,
      "step": 2919
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9862,
      "step": 2920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7443,
      "step": 2921
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7656,
      "step": 2922
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8006,
      "step": 2923
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9015,
      "step": 2924
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7512,
      "step": 2925
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6401,
      "step": 2926
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6969,
      "step": 2927
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.658,
      "step": 2928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9087,
      "step": 2929
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7381,
      "step": 2930
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9261,
      "step": 2931
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8332,
      "step": 2932
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.8891,
      "step": 2933
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6632,
      "step": 2934
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.9032,
      "step": 2935
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9133,
      "step": 2936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.694,
      "step": 2937
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9093,
      "step": 2938
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.817,
      "step": 2939
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9667,
      "step": 2940
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6794,
      "step": 2941
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8412,
      "step": 2942
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7151,
      "step": 2943
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7756,
      "step": 2944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6441,
      "step": 2945
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6674,
      "step": 2946
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6487,
      "step": 2947
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8969,
      "step": 2948
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7883,
      "step": 2949
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7675,
      "step": 2950
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8792,
      "step": 2951
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9758,
      "step": 2952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5819,
      "step": 2953
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8457,
      "step": 2954
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7322,
      "step": 2955
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7617,
      "step": 2956
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6863,
      "step": 2957
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.706,
      "step": 2958
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.663,
      "step": 2959
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7078,
      "step": 2960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8218,
      "step": 2961
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9403,
      "step": 2962
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8158,
      "step": 2963
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8269,
      "step": 2964
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9624,
      "step": 2965
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8321,
      "step": 2966
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.944,
      "step": 2967
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 2968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 2969
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7698,
      "step": 2970
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 1.0871,
      "step": 2971
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8547,
      "step": 2972
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8711,
      "step": 2973
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5715,
      "step": 2974
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8771,
      "step": 2975
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6301,
      "step": 2976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8078,
      "step": 2977
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7224,
      "step": 2978
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.989,
      "step": 2979
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.847,
      "step": 2980
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9608,
      "step": 2981
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9068,
      "step": 2982
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7749,
      "step": 2983
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 1.0827,
      "step": 2984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.9586,
      "step": 2985
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5333,
      "step": 2986
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.8201,
      "step": 2987
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7648,
      "step": 2988
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8208,
      "step": 2989
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9312,
      "step": 2990
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9856,
      "step": 2991
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7667,
      "step": 2992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8876,
      "step": 2993
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 1.0616,
      "step": 2994
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7746,
      "step": 2995
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8519,
      "step": 2996
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8316,
      "step": 2997
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9393,
      "step": 2998
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7587,
      "step": 2999
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8048,
      "step": 3000
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.8166067004203796,
      "eval_runtime": 107.7872,
      "eval_samples_per_second": 9.278,
      "eval_steps_per_second": 4.639,
      "step": 3000
    },
    {
      "epoch": 0.58,
      "mmlu_eval_accuracy": 0.5289063988783227,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.5,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.09090909090909091,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "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": 0.22727272727272727,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.7666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.8,
      "mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.631578947368421,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.5588235294117647,
      "mmlu_eval_accuracy_prehistory": 0.7714285714285715,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.35294117647058826,
      "mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.37037037037037035,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
      "mmlu_loss": 0.8663900064659492,
      "step": 3000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8332,
      "step": 3001
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6512,
      "step": 3002
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8917,
      "step": 3003
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6116,
      "step": 3004
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7517,
      "step": 3005
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8221,
      "step": 3006
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8256,
      "step": 3007
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8428,
      "step": 3008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.679,
      "step": 3009
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7836,
      "step": 3010
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8759,
      "step": 3011
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.928,
      "step": 3012
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9425,
      "step": 3013
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9957,
      "step": 3014
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7464,
      "step": 3015
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7922,
      "step": 3016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 1.0403,
      "step": 3017
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5892,
      "step": 3018
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7488,
      "step": 3019
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 1.0873,
      "step": 3020
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9428,
      "step": 3021
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8034,
      "step": 3022
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8666,
      "step": 3023
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8585,
      "step": 3024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9168,
      "step": 3025
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9171,
      "step": 3026
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8569,
      "step": 3027
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7058,
      "step": 3028
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9012,
      "step": 3029
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 1.143,
      "step": 3030
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7926,
      "step": 3031
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4878,
      "step": 3032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.9964,
      "step": 3033
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6108,
      "step": 3034
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7951,
      "step": 3035
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8481,
      "step": 3036
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8694,
      "step": 3037
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 1.0437,
      "step": 3038
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7478,
      "step": 3039
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9139,
      "step": 3040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7026,
      "step": 3041
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8673,
      "step": 3042
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8485,
      "step": 3043
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8066,
      "step": 3044
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9772,
      "step": 3045
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8762,
      "step": 3046
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7221,
      "step": 3047
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6303,
      "step": 3048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9211,
      "step": 3049
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7694,
      "step": 3050
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6278,
      "step": 3051
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6648,
      "step": 3052
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6128,
      "step": 3053
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8237,
      "step": 3054
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6042,
      "step": 3055
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 1.089,
      "step": 3056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6365,
      "step": 3057
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7232,
      "step": 3058
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7746,
      "step": 3059
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9566,
      "step": 3060
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6955,
      "step": 3061
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.836,
      "step": 3062
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6786,
      "step": 3063
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 1.1291,
      "step": 3064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8808,
      "step": 3065
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8526,
      "step": 3066
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7458,
      "step": 3067
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9741,
      "step": 3068
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 1.0478,
      "step": 3069
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6141,
      "step": 3070
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9822,
      "step": 3071
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 1.045,
      "step": 3072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 1.0134,
      "step": 3073
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9526,
      "step": 3074
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5294,
      "step": 3075
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5743,
      "step": 3076
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8331,
      "step": 3077
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7216,
      "step": 3078
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7696,
      "step": 3079
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.9013,
      "step": 3080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.816,
      "step": 3081
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6218,
      "step": 3082
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.801,
      "step": 3083
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8655,
      "step": 3084
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7076,
      "step": 3085
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8579,
      "step": 3086
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 1.0852,
      "step": 3087
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6902,
      "step": 3088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8049,
      "step": 3089
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7819,
      "step": 3090
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7417,
      "step": 3091
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6342,
      "step": 3092
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8023,
      "step": 3093
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9196,
      "step": 3094
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7899,
      "step": 3095
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.976,
      "step": 3096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8742,
      "step": 3097
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 3098
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7002,
      "step": 3099
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.879,
      "step": 3100
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7912,
      "step": 3101
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.666,
      "step": 3102
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5824,
      "step": 3103
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6956,
      "step": 3104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7529,
      "step": 3105
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9796,
      "step": 3106
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.966,
      "step": 3107
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8376,
      "step": 3108
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7305,
      "step": 3109
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6761,
      "step": 3110
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.3481,
      "step": 3111
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.758,
      "step": 3112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9088,
      "step": 3113
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8613,
      "step": 3114
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9453,
      "step": 3115
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5693,
      "step": 3116
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8315,
      "step": 3117
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7907,
      "step": 3118
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8539,
      "step": 3119
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4608,
      "step": 3120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7684,
      "step": 3121
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3658,
      "step": 3122
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9837,
      "step": 3123
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.0685,
      "step": 3124
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8169,
      "step": 3125
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7939,
      "step": 3126
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.0271,
      "step": 3127
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.0376,
      "step": 3128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7362,
      "step": 3129
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9716,
      "step": 3130
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6823,
      "step": 3131
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.688,
      "step": 3132
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8793,
      "step": 3133
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7037,
      "step": 3134
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5463,
      "step": 3135
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.982,
      "step": 3136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6707,
      "step": 3137
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.0237,
      "step": 3138
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.787,
      "step": 3139
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7696,
      "step": 3140
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.8265,
      "step": 3141
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.0936,
      "step": 3142
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9071,
      "step": 3143
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8625,
      "step": 3144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.722,
      "step": 3145
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8443,
      "step": 3146
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.884,
      "step": 3147
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6677,
      "step": 3148
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 1.0783,
      "step": 3149
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7543,
      "step": 3150
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8827,
      "step": 3151
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9072,
      "step": 3152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8797,
      "step": 3153
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9065,
      "step": 3154
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.904,
      "step": 3155
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9015,
      "step": 3156
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7443,
      "step": 3157
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5642,
      "step": 3158
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5282,
      "step": 3159
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7929,
      "step": 3160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 1.0713,
      "step": 3161
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 1.2525,
      "step": 3162
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9277,
      "step": 3163
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6107,
      "step": 3164
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9241,
      "step": 3165
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8229,
      "step": 3166
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6105,
      "step": 3167
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4618,
      "step": 3168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8767,
      "step": 3169
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7142,
      "step": 3170
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8898,
      "step": 3171
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7706,
      "step": 3172
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8258,
      "step": 3173
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 1.0231,
      "step": 3174
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8864,
      "step": 3175
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 1.0167,
      "step": 3176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6832,
      "step": 3177
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6807,
      "step": 3178
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8371,
      "step": 3179
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 1.0118,
      "step": 3180
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6449,
      "step": 3181
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9208,
      "step": 3182
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7697,
      "step": 3183
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8918,
      "step": 3184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9904,
      "step": 3185
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9343,
      "step": 3186
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.999,
      "step": 3187
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7611,
      "step": 3188
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9178,
      "step": 3189
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6474,
      "step": 3190
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7067,
      "step": 3191
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8426,
      "step": 3192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9448,
      "step": 3193
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7941,
      "step": 3194
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.9277,
      "step": 3195
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8394,
      "step": 3196
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8936,
      "step": 3197
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8054,
      "step": 3198
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9134,
      "step": 3199
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9001,
      "step": 3200
    },
    {
      "epoch": 0.62,
      "eval_loss": 0.8203288316726685,
      "eval_runtime": 107.9739,
      "eval_samples_per_second": 9.261,
      "eval_steps_per_second": 4.631,
      "step": 3200
    },
    {
      "epoch": 0.62,
      "mmlu_eval_accuracy": 0.5014796620409542,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.5,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.09090909090909091,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.2727272727272727,
      "mmlu_eval_accuracy_conceptual_physics": 0.3076923076923077,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.22727272727272727,
      "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": 0.4418604651162791,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.7666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.31,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.5294117647058824,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.35294117647058826,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5362318840579711,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.37037037037037035,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.5789473684210527,
      "mmlu_loss": 0.9854449025566833,
      "step": 3200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8048,
      "step": 3201
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.713,
      "step": 3202
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7551,
      "step": 3203
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9197,
      "step": 3204
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9168,
      "step": 3205
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6971,
      "step": 3206
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6683,
      "step": 3207
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7632,
      "step": 3208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8761,
      "step": 3209
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9267,
      "step": 3210
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8419,
      "step": 3211
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5485,
      "step": 3212
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5954,
      "step": 3213
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8521,
      "step": 3214
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8348,
      "step": 3215
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9298,
      "step": 3216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7824,
      "step": 3217
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7962,
      "step": 3218
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8071,
      "step": 3219
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6494,
      "step": 3220
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8515,
      "step": 3221
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8619,
      "step": 3222
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8646,
      "step": 3223
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9403,
      "step": 3224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9984,
      "step": 3225
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8697,
      "step": 3226
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8117,
      "step": 3227
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 1.0824,
      "step": 3228
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9783,
      "step": 3229
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9072,
      "step": 3230
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 1.0369,
      "step": 3231
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7868,
      "step": 3232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7621,
      "step": 3233
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9334,
      "step": 3234
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9048,
      "step": 3235
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 1.1183,
      "step": 3236
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7068,
      "step": 3237
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7554,
      "step": 3238
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6915,
      "step": 3239
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8555,
      "step": 3240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8476,
      "step": 3241
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8826,
      "step": 3242
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.802,
      "step": 3243
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8403,
      "step": 3244
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8538,
      "step": 3245
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7933,
      "step": 3246
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 3247
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8843,
      "step": 3248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8104,
      "step": 3249
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7317,
      "step": 3250
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7074,
      "step": 3251
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7344,
      "step": 3252
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 3253
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9466,
      "step": 3254
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6532,
      "step": 3255
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7269,
      "step": 3256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9125,
      "step": 3257
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6542,
      "step": 3258
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6986,
      "step": 3259
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6692,
      "step": 3260
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8612,
      "step": 3261
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9947,
      "step": 3262
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.642,
      "step": 3263
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5641,
      "step": 3264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7663,
      "step": 3265
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7064,
      "step": 3266
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8531,
      "step": 3267
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8515,
      "step": 3268
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6466,
      "step": 3269
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8266,
      "step": 3270
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7749,
      "step": 3271
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7967,
      "step": 3272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.839,
      "step": 3273
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6899,
      "step": 3274
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8346,
      "step": 3275
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7593,
      "step": 3276
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7418,
      "step": 3277
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9702,
      "step": 3278
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9786,
      "step": 3279
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6282,
      "step": 3280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7217,
      "step": 3281
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6456,
      "step": 3282
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9597,
      "step": 3283
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.834,
      "step": 3284
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7665,
      "step": 3285
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6613,
      "step": 3286
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8632,
      "step": 3287
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6328,
      "step": 3288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8461,
      "step": 3289
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5678,
      "step": 3290
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9212,
      "step": 3291
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9335,
      "step": 3292
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8,
      "step": 3293
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7253,
      "step": 3294
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8658,
      "step": 3295
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.564,
      "step": 3296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7528,
      "step": 3297
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8072,
      "step": 3298
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7911,
      "step": 3299
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.9006,
      "step": 3300
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8217,
      "step": 3301
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6853,
      "step": 3302
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.846,
      "step": 3303
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8685,
      "step": 3304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6958,
      "step": 3305
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7571,
      "step": 3306
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 1.0194,
      "step": 3307
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7473,
      "step": 3308
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 1.012,
      "step": 3309
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8604,
      "step": 3310
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8453,
      "step": 3311
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.9657,
      "step": 3312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8871,
      "step": 3313
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7822,
      "step": 3314
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.671,
      "step": 3315
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7361,
      "step": 3316
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8157,
      "step": 3317
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7099,
      "step": 3318
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8356,
      "step": 3319
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5088,
      "step": 3320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.77,
      "step": 3321
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7494,
      "step": 3322
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6109,
      "step": 3323
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8265,
      "step": 3324
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8706,
      "step": 3325
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6647,
      "step": 3326
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.9286,
      "step": 3327
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.9484,
      "step": 3328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6117,
      "step": 3329
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8205,
      "step": 3330
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8219,
      "step": 3331
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.9009,
      "step": 3332
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8476,
      "step": 3333
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8958,
      "step": 3334
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7447,
      "step": 3335
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6864,
      "step": 3336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7936,
      "step": 3337
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 1.0037,
      "step": 3338
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8569,
      "step": 3339
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8503,
      "step": 3340
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 1.1016,
      "step": 3341
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8246,
      "step": 3342
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8719,
      "step": 3343
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6196,
      "step": 3344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7195,
      "step": 3345
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6417,
      "step": 3346
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8382,
      "step": 3347
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6826,
      "step": 3348
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 1.0288,
      "step": 3349
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.8499,
      "step": 3350
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7513,
      "step": 3351
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7688,
      "step": 3352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7236,
      "step": 3353
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6513,
      "step": 3354
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7538,
      "step": 3355
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7723,
      "step": 3356
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9084,
      "step": 3357
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7984,
      "step": 3358
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9541,
      "step": 3359
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6101,
      "step": 3360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8465,
      "step": 3361
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6854,
      "step": 3362
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.591,
      "step": 3363
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.76,
      "step": 3364
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 1.0425,
      "step": 3365
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7886,
      "step": 3366
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7407,
      "step": 3367
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9116,
      "step": 3368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9396,
      "step": 3369
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7826,
      "step": 3370
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 1.0962,
      "step": 3371
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 1.0537,
      "step": 3372
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8391,
      "step": 3373
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.916,
      "step": 3374
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8173,
      "step": 3375
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7085,
      "step": 3376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.825,
      "step": 3377
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7811,
      "step": 3378
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 3379
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 1.0074,
      "step": 3380
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6799,
      "step": 3381
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9091,
      "step": 3382
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.946,
      "step": 3383
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8527,
      "step": 3384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7032,
      "step": 3385
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8555,
      "step": 3386
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6733,
      "step": 3387
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7648,
      "step": 3388
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.933,
      "step": 3389
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5233,
      "step": 3390
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9788,
      "step": 3391
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.911,
      "step": 3392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8172,
      "step": 3393
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.9803,
      "step": 3394
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7733,
      "step": 3395
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8746,
      "step": 3396
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.8486,
      "step": 3397
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.916,
      "step": 3398
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7851,
      "step": 3399
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 1.0269,
      "step": 3400
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.8125561475753784,
      "eval_runtime": 107.974,
      "eval_samples_per_second": 9.261,
      "eval_steps_per_second": 4.631,
      "step": 3400
    },
    {
      "epoch": 0.65,
      "mmlu_eval_accuracy": 0.5759145815606099,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.36363636363636365,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.7272727272727273,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "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": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.5217391304347826,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.76,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
      "mmlu_eval_accuracy_moral_disputes": 0.631578947368421,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.6470588235294118,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.35294117647058826,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5072463768115942,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
      "mmlu_eval_accuracy_sociology": 0.8181818181818182,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.3888888888888889,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8991479393551929,
      "step": 3400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6861,
      "step": 3401
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7158,
      "step": 3402
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.736,
      "step": 3403
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 1.0602,
      "step": 3404
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 1.03,
      "step": 3405
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7491,
      "step": 3406
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8093,
      "step": 3407
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6717,
      "step": 3408
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.9773,
      "step": 3409
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7336,
      "step": 3410
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5733,
      "step": 3411
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.809,
      "step": 3412
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5233,
      "step": 3413
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7855,
      "step": 3414
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.9482,
      "step": 3415
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7062,
      "step": 3416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6862,
      "step": 3417
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.926,
      "step": 3418
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.9572,
      "step": 3419
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8331,
      "step": 3420
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7681,
      "step": 3421
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.782,
      "step": 3422
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.9879,
      "step": 3423
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8135,
      "step": 3424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3852,
      "step": 3425
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8641,
      "step": 3426
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4457,
      "step": 3427
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8779,
      "step": 3428
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8967,
      "step": 3429
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.9401,
      "step": 3430
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8918,
      "step": 3431
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6811,
      "step": 3432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8442,
      "step": 3433
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6976,
      "step": 3434
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8064,
      "step": 3435
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7255,
      "step": 3436
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6566,
      "step": 3437
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8223,
      "step": 3438
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6146,
      "step": 3439
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 1.0354,
      "step": 3440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8094,
      "step": 3441
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 1.2345,
      "step": 3442
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7732,
      "step": 3443
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7612,
      "step": 3444
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8174,
      "step": 3445
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8932,
      "step": 3446
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5656,
      "step": 3447
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7289,
      "step": 3448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8909,
      "step": 3449
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7303,
      "step": 3450
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8687,
      "step": 3451
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8931,
      "step": 3452
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7083,
      "step": 3453
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4846,
      "step": 3454
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.838,
      "step": 3455
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8588,
      "step": 3456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5354,
      "step": 3457
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7602,
      "step": 3458
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6431,
      "step": 3459
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7518,
      "step": 3460
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8351,
      "step": 3461
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7377,
      "step": 3462
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6603,
      "step": 3463
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.9525,
      "step": 3464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5739,
      "step": 3465
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8416,
      "step": 3466
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6703,
      "step": 3467
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.0503,
      "step": 3468
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.0621,
      "step": 3469
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.902,
      "step": 3470
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.9035,
      "step": 3471
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6753,
      "step": 3472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7314,
      "step": 3473
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8653,
      "step": 3474
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.9395,
      "step": 3475
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7313,
      "step": 3476
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8813,
      "step": 3477
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5559,
      "step": 3478
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.0073,
      "step": 3479
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7678,
      "step": 3480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8366,
      "step": 3481
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5225,
      "step": 3482
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7702,
      "step": 3483
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.9598,
      "step": 3484
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8214,
      "step": 3485
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8572,
      "step": 3486
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8979,
      "step": 3487
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8559,
      "step": 3488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6136,
      "step": 3489
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.985,
      "step": 3490
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.639,
      "step": 3491
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8172,
      "step": 3492
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.691,
      "step": 3493
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.9415,
      "step": 3494
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6757,
      "step": 3495
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.0516,
      "step": 3496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7981,
      "step": 3497
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8233,
      "step": 3498
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.9028,
      "step": 3499
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7491,
      "step": 3500
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.0109,
      "step": 3501
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 3502
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7411,
      "step": 3503
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7745,
      "step": 3504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6562,
      "step": 3505
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.0137,
      "step": 3506
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9517,
      "step": 3507
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8593,
      "step": 3508
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9196,
      "step": 3509
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6402,
      "step": 3510
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7402,
      "step": 3511
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9172,
      "step": 3512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7093,
      "step": 3513
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.821,
      "step": 3514
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7624,
      "step": 3515
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8305,
      "step": 3516
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 1.0366,
      "step": 3517
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7253,
      "step": 3518
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.648,
      "step": 3519
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 1.1089,
      "step": 3520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.942,
      "step": 3521
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8691,
      "step": 3522
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6725,
      "step": 3523
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9567,
      "step": 3524
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8862,
      "step": 3525
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8479,
      "step": 3526
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7917,
      "step": 3527
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 1.0296,
      "step": 3528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8646,
      "step": 3529
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6045,
      "step": 3530
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9432,
      "step": 3531
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7775,
      "step": 3532
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 1.1202,
      "step": 3533
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.914,
      "step": 3534
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.639,
      "step": 3535
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7982,
      "step": 3536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9486,
      "step": 3537
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7927,
      "step": 3538
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7802,
      "step": 3539
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8293,
      "step": 3540
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.724,
      "step": 3541
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5588,
      "step": 3542
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.711,
      "step": 3543
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8779,
      "step": 3544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9088,
      "step": 3545
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7309,
      "step": 3546
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8989,
      "step": 3547
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 3548
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9089,
      "step": 3549
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8557,
      "step": 3550
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6739,
      "step": 3551
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6595,
      "step": 3552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7567,
      "step": 3553
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8863,
      "step": 3554
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.848,
      "step": 3555
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7775,
      "step": 3556
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8092,
      "step": 3557
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9254,
      "step": 3558
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7224,
      "step": 3559
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8695,
      "step": 3560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9882,
      "step": 3561
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8697,
      "step": 3562
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8749,
      "step": 3563
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8971,
      "step": 3564
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.576,
      "step": 3565
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8572,
      "step": 3566
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6785,
      "step": 3567
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8161,
      "step": 3568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7151,
      "step": 3569
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9046,
      "step": 3570
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6436,
      "step": 3571
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5488,
      "step": 3572
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6604,
      "step": 3573
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8842,
      "step": 3574
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7987,
      "step": 3575
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7872,
      "step": 3576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7358,
      "step": 3577
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7841,
      "step": 3578
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9804,
      "step": 3579
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9091,
      "step": 3580
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9659,
      "step": 3581
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8838,
      "step": 3582
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6954,
      "step": 3583
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7394,
      "step": 3584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7543,
      "step": 3585
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9031,
      "step": 3586
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.93,
      "step": 3587
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7872,
      "step": 3588
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8481,
      "step": 3589
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9,
      "step": 3590
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5039,
      "step": 3591
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6818,
      "step": 3592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6363,
      "step": 3593
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9755,
      "step": 3594
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8166,
      "step": 3595
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7076,
      "step": 3596
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.812,
      "step": 3597
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9931,
      "step": 3598
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8344,
      "step": 3599
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8165,
      "step": 3600
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.8190118670463562,
      "eval_runtime": 107.9796,
      "eval_samples_per_second": 9.261,
      "eval_steps_per_second": 4.631,
      "step": 3600
    },
    {
      "epoch": 0.69,
      "mmlu_eval_accuracy": 0.5701111517034201,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.18181818181818182,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.34615384615384615,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.9545454545454546,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8571428571428571,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.38,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.35294117647058826,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.6231884057971014,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 0.7736761521586939,
      "step": 3600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8928,
      "step": 3601
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 1.4089,
      "step": 3602
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8505,
      "step": 3603
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6933,
      "step": 3604
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9077,
      "step": 3605
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7148,
      "step": 3606
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6721,
      "step": 3607
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6426,
      "step": 3608
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9011,
      "step": 3609
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9639,
      "step": 3610
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8004,
      "step": 3611
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6992,
      "step": 3612
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8213,
      "step": 3613
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7267,
      "step": 3614
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7079,
      "step": 3615
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7329,
      "step": 3616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6676,
      "step": 3617
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7143,
      "step": 3618
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7488,
      "step": 3619
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8854,
      "step": 3620
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7882,
      "step": 3621
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6571,
      "step": 3622
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8034,
      "step": 3623
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9483,
      "step": 3624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8828,
      "step": 3625
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8699,
      "step": 3626
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8094,
      "step": 3627
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6757,
      "step": 3628
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 3629
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 1.099,
      "step": 3630
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9265,
      "step": 3631
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9114,
      "step": 3632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8891,
      "step": 3633
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8507,
      "step": 3634
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.89,
      "step": 3635
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7804,
      "step": 3636
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8288,
      "step": 3637
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.781,
      "step": 3638
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5591,
      "step": 3639
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6454,
      "step": 3640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 3641
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8536,
      "step": 3642
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9455,
      "step": 3643
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9775,
      "step": 3644
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 1.0442,
      "step": 3645
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6728,
      "step": 3646
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7347,
      "step": 3647
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8721,
      "step": 3648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8134,
      "step": 3649
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.899,
      "step": 3650
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8165,
      "step": 3651
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8974,
      "step": 3652
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9367,
      "step": 3653
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7808,
      "step": 3654
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9169,
      "step": 3655
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 1.2828,
      "step": 3656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9618,
      "step": 3657
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8014,
      "step": 3658
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5919,
      "step": 3659
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9309,
      "step": 3660
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.906,
      "step": 3661
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.8457,
      "step": 3662
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7761,
      "step": 3663
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.801,
      "step": 3664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.787,
      "step": 3665
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8847,
      "step": 3666
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 1.0727,
      "step": 3667
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8835,
      "step": 3668
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6365,
      "step": 3669
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 1.0369,
      "step": 3670
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8553,
      "step": 3671
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.789,
      "step": 3672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6848,
      "step": 3673
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8462,
      "step": 3674
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8513,
      "step": 3675
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.757,
      "step": 3676
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8101,
      "step": 3677
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 1.0528,
      "step": 3678
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6893,
      "step": 3679
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8252,
      "step": 3680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9557,
      "step": 3681
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7243,
      "step": 3682
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6056,
      "step": 3683
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9589,
      "step": 3684
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.73,
      "step": 3685
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8569,
      "step": 3686
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8113,
      "step": 3687
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.732,
      "step": 3688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7258,
      "step": 3689
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7594,
      "step": 3690
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6829,
      "step": 3691
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7297,
      "step": 3692
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8057,
      "step": 3693
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8363,
      "step": 3694
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9062,
      "step": 3695
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9615,
      "step": 3696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 1.0843,
      "step": 3697
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9988,
      "step": 3698
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7707,
      "step": 3699
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.809,
      "step": 3700
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9604,
      "step": 3701
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5458,
      "step": 3702
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6301,
      "step": 3703
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7821,
      "step": 3704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9472,
      "step": 3705
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8347,
      "step": 3706
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.975,
      "step": 3707
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6815,
      "step": 3708
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 1.2374,
      "step": 3709
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7343,
      "step": 3710
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.9115,
      "step": 3711
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5675,
      "step": 3712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8207,
      "step": 3713
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7558,
      "step": 3714
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8275,
      "step": 3715
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8973,
      "step": 3716
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7335,
      "step": 3717
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6125,
      "step": 3718
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8739,
      "step": 3719
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7939,
      "step": 3720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9537,
      "step": 3721
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7405,
      "step": 3722
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6134,
      "step": 3723
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5236,
      "step": 3724
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5354,
      "step": 3725
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9592,
      "step": 3726
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7938,
      "step": 3727
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6966,
      "step": 3728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8909,
      "step": 3729
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8338,
      "step": 3730
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6698,
      "step": 3731
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5807,
      "step": 3732
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8391,
      "step": 3733
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7371,
      "step": 3734
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8137,
      "step": 3735
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9164,
      "step": 3736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4544,
      "step": 3737
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6384,
      "step": 3738
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 1.0316,
      "step": 3739
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6451,
      "step": 3740
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6656,
      "step": 3741
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7284,
      "step": 3742
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8457,
      "step": 3743
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8141,
      "step": 3744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9198,
      "step": 3745
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5447,
      "step": 3746
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8759,
      "step": 3747
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8767,
      "step": 3748
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.835,
      "step": 3749
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9485,
      "step": 3750
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7922,
      "step": 3751
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.612,
      "step": 3752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7086,
      "step": 3753
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9083,
      "step": 3754
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4989,
      "step": 3755
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7845,
      "step": 3756
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8508,
      "step": 3757
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 1.0015,
      "step": 3758
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 1.0442,
      "step": 3759
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7172,
      "step": 3760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.9004,
      "step": 3761
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7481,
      "step": 3762
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8149,
      "step": 3763
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.956,
      "step": 3764
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7893,
      "step": 3765
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8403,
      "step": 3766
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8708,
      "step": 3767
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8063,
      "step": 3768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5505,
      "step": 3769
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7515,
      "step": 3770
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9556,
      "step": 3771
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8703,
      "step": 3772
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7411,
      "step": 3773
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9373,
      "step": 3774
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.582,
      "step": 3775
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6631,
      "step": 3776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7543,
      "step": 3777
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7449,
      "step": 3778
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8922,
      "step": 3779
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9301,
      "step": 3780
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8565,
      "step": 3781
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6146,
      "step": 3782
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8204,
      "step": 3783
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7027,
      "step": 3784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9535,
      "step": 3785
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7605,
      "step": 3786
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.817,
      "step": 3787
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8344,
      "step": 3788
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9315,
      "step": 3789
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7634,
      "step": 3790
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7785,
      "step": 3791
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7717,
      "step": 3792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9455,
      "step": 3793
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 1.1881,
      "step": 3794
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9221,
      "step": 3795
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9936,
      "step": 3796
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9299,
      "step": 3797
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8468,
      "step": 3798
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6633,
      "step": 3799
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7883,
      "step": 3800
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.820205807685852,
      "eval_runtime": 107.9092,
      "eval_samples_per_second": 9.267,
      "eval_steps_per_second": 4.634,
      "step": 3800
    },
    {
      "epoch": 0.73,
      "mmlu_eval_accuracy": 0.5718211338252034,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.5,
      "mmlu_eval_accuracy_business_ethics": 0.36363636363636365,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.5,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.6363636363636364,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7093023255813954,
      "mmlu_eval_accuracy_moral_disputes": 0.631578947368421,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.4444444444444444,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 0.8523663388803484,
      "step": 3800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9223,
      "step": 3801
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8129,
      "step": 3802
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 1.255,
      "step": 3803
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9389,
      "step": 3804
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.9257,
      "step": 3805
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.687,
      "step": 3806
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 1.1002,
      "step": 3807
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.792,
      "step": 3808
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7536,
      "step": 3809
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8713,
      "step": 3810
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 1.0122,
      "step": 3811
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8524,
      "step": 3812
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6172,
      "step": 3813
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7279,
      "step": 3814
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6863,
      "step": 3815
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6556,
      "step": 3816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7354,
      "step": 3817
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8721,
      "step": 3818
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6743,
      "step": 3819
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8127,
      "step": 3820
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 1.0172,
      "step": 3821
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8414,
      "step": 3822
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 1.261,
      "step": 3823
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9201,
      "step": 3824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8414,
      "step": 3825
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7791,
      "step": 3826
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 1.0257,
      "step": 3827
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6868,
      "step": 3828
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8814,
      "step": 3829
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6735,
      "step": 3830
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7997,
      "step": 3831
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 1.541,
      "step": 3832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9176,
      "step": 3833
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5865,
      "step": 3834
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7504,
      "step": 3835
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8219,
      "step": 3836
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7177,
      "step": 3837
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7121,
      "step": 3838
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6279,
      "step": 3839
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8335,
      "step": 3840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8002,
      "step": 3841
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7154,
      "step": 3842
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7241,
      "step": 3843
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7542,
      "step": 3844
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8871,
      "step": 3845
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8838,
      "step": 3846
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9575,
      "step": 3847
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5727,
      "step": 3848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6634,
      "step": 3849
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9587,
      "step": 3850
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7522,
      "step": 3851
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8465,
      "step": 3852
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 1.0198,
      "step": 3853
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5812,
      "step": 3854
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7911,
      "step": 3855
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8589,
      "step": 3856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7468,
      "step": 3857
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7504,
      "step": 3858
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7628,
      "step": 3859
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.698,
      "step": 3860
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8974,
      "step": 3861
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.8701,
      "step": 3862
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9051,
      "step": 3863
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7098,
      "step": 3864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7049,
      "step": 3865
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9851,
      "step": 3866
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7782,
      "step": 3867
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.9136,
      "step": 3868
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6421,
      "step": 3869
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7885,
      "step": 3870
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7346,
      "step": 3871
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.789,
      "step": 3872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7492,
      "step": 3873
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9007,
      "step": 3874
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8005,
      "step": 3875
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7657,
      "step": 3876
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8303,
      "step": 3877
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.95,
      "step": 3878
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8703,
      "step": 3879
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5985,
      "step": 3880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6708,
      "step": 3881
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4434,
      "step": 3882
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7686,
      "step": 3883
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6875,
      "step": 3884
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 1.2572,
      "step": 3885
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8105,
      "step": 3886
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6134,
      "step": 3887
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8419,
      "step": 3888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9802,
      "step": 3889
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 1.0512,
      "step": 3890
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7326,
      "step": 3891
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7864,
      "step": 3892
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7604,
      "step": 3893
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7117,
      "step": 3894
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.74,
      "step": 3895
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6474,
      "step": 3896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6942,
      "step": 3897
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7255,
      "step": 3898
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 1.0974,
      "step": 3899
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8283,
      "step": 3900
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7745,
      "step": 3901
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8317,
      "step": 3902
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9907,
      "step": 3903
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4976,
      "step": 3904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7092,
      "step": 3905
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6622,
      "step": 3906
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9098,
      "step": 3907
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8724,
      "step": 3908
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9539,
      "step": 3909
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6616,
      "step": 3910
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 1.0611,
      "step": 3911
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8978,
      "step": 3912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8807,
      "step": 3913
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6667,
      "step": 3914
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9238,
      "step": 3915
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7375,
      "step": 3916
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.9801,
      "step": 3917
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6797,
      "step": 3918
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7256,
      "step": 3919
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8989,
      "step": 3920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7679,
      "step": 3921
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.662,
      "step": 3922
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6603,
      "step": 3923
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9439,
      "step": 3924
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7535,
      "step": 3925
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.399,
      "step": 3926
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5882,
      "step": 3927
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9179,
      "step": 3928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7584,
      "step": 3929
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8492,
      "step": 3930
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7141,
      "step": 3931
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.795,
      "step": 3932
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8074,
      "step": 3933
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.868,
      "step": 3934
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4912,
      "step": 3935
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 1.0423,
      "step": 3936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8398,
      "step": 3937
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8081,
      "step": 3938
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8666,
      "step": 3939
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 1.0106,
      "step": 3940
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9221,
      "step": 3941
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8795,
      "step": 3942
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6879,
      "step": 3943
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6755,
      "step": 3944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9785,
      "step": 3945
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8128,
      "step": 3946
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7456,
      "step": 3947
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7358,
      "step": 3948
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8538,
      "step": 3949
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9414,
      "step": 3950
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8019,
      "step": 3951
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6728,
      "step": 3952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7793,
      "step": 3953
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9775,
      "step": 3954
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.747,
      "step": 3955
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8564,
      "step": 3956
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8589,
      "step": 3957
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8922,
      "step": 3958
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9449,
      "step": 3959
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8319,
      "step": 3960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8275,
      "step": 3961
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9519,
      "step": 3962
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8512,
      "step": 3963
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.708,
      "step": 3964
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7343,
      "step": 3965
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7775,
      "step": 3966
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 1.0233,
      "step": 3967
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8113,
      "step": 3968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9848,
      "step": 3969
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7191,
      "step": 3970
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.9043,
      "step": 3971
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.624,
      "step": 3972
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7307,
      "step": 3973
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 1.1121,
      "step": 3974
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6765,
      "step": 3975
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6316,
      "step": 3976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 1.0722,
      "step": 3977
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6493,
      "step": 3978
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8608,
      "step": 3979
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8662,
      "step": 3980
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9331,
      "step": 3981
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9045,
      "step": 3982
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7782,
      "step": 3983
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8305,
      "step": 3984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9335,
      "step": 3985
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7542,
      "step": 3986
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9233,
      "step": 3987
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.901,
      "step": 3988
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8528,
      "step": 3989
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8379,
      "step": 3990
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.797,
      "step": 3991
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8037,
      "step": 3992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8963,
      "step": 3993
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7475,
      "step": 3994
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7335,
      "step": 3995
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9124,
      "step": 3996
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9603,
      "step": 3997
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9771,
      "step": 3998
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8798,
      "step": 3999
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9862,
      "step": 4000
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.8221142292022705,
      "eval_runtime": 107.8144,
      "eval_samples_per_second": 9.275,
      "eval_steps_per_second": 4.638,
      "step": 4000
    },
    {
      "epoch": 0.77,
      "mmlu_eval_accuracy": 0.5600923329678102,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "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": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.34615384615384615,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "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": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.7272727272727273,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.36470588235294116,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 0.8262668283700009,
      "step": 4000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8867,
      "step": 4001
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7334,
      "step": 4002
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6826,
      "step": 4003
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.893,
      "step": 4004
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7477,
      "step": 4005
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9147,
      "step": 4006
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9845,
      "step": 4007
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6748,
      "step": 4008
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8786,
      "step": 4009
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 1.0767,
      "step": 4010
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6939,
      "step": 4011
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7368,
      "step": 4012
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.911,
      "step": 4013
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8293,
      "step": 4014
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.9974,
      "step": 4015
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4812,
      "step": 4016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8483,
      "step": 4017
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8799,
      "step": 4018
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6526,
      "step": 4019
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7268,
      "step": 4020
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 1.0066,
      "step": 4021
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8232,
      "step": 4022
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7561,
      "step": 4023
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8018,
      "step": 4024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8316,
      "step": 4025
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6801,
      "step": 4026
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7252,
      "step": 4027
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6324,
      "step": 4028
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.0119,
      "step": 4029
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9082,
      "step": 4030
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9898,
      "step": 4031
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5007,
      "step": 4032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.816,
      "step": 4033
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9198,
      "step": 4034
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9887,
      "step": 4035
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7643,
      "step": 4036
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6865,
      "step": 4037
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7396,
      "step": 4038
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7899,
      "step": 4039
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8261,
      "step": 4040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8384,
      "step": 4041
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.0225,
      "step": 4042
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8498,
      "step": 4043
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9356,
      "step": 4044
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6719,
      "step": 4045
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6078,
      "step": 4046
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5892,
      "step": 4047
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9048,
      "step": 4048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7285,
      "step": 4049
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8593,
      "step": 4050
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9923,
      "step": 4051
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8972,
      "step": 4052
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 4053
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8651,
      "step": 4054
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9147,
      "step": 4055
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.2438,
      "step": 4056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8914,
      "step": 4057
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7692,
      "step": 4058
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5885,
      "step": 4059
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7558,
      "step": 4060
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.0149,
      "step": 4061
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.003,
      "step": 4062
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9532,
      "step": 4063
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9818,
      "step": 4064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8491,
      "step": 4065
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9772,
      "step": 4066
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7802,
      "step": 4067
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.676,
      "step": 4068
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9319,
      "step": 4069
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.0225,
      "step": 4070
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8158,
      "step": 4071
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8873,
      "step": 4072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6674,
      "step": 4073
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6562,
      "step": 4074
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7031,
      "step": 4075
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8637,
      "step": 4076
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 1.0801,
      "step": 4077
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.9089,
      "step": 4078
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8394,
      "step": 4079
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5542,
      "step": 4080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9428,
      "step": 4081
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9019,
      "step": 4082
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7525,
      "step": 4083
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7045,
      "step": 4084
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 1.0369,
      "step": 4085
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8201,
      "step": 4086
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6975,
      "step": 4087
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6207,
      "step": 4088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8521,
      "step": 4089
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5928,
      "step": 4090
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 1.0153,
      "step": 4091
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7057,
      "step": 4092
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.771,
      "step": 4093
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 1.009,
      "step": 4094
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6351,
      "step": 4095
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.793,
      "step": 4096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 1.0179,
      "step": 4097
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8492,
      "step": 4098
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7557,
      "step": 4099
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.812,
      "step": 4100
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9738,
      "step": 4101
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6942,
      "step": 4102
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8358,
      "step": 4103
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7912,
      "step": 4104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8334,
      "step": 4105
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7372,
      "step": 4106
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6961,
      "step": 4107
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8259,
      "step": 4108
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7566,
      "step": 4109
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8081,
      "step": 4110
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9948,
      "step": 4111
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6058,
      "step": 4112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9171,
      "step": 4113
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7995,
      "step": 4114
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.615,
      "step": 4115
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5189,
      "step": 4116
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7135,
      "step": 4117
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6549,
      "step": 4118
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9093,
      "step": 4119
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8177,
      "step": 4120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8415,
      "step": 4121
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6414,
      "step": 4122
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.9301,
      "step": 4123
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7335,
      "step": 4124
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.82,
      "step": 4125
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.847,
      "step": 4126
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5162,
      "step": 4127
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.8265,
      "step": 4128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7324,
      "step": 4129
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7617,
      "step": 4130
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.72,
      "step": 4131
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7508,
      "step": 4132
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 1.0467,
      "step": 4133
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9665,
      "step": 4134
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8059,
      "step": 4135
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7991,
      "step": 4136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6642,
      "step": 4137
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8695,
      "step": 4138
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7287,
      "step": 4139
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7461,
      "step": 4140
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.725,
      "step": 4141
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7178,
      "step": 4142
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8157,
      "step": 4143
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9665,
      "step": 4144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7622,
      "step": 4145
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 1.0165,
      "step": 4146
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7412,
      "step": 4147
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8618,
      "step": 4148
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7178,
      "step": 4149
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7646,
      "step": 4150
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9048,
      "step": 4151
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9065,
      "step": 4152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8277,
      "step": 4153
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8752,
      "step": 4154
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6315,
      "step": 4155
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7579,
      "step": 4156
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8282,
      "step": 4157
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7334,
      "step": 4158
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 1.0306,
      "step": 4159
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9014,
      "step": 4160
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9001,
      "step": 4161
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7869,
      "step": 4162
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7677,
      "step": 4163
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 1.1934,
      "step": 4164
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6869,
      "step": 4165
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7307,
      "step": 4166
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6351,
      "step": 4167
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.75,
      "step": 4168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6447,
      "step": 4169
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9451,
      "step": 4170
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7291,
      "step": 4171
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.778,
      "step": 4172
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5472,
      "step": 4173
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5262,
      "step": 4174
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 1.2268,
      "step": 4175
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7339,
      "step": 4176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8139,
      "step": 4177
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9853,
      "step": 4178
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9272,
      "step": 4179
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8554,
      "step": 4180
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6984,
      "step": 4181
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.9198,
      "step": 4182
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.771,
      "step": 4183
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 1.1079,
      "step": 4184
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.675,
      "step": 4185
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7779,
      "step": 4186
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9008,
      "step": 4187
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6691,
      "step": 4188
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8981,
      "step": 4189
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8293,
      "step": 4190
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6914,
      "step": 4191
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5825,
      "step": 4192
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8471,
      "step": 4193
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8446,
      "step": 4194
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 1.0037,
      "step": 4195
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6958,
      "step": 4196
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 4197
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8914,
      "step": 4198
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7749,
      "step": 4199
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8109,
      "step": 4200
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.8209733963012695,
      "eval_runtime": 107.9479,
      "eval_samples_per_second": 9.264,
      "eval_steps_per_second": 4.632,
      "step": 4200
    },
    {
      "epoch": 0.81,
      "mmlu_eval_accuracy": 0.5458076663900231,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "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": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.75,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.46875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5116279069767442,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.5217391304347826,
      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.6923076923076923,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.8,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.26,
      "mmlu_eval_accuracy_nutrition": 0.5151515151515151,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.4,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.36470588235294116,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.6231884057971014,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "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": 0.7894736842105263,
      "mmlu_loss": 1.2268626109037948,
      "step": 4200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7857,
      "step": 4201
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 4202
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8666,
      "step": 4203
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.581,
      "step": 4204
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8737,
      "step": 4205
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7913,
      "step": 4206
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7143,
      "step": 4207
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9991,
      "step": 4208
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7904,
      "step": 4209
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8881,
      "step": 4210
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9593,
      "step": 4211
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7709,
      "step": 4212
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8464,
      "step": 4213
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9528,
      "step": 4214
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6889,
      "step": 4215
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6915,
      "step": 4216
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6666,
      "step": 4217
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6561,
      "step": 4218
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.809,
      "step": 4219
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7873,
      "step": 4220
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6726,
      "step": 4221
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8009,
      "step": 4222
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.701,
      "step": 4223
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9206,
      "step": 4224
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7083,
      "step": 4225
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9042,
      "step": 4226
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 1.0602,
      "step": 4227
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9032,
      "step": 4228
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6832,
      "step": 4229
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.9716,
      "step": 4230
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8111,
      "step": 4231
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.994,
      "step": 4232
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7791,
      "step": 4233
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6293,
      "step": 4234
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9417,
      "step": 4235
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8317,
      "step": 4236
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8168,
      "step": 4237
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9155,
      "step": 4238
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.848,
      "step": 4239
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8142,
      "step": 4240
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 1.0177,
      "step": 4241
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7935,
      "step": 4242
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8016,
      "step": 4243
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8109,
      "step": 4244
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8377,
      "step": 4245
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7329,
      "step": 4246
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9835,
      "step": 4247
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8914,
      "step": 4248
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 1.161,
      "step": 4249
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8531,
      "step": 4250
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8545,
      "step": 4251
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 1.0484,
      "step": 4252
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8509,
      "step": 4253
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8513,
      "step": 4254
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6977,
      "step": 4255
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8502,
      "step": 4256
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8315,
      "step": 4257
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8548,
      "step": 4258
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 1.1189,
      "step": 4259
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9019,
      "step": 4260
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9567,
      "step": 4261
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9088,
      "step": 4262
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4766,
      "step": 4263
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8536,
      "step": 4264
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 1.1026,
      "step": 4265
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8646,
      "step": 4266
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9795,
      "step": 4267
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5778,
      "step": 4268
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9357,
      "step": 4269
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 4270
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7539,
      "step": 4271
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.915,
      "step": 4272
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7837,
      "step": 4273
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8988,
      "step": 4274
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7436,
      "step": 4275
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.783,
      "step": 4276
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8758,
      "step": 4277
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6019,
      "step": 4278
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8301,
      "step": 4279
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.902,
      "step": 4280
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7531,
      "step": 4281
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8894,
      "step": 4282
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8064,
      "step": 4283
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6957,
      "step": 4284
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.9954,
      "step": 4285
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6959,
      "step": 4286
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6752,
      "step": 4287
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 1.1736,
      "step": 4288
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8235,
      "step": 4289
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7767,
      "step": 4290
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7308,
      "step": 4291
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6676,
      "step": 4292
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9154,
      "step": 4293
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7405,
      "step": 4294
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8408,
      "step": 4295
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7692,
      "step": 4296
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.921,
      "step": 4297
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9396,
      "step": 4298
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 4299
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7232,
      "step": 4300
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9413,
      "step": 4301
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7114,
      "step": 4302
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5683,
      "step": 4303
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8021,
      "step": 4304
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7654,
      "step": 4305
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 1.0343,
      "step": 4306
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9243,
      "step": 4307
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8466,
      "step": 4308
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5035,
      "step": 4309
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9171,
      "step": 4310
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7806,
      "step": 4311
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7898,
      "step": 4312
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9763,
      "step": 4313
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5787,
      "step": 4314
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9837,
      "step": 4315
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7055,
      "step": 4316
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9861,
      "step": 4317
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8911,
      "step": 4318
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6779,
      "step": 4319
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9703,
      "step": 4320
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.658,
      "step": 4321
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.833,
      "step": 4322
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6314,
      "step": 4323
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7123,
      "step": 4324
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7282,
      "step": 4325
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5779,
      "step": 4326
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8338,
      "step": 4327
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7974,
      "step": 4328
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6326,
      "step": 4329
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8585,
      "step": 4330
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8525,
      "step": 4331
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7261,
      "step": 4332
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8752,
      "step": 4333
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.909,
      "step": 4334
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.9945,
      "step": 4335
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8313,
      "step": 4336
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6582,
      "step": 4337
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7005,
      "step": 4338
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.891,
      "step": 4339
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9758,
      "step": 4340
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.1343,
      "step": 4341
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.0118,
      "step": 4342
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9566,
      "step": 4343
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.005,
      "step": 4344
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6707,
      "step": 4345
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8093,
      "step": 4346
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7232,
      "step": 4347
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.988,
      "step": 4348
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6333,
      "step": 4349
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8303,
      "step": 4350
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7988,
      "step": 4351
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8346,
      "step": 4352
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7743,
      "step": 4353
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8662,
      "step": 4354
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7506,
      "step": 4355
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7852,
      "step": 4356
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.637,
      "step": 4357
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.792,
      "step": 4358
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8164,
      "step": 4359
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.733,
      "step": 4360
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9265,
      "step": 4361
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.1875,
      "step": 4362
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6939,
      "step": 4363
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7582,
      "step": 4364
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.85,
      "step": 4365
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8048,
      "step": 4366
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.767,
      "step": 4367
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9321,
      "step": 4368
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.813,
      "step": 4369
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8069,
      "step": 4370
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7865,
      "step": 4371
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5636,
      "step": 4372
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.0375,
      "step": 4373
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9172,
      "step": 4374
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6471,
      "step": 4375
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8025,
      "step": 4376
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6284,
      "step": 4377
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.0684,
      "step": 4378
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8451,
      "step": 4379
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 1.0077,
      "step": 4380
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6536,
      "step": 4381
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7798,
      "step": 4382
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7227,
      "step": 4383
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5544,
      "step": 4384
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.865,
      "step": 4385
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9888,
      "step": 4386
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9199,
      "step": 4387
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7,
      "step": 4388
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.8729,
      "step": 4389
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.5679,
      "step": 4390
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8715,
      "step": 4391
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6599,
      "step": 4392
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8426,
      "step": 4393
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 1.0103,
      "step": 4394
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7887,
      "step": 4395
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6493,
      "step": 4396
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8011,
      "step": 4397
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 1.0113,
      "step": 4398
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.9734,
      "step": 4399
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8383,
      "step": 4400
    },
    {
      "epoch": 0.85,
      "eval_loss": 0.8198485970497131,
      "eval_runtime": 107.9191,
      "eval_samples_per_second": 9.266,
      "eval_steps_per_second": 4.633,
      "step": 4400
    },
    {
      "epoch": 0.85,
      "mmlu_eval_accuracy": 0.5513569197824587,
      "mmlu_eval_accuracy_abstract_algebra": 0.09090909090909091,
      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.46511627906976744,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.3333333333333333,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 0.8504945156760222,
      "step": 4400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8868,
      "step": 4401
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7877,
      "step": 4402
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 1.0112,
      "step": 4403
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8758,
      "step": 4404
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7286,
      "step": 4405
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6897,
      "step": 4406
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8559,
      "step": 4407
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8623,
      "step": 4408
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.9137,
      "step": 4409
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6507,
      "step": 4410
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.986,
      "step": 4411
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7837,
      "step": 4412
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7353,
      "step": 4413
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.9019,
      "step": 4414
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7218,
      "step": 4415
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6246,
      "step": 4416
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.5961,
      "step": 4417
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.897,
      "step": 4418
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6939,
      "step": 4419
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 1.0046,
      "step": 4420
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8164,
      "step": 4421
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8498,
      "step": 4422
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.787,
      "step": 4423
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6893,
      "step": 4424
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7449,
      "step": 4425
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7532,
      "step": 4426
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7267,
      "step": 4427
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7104,
      "step": 4428
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8861,
      "step": 4429
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8043,
      "step": 4430
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.9941,
      "step": 4431
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7642,
      "step": 4432
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.9499,
      "step": 4433
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8731,
      "step": 4434
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.922,
      "step": 4435
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.648,
      "step": 4436
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.834,
      "step": 4437
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7216,
      "step": 4438
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.8451,
      "step": 4439
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.6982,
      "step": 4440
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.7314,
      "step": 4441
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9021,
      "step": 4442
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.756,
      "step": 4443
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.803,
      "step": 4444
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9397,
      "step": 4445
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8824,
      "step": 4446
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7651,
      "step": 4447
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8221,
      "step": 4448
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8841,
      "step": 4449
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.725,
      "step": 4450
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8952,
      "step": 4451
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7927,
      "step": 4452
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6365,
      "step": 4453
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7979,
      "step": 4454
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8877,
      "step": 4455
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 1.008,
      "step": 4456
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7158,
      "step": 4457
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8695,
      "step": 4458
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.5716,
      "step": 4459
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9393,
      "step": 4460
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8769,
      "step": 4461
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8844,
      "step": 4462
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9459,
      "step": 4463
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9366,
      "step": 4464
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6147,
      "step": 4465
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7443,
      "step": 4466
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7938,
      "step": 4467
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 1.0524,
      "step": 4468
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6873,
      "step": 4469
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6799,
      "step": 4470
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8799,
      "step": 4471
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.765,
      "step": 4472
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7215,
      "step": 4473
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8677,
      "step": 4474
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.5468,
      "step": 4475
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6193,
      "step": 4476
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8918,
      "step": 4477
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6848,
      "step": 4478
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.848,
      "step": 4479
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9481,
      "step": 4480
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.894,
      "step": 4481
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 1.0408,
      "step": 4482
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7589,
      "step": 4483
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7012,
      "step": 4484
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9108,
      "step": 4485
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7318,
      "step": 4486
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.873,
      "step": 4487
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.7229,
      "step": 4488
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8625,
      "step": 4489
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.9901,
      "step": 4490
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6175,
      "step": 4491
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.6223,
      "step": 4492
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.8114,
      "step": 4493
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8278,
      "step": 4494
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7382,
      "step": 4495
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9622,
      "step": 4496
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8255,
      "step": 4497
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8458,
      "step": 4498
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 1.0234,
      "step": 4499
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6267,
      "step": 4500
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7769,
      "step": 4501
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8266,
      "step": 4502
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7943,
      "step": 4503
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.5807,
      "step": 4504
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6664,
      "step": 4505
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7788,
      "step": 4506
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7794,
      "step": 4507
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6964,
      "step": 4508
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 1.1599,
      "step": 4509
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8728,
      "step": 4510
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6884,
      "step": 4511
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8816,
      "step": 4512
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7736,
      "step": 4513
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6896,
      "step": 4514
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7128,
      "step": 4515
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9099,
      "step": 4516
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7547,
      "step": 4517
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 1.0283,
      "step": 4518
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9545,
      "step": 4519
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.944,
      "step": 4520
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.751,
      "step": 4521
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9541,
      "step": 4522
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7989,
      "step": 4523
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9088,
      "step": 4524
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7146,
      "step": 4525
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.5908,
      "step": 4526
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9894,
      "step": 4527
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.685,
      "step": 4528
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.4707,
      "step": 4529
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6972,
      "step": 4530
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7272,
      "step": 4531
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7192,
      "step": 4532
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8086,
      "step": 4533
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6369,
      "step": 4534
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 1.2391,
      "step": 4535
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7824,
      "step": 4536
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9334,
      "step": 4537
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7563,
      "step": 4538
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8377,
      "step": 4539
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.8715,
      "step": 4540
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7656,
      "step": 4541
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.6532,
      "step": 4542
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.9611,
      "step": 4543
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7108,
      "step": 4544
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.7664,
      "step": 4545
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9456,
      "step": 4546
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9645,
      "step": 4547
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7827,
      "step": 4548
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6588,
      "step": 4549
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.873,
      "step": 4550
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9653,
      "step": 4551
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7785,
      "step": 4552
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8918,
      "step": 4553
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.929,
      "step": 4554
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8794,
      "step": 4555
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7883,
      "step": 4556
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.5826,
      "step": 4557
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9126,
      "step": 4558
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6188,
      "step": 4559
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8263,
      "step": 4560
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8448,
      "step": 4561
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7585,
      "step": 4562
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8439,
      "step": 4563
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9045,
      "step": 4564
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7819,
      "step": 4565
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8286,
      "step": 4566
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9736,
      "step": 4567
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8868,
      "step": 4568
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7894,
      "step": 4569
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6359,
      "step": 4570
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8261,
      "step": 4571
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6165,
      "step": 4572
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.662,
      "step": 4573
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9108,
      "step": 4574
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6238,
      "step": 4575
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6855,
      "step": 4576
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7809,
      "step": 4577
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8394,
      "step": 4578
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7894,
      "step": 4579
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7226,
      "step": 4580
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 1.0343,
      "step": 4581
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.5678,
      "step": 4582
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.746,
      "step": 4583
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9877,
      "step": 4584
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.838,
      "step": 4585
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9743,
      "step": 4586
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.694,
      "step": 4587
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7826,
      "step": 4588
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8165,
      "step": 4589
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8191,
      "step": 4590
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.9511,
      "step": 4591
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8271,
      "step": 4592
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7749,
      "step": 4593
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8525,
      "step": 4594
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.6422,
      "step": 4595
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.8507,
      "step": 4596
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.7377,
      "step": 4597
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.6867,
      "step": 4598
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.8178,
      "step": 4599
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.6655,
      "step": 4600
    },
    {
      "epoch": 0.89,
      "eval_loss": 0.8238421082496643,
      "eval_runtime": 108.0185,
      "eval_samples_per_second": 9.258,
      "eval_steps_per_second": 4.629,
      "step": 4600
    },
    {
      "epoch": 0.89,
      "mmlu_eval_accuracy": 0.5258230508843268,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.4375,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.09090909090909091,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.36363636363636365,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.46875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.4418604651162791,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.75,
      "mmlu_eval_accuracy_high_school_statistics": 0.21739130434782608,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.7692307692307693,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.7272727272727273,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7093023255813954,
      "mmlu_eval_accuracy_moral_disputes": 0.6578947368421053,
      "mmlu_eval_accuracy_moral_scenarios": 0.35,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.5882352941176471,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.3411764705882353,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.6818181818181818,
      "mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.3888888888888889,
      "mmlu_eval_accuracy_world_religions": 0.5789473684210527,
      "mmlu_loss": 1.100015012581417,
      "step": 4600
    }
  ],
  "logging_steps": 1,
  "max_steps": 15585,
  "num_train_epochs": 3,
  "save_steps": 200,
  "total_flos": 6.96052878248706e+17,
  "trial_name": null,
  "trial_params": null
}